{ "status": "partial", "applicable_count": 100, "completed_count": 99, "failed_count": 1, "schema_version": 3, "config_hash": "b9c71c81b67c097c", "data": { "contis": { "S1sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S1sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L01B02.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the theater’s backstage waiting room, beside the chair used before the performance.\n\nSHOT TEXT (authoritative, Korean): 어두운 대기실 의자에 앉아 굳은 표정으로 바닥을 응시하는 김광석(광석·광돌·김광돌)의 상반신\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At seated chest height inside the dressing room, the camera completes its slow dolly toward 김광석, settling into a close upper-body frame while preserving his downward eyeline. He occupies the center-right with hunched shoulders and clasped tension, while the chair and a narrow portion of the dark waiting area remain visible around him.\n- FRAMING SCALE: medium shot\n- KEY BACKGROUND ELEMENTS: dressing-room chair (Occupied by 김광석) — Its front and one side are visible behind 김광석's bent torso; used as Supports the seated posture and remains partially visible beneath and behind the subject; dressing-room interior (Dark); used as Provides restrained spatial context around the isolated seated figure.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L01B02.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "5e7f3627-06ab-45ab-ab74-7f81a2d0f66a" }, "S1sh9": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S1sh9.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L01B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: At the theater’s backstage-to-stage threshold, facing the brightly lit stage floor.\n\nSHOT TEXT (authoritative, Korean): 밝은 조명이 쏟아지는 무대 바닥을 향해 한쪽 발을 공중에 든 채 디디려는 찰나의 김광석(광석·광돌·김광돌)\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From near stage-floor height, slightly ahead and to the side, the tracking camera suspends the slow-motion instant at the peak of 김광석's step. His lifted foot and weighted supporting leg occupy the middle and lower frame, with the brightly illuminated stage floor retained around the point where the foot will land.\n- FRAMING SCALE: insert close-up on a detail\n- KEY BACKGROUND ELEMENTS: stage floor (Brightly illuminated); used as Forms the landing plane beneath the suspended foot and indicates the direction of advance.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L01B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "7597bee5-ee05-4ba2-abd5-8e40dc851945" }, "S1sh11": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S2sh3": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S2sh3.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L06B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the neighborhood electronics repair shop, at the cluttered radio-repair workbench.\n\nSHOT TEXT (authoritative, Korean): 구겨진 천 원짜리 지폐를 쥔 광석의 아버지의 손이 김광석(광석·광돌·김광돌)의 펼친 손바닥에 맞닿은 모습\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At waist height inside the shop, the completed dolly-in isolates the exact contact between the father's offering hand and 김광석's open palm. The crumpled banknote lies across the center of the frame, with both hands and a narrow edge of the workbench providing scale and physical context.\n- FRAMING SCALE: insert close-up on a detail\n- KEY BACKGROUND ELEMENTS: crumpled one-thousand-won banknote (Crumpled and being handed over) — One printed face is visible to the camera across 김광석's palm, with the denomination legible within the creases; used as Primary focal object positioned between the two touching hands; workbench (Used as the father's repair station) — The upper working face is seen obliquely beneath the hands; used as A narrow background edge anchors the exchange within the repair workspace.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L06B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:광석의 아버지", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/b6968f65-4381-4394-9823-7a248465e015.png" ] ], "asset_id": "418c7ceb-3c56-48d1-a31c-1485d07f7616" }, "S2sh6": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S2sh10": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S2sh10.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L06B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Outside on a steep hillside-neighborhood stairway, directly behind a handcart stuck at the steps.\n\nSHOT TEXT (authoritative, Korean): 리어카 뒤편에 양손을 대고 한쪽 발에 체중을 실어 강하게 밀어붙이는 mid-action 자세의 김광석(광석·광돌·김광돌)\n\nCARRIED STATE (persist exactly): Gwangseok remains wearing the guitar bag while he pushes the cart.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From a low rear-quarter position behind and slightly beside the cart, the camera tracks parallel to 김광석's upward effort. His full extended body line runs diagonally through the frame—both hands planted on the cart, one foot bearing his weight—while the steps ahead establish the resistance he is overcoming.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: handcart (Being pushed past the obstruction) — Its rear faces the camera while its front points upward along the route; used as Primary action anchor receiving the pressure of both hands; neighborhood steps (Steep) — The steps recede upward from the lower foreground toward the upper background; used as Forms the rising path and makes the weighted posture legible; guitar bag (Carried while he pushes) — The back and side of the carried bag are visible along his bent torso; used as Maintains continuity with 김광석's departure from the shop.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L06B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "0f186e6f-628e-4f98-9000-7786f1b385ae" }, "S3sh4": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S4sh4": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S4sh4.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L11B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the tiered university lecture hall, among rows of seated students.\n\nSHOT TEXT (authoritative, Korean): 학생들 틈바구니에서 스프링처럼 벌떡 일어선 채 바짝 얼어붙은 자세의 김광석(광석·광돌·김광돌)\n\nCARRIED STATE (persist exactly): Gwangseok is still carrying his guitar in its bag as he stands among the students.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From student eye level within the seating area, the lateral-forward track steadies at its endpoint to hold 김광석 abruptly risen among seated classmates. His rigid figure occupies the middle of the frame with the upright guitar bag extending behind him, while staggered desks and seated students preserve the tiered classroom depth toward the off-screen professor.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: tiered classroom desks (Occupied by seated students) — Desktop fronts run diagonally across the frame and recede toward the classroom front; used as Creates repeating foreground and background layers around the suddenly risen student; guitar bag (Carried upright on his back) — Its long profile rises above the seated row behind 김광석; used as Makes the professor's sailboat comparison visually legible behind 김광석.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L11B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "ccb74095-f973-4c8a-8b6d-4e1c9937146a" }, "S4sh9": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S5sh6": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S5sh6.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L12B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the music café’s back office, beside the manager’s desk where payment is being handed over.\n\nSHOT TEXT (authoritative, Korean): 업소부장의 손아귀에 쥔 지폐 끝을 꼬집듯 잡아챈 채 서로 팽팽하게 잡아당기고 있는 김광석(광석·광돌·김광돌)의 자세\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From waist height between the two men, dolly into a tight close frame with the taut banknotes centered and both opposing hands fully legible. 김광석's pinching grip and braced torso occupy the left side while 한국인 업소부장's clenched hand enters from the right, keeping their faces outside the upper boundary and making the physical stalemate the sole emphasis.\n- FRAMING SCALE: close-up\n- KEY BACKGROUND ELEMENTS: 지폐 (Held taut between 김광석 and 한국인 업소부장) — The printed face lies obliquely toward the camera between the two grips, with both ends visible; used as Central focal object connecting the opposing hands and revealing the struggle.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L12B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "95147db6-8b13-49d3-910b-6f9482aeed90" }, "S5sh10": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S5sh13": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S5sh13.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L12B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: At the music café’s front counter telephone, with the noisy venue audible around it.\n\nSHOT TEXT (authoritative, Korean): 눈을 반짝이며 수화기에 대고 쾌활한 기색으로 큰 입 모양을 한 채 고정된 김광석(광석·광돌·김광돌)의 클로즈업\n\nCARRIED STATE (persist exactly): The money remains in Gwangseok's front shirt pocket while he takes the call.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Complete the eye-level dolly-in at the counter as a fixed close-up, with 김광석's animated eyes and widely shaped mouth centered beside the telephone receiver. His receiver hand stays along one edge and the hand covering his other ear enters from the opposite side, isolating his sudden buoyancy from the office behind him.\n- FRAMING SCALE: close-up\n- FRAME LAYOUT: 김광석 in the middle-center of the frame, foreground.\n- KEY BACKGROUND ELEMENTS: 전화 수화기 (Held in use during the call) — Its speaking end angles toward 김광석's mouth while the listening end rests at his ear; used as Visible beside 김광석's cheek as the immediate context for his loud speech; 카운터 (In use as the telephone station) — Only its upper edge is visible behind and below 김광석; used as Soft contextual boundary behind the close-up.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L12B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "d6775b66-ce1f-4753-9a7d-17539a6bb608" }, "S6sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S6sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L13B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: At a table along the dance hall’s edge, looking toward the crowded stage floor.\n\nSHOT TEXT (authoritative, Korean): 홀 구석 테이블에 턱을 괸 채 맥주잔을 앞에 두고 심드렁한 눈빛으로 스테이지를 쳐다보는 김광석(광석·광돌·김광돌)\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At seated eye level across the corner table, hold a static medium composition with 김광석 offset to one side, his cheek weighted into his palm and the beer glass between him and the lens. Preserve the stage beyond his turned eyeline so his stillness reads against the activity he watches, without shifting the established table axis.\n- FRAMING SCALE: medium shot\n- FRAME LAYOUT: 김광석 in the middle-left of the frame, midground, looks toward performance stage; performance stage in the middle-right of the frame, background.\n- KEY BACKGROUND ELEMENTS: 홀 구석 테이블 (Occupied by 김광석) — The near edge faces camera, with 김광석 seated across it; used as Foreground spatial divider supporting 김광석's slouched posture; 맥주잔 (Placed in front of 김광석); used as Small foreground cue to his passive drinking; 스테이지 (Occupied by dancing university students) — Its audience-facing side lies beyond 김광석's line of sight; used as Background destination of 김광석's gaze.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L13B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "7cb777bc-a82d-45ef-b2a8-a47a51b3215b" }, "S6sh6": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S6sh10": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S7sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S7sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L14B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Around a food-and-drink table inside the smoky student tavern.\n\nSHOT TEXT (authoritative, Korean): 자욱한 담배 연기 속, 안주와 막걸리가 놓인 테이블에 둘러앉은 김창기(창기), 유준호(준호·준열), 박경찬(경찬), 김광석(광석·광돌·김광돌)의 와이드 샷\n\nCARRIED STATE (persist exactly): The makgeolli and shared dishes remain arranged on the table amid the cigarette smoke.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From seated eye level on the open side of the table, hold a static wide frame containing all four men and the complete shared tabletop arrangement. The table forms a low visual anchor while 김창기, 유준호, 박경찬, and 김광석 occupy distinct positions around it, their varied seated gestures emerging through the dense smoke without privileging one introduction yet.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 테이블 (Set with food and drink) — Its open side faces the camera, allowing all four seated positions and the tabletop to remain visible; used as Central spatial anchor binding the four men into one conversational group; 막걸리와 안주 (김치찌개, 두부김치, 도토리묵과 함께 arranged on the table); used as Visible tabletop spread establishing the drinking gathering; 담배 연기 (Dense throughout the room); used as Atmospheric layer softening the depth between the camera and the group.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L14B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ] ], "asset_id": "35a3a8ca-5722-4ff1-896d-dbabca5dbb54" }, "S7sh8": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S7sh13": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S8sh8": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S8sh8.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_university_library_exterior_sel.png", "prompt": "The attached image is the LOCATION PHOTOGRAPH (loose spatial reference only).\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: On the outdoor approach immediately in front of the university library entrance, beneath falling leaflets.\n\nSHOT TEXT (authoritative, Korean): 허공에 수많은 전단이 흩날리는 아래로, 도서관 입구를 향해 한 발은 땅을 강하게 디디고 다른 발은 굽혀 든 전력 질주 자세의 무리의 한국인 대학생들\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Track at waist height parallel to the charging 한국인 대학생 무리, using a wide side view that catches front feet striking the route and trailing knees bent beneath them. Keep the runners traveling laterally toward the 도서관 입구 in the background while falling leaflets occupy the upper portion of the frame, preserving enough campus space to make their destination and collective momentum clear.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 전단 (Falling through the air) — Multiple printed faces and reverse sides turn toward the camera at different angles as they fall; specific text is not legible in this framing; used as Upper-frame moving layer above the running students; 도서관 입구 (Students are running toward it) — The entrance-facing side is visible ahead of the runners along their route; used as Background destination anchoring the students' running direction; 도서관 앞 이동로 (Occupied by the running student group); used as Ground plane defining the lateral tracking path and footfalls.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_university_library_exterior_sel.png" ] ], "asset_id": "5eaf38ad-ec64-4682-b007-d8c12e036328" }, "S8sh14": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S8sh17": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S9sh10": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S9sh10.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L17B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: At center stage inside the dim beer hall, under a focused performance light.\n\nSHOT TEXT (authoritative, Korean): 화려한 핀 조명을 받으며 무대 중앙에서 눈을 감고 입을 벌려 노래하는 정승찬(승찬)의 상반신\n\nCARRIED STATE (persist exactly): Gwangseok still has the thin guitar case containing the guitar with its broken neck.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Cut to the frontal stage axis at chest height and tighten to 정승찬's upper body, keeping him centered while his lifted chest and open mouth register the sustained lyric. Use a steady performance portrait with shallow background separation, preserving the direct relationship between singer and unseen audience without turning the moment into a posed still.\n- FRAMING SCALE: medium shot\n- KEY BACKGROUND ELEMENTS: stage (occupied by 정승찬) — The front performance side faces the camera on the central axis; used as Provides the centered performance plane behind the singer.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L17B01.png" ], [ "char:정승찬", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/ece62b7c-1dd1-4871-8e5c-0ef3c926acc5.png" ] ], "asset_id": "5fd6b8b4-fe2b-4149-94c3-5c074b6da06b" }, "S9sh13": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S9sh21": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S10sh3": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S10sh3.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L18B01.png", "prompt": "The attached image is the LOCATION PHOTOGRAPH (loose spatial reference only).\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: At the sales counter inside the instrument shop, over an opened guitar case.\n\nSHOT TEXT (authoritative, Korean): 매대 위에 놓인 부러진 기타의 목 부분을 심각한 표정으로 굽어살펴보는 한국인 악기점 주인\n\nCARRIED STATE (persist exactly): The guitar remains out of its case with the neck broken off, while the Me-ari songbook that fell from the case remains nearby.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From the aisle side of the counter, finish the dolly-in just beyond the flow's final distance and angle slightly downward into a medium three-quarter view of the 한국인 악기점 주인. Place his bent face in the upper half and the broken neck directly beneath his gaze in the lower center, preserving enough of the counter to make the examination tactile and credible.\n- FRAMING SCALE: medium shot\n- KEY BACKGROUND ELEMENTS: shop counter (supporting the broken guitar) — Its aisle-facing edge runs across the foreground while the top recedes toward the owner; used as Carries the damaged instrument beneath the owner's examination; broken guitar neck (separated from the guitar body) — Both fractured sections lie across the counter with the break visible to camera; used as The damaged section is retained directly beneath the owner's gaze.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L18B01.png" ] ], "asset_id": "f4558dbe-3721-40fd-9b6b-322b772d8c9a" }, "S10sh11": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S10sh14": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S11sh6": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S11sh6.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: Inside the small storage-room conversion used as a bedroom, against its close wall.\n\nSHOT TEXT (authoritative, Korean): 창고를 개조한 좁은 방의 벽에 등을 기댄 채 굳은 표정으로 통기타 줄을 막 튕겨낸 mid-action 순간의 김광석(광석·광돌·김광돌)의 상반신\n\nCARRIED STATE (persist exactly): Gwangseok has the guitar in his arms and the Me-ari songbook remains with him in the room.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Begin the inward dolly from seated chest height in a medium upper-body view, with 김광석 backed against the wall and the guitar spanning the lower half of his torso. Freeze the keyframe at the instant his striking hand clears the strings, using a slight three-quarter axis so both his hardened expression and the completed gesture remain readable.\n- FRAMING SCALE: medium shot\n- KEY BACKGROUND ELEMENTS: wall of the converted room (directly behind 김광석); used as Provides the close spatial boundary behind 김광석 and emphasizes the room's confinement; acoustic guitar (held in playing position) — Its front is angled toward camera while the strings lie beneath his raised hand; used as Occupies the lower portion of the portrait and records the final chord gesture.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "270f4f24-693a-4062-bdcd-028da2dd0004" }, "S11sh8": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S12sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S12sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L24B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: In the building’s interior club-room corridor, directly outside the door marked 207.\n\nSHOT TEXT (authoritative, Korean): 동아리 방들이 늘어선 실내 복도 한가운데서 '207호' 팻말이 붙은 문을 올려다보며 굳어있는 김광석(광석·광돌·김광돌)의 모습\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From the corridor center at chest height, hold a static medium view with 김광석 in the middle-left and the closed 207 door in the right midground. Keep him in a hesitant three-quarter profile as his stopped body and lifted gaze form a diagonal toward the room-number sign above the door.\n- FRAMING SCALE: medium shot\n- FRAME LAYOUT: 김광석 in the middle-left of the frame, midground, looks toward 207 room-number sign; 207 room-number sign in the upper-right of the frame, background.\n- KEY BACKGROUND ELEMENTS: 207호 출입문 (closed) — The corridor-facing side and its handle side are visible to camera; used as Provides the destination beside 김광석 and anchors his hesitation; 207호 팻말 (attached above the door) — Its written front face bearing '207호' is visible to camera; used as Creates the elevated visual target at the end of 김광석's eyeline; 복도와 동아리방 출입문들 (visible) — The doors recede along both sides of the corridor; used as Extends behind 김광석 to establish the row of club rooms.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L24B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "b0d0a2b2-31c8-4a34-b967-b1c1755fd330" }, "S13sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S13sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L25B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside a modest club room, between the open doorway and the large seminar table; daylight reaches the room from the building interior.\n\nSHOT TEXT (authoritative, Korean): '고뇌하는 마음으로 노래를!' 구호가 적힌 칠판 앞 커다란 책상에 모여 앉은 한국인 대학생들 뒤로, 열린 문간에 뻘쭘하게 서 있는 김광석(광석·광돌·김광돌)\n\nCARRIED STATE (persist exactly): The slogan “Sing with an anguished heart!” remains written on the corner blackboard.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From the opposite corner at standing chest height, hold a frontal wide view across the large table, placing the seated seminar group through the lower and middle frame, the slogan-bearing blackboard behind them, and 김광석 isolated in the open doorway at the rear edge. Let the doorway frame his awkward pause while the table binds the other students into a compact discussion circle.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: 김광석 in the upper-right of the frame, background; large seminar table in the middle-center of the frame, midground; open doorway in the upper-right of the frame, background.\n- KEY BACKGROUND ELEMENTS: 열린 출입문 (open) — The opening faces the room camera axis, with 김광석 held inside it; used as Frames 김광석 separately from the established seminar group; 칠판 (slogan written on one side) — The written face showing '고뇌하는 마음으로 노래를!' is visible to camera; used as Carries the ideological slogan above the seated group; 커다란 책상 (surrounded by seated students) — Its long side runs across the lower-middle frame; used as Organizes the students into a single seminar formation; 세미나 유인물 (spread on the table) — Printed faces lie upward on the tabletop, with text present but not isolated for reading at this distance; used as Provides visible evidence of the seminar activity without becoming the primary subject.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L25B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "b6926608-7c44-4f30-a8f0-4714142c1ef0" }, "S13sh7": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S14sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S14sh1.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: Around a low dining table inside the smoke-filled Chinese restaurant.\n\nSHOT TEXT (authoritative, Korean): 자욱한 담배 연기로 가득 찬 중국집의 앉은뱅이 상을 둘러싼 한국인 대학생들 사이에서, 눈을 지그시 감은 채 노래하는 송영숙(영숙)의 상반신\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At seated eye level beyond the low table, complete the slow approach in a medium upper-body view of 송영숙, centered slightly off-axis with her eyes gently closed as she sings. Preserve partial silhouettes and shoulders of seated students at both frame edges, allowing the smoke between camera and singer to soften depth without obscuring her expression.\n- FRAMING SCALE: medium shot\n- KEY BACKGROUND ELEMENTS: 자욱한 담배 연기 (filling the room); used as Creates visible depth layers between the camera, surrounding students, and 송영숙; 앉은뱅이 상 (surrounded by seated members) — Its near edge cuts across the lower frame, viewed from the outer side of the group; used as Establishes the low seating arrangement beneath the singer and surrounding members.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "d786eed9-630e-44bc-bf81-1c4bf7673bef" }, "S14sh7": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S15sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S15sh1.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: On the front platform of the night-school classroom, directly before the banner-covered blackboard.\n\nSHOT TEXT (authoritative, Korean): 칠판에 붙은 '제5회 민들레 야학 졸업식' 현수막 앞, 단상에 서 있는 김광석(광석·광돌·김광돌), 송영숙(영숙), 정승찬(승찬), 한국인 회원들의 와이드 샷\n\nCARRIED STATE (persist exactly): Youngsook is wearing the handkerchief around her neck at the start of the performance.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From the rear center of the classroom at standing eye level, hold a straight-on wide composition with the platform spanning the middle frame and the handwritten graduation banner centered above it. Place 정승찬 slightly forward of the choir line while 김광석, 송영숙, and the other members remain grouped behind him, allowing the formal symmetry to read as a modest graduation performance rather than a posed portrait.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: graduation banner in the upper-center of the frame, background; 김광석 in the middle-left of the frame, midground; 송영숙 in the middle-right of the frame, midground.\n- KEY BACKGROUND ELEMENTS: 제5회 민들레 야학 졸업식 현수막 (attached to the blackboard) — The handwritten front face reading '제5회 민들레 야학 졸업식' is fully visible to camera; used as Supplies the event title above the performers and anchors the central axis; 칠판 (banner attached) — Its classroom-facing side is visible behind the banner; used as Forms the background plane for the banner and choir; 단상 (occupied by the choir) — Its front edge faces the camera across the width of the middle frame; used as Elevates and groups the performers across the classroom front.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "5543660e-e35c-42eb-961b-ab1cacde9132" }, "S15sh7": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S16sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S16sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L28B02.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Around a table in the café’s gathering area, with guitars held among the seated friends.\n\nSHOT TEXT (authoritative, Korean): 카페 테이블에 모여 앉아 각자 무릎 위에 기타를 올려둔 김광석(광석·광돌·김광돌), 김창기(창기), 유준호(준호·준열), 박경찬(경찬)의 와이드 샷\n\nCARRIED STATE (persist exactly): Gwangseok, Changgi, Junho, and Gyeongchan each keep a guitar on or across their laps at the table.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At seated eye level several meters from the table, hold a wide frontal composition containing all four men and their guitars, with the table centered low and open café space around the group. Vary their orientations within the shared frame: 김광석 turns toward the off-screen television direction while 김창기, 유준호, and 박경찬 remain absorbed in their instruments and conversation.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 카페 테이블 (surrounded by the group) — The near side faces camera, with the four men distributed around it; used as Binds the four musicians into a single conversational arrangement; 기타들 (being lightly played) — Each guitar rests across its player's knees with the string side angled partly toward camera; used as Makes the informal playing session immediately legible without excessive foreground enlargement.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L28B02.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ] ], "asset_id": "34ef0e29-c111-48a9-b77c-e32ca936e78d" }, "S16sh11": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S16sh15": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S17sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S17sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L30B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: On the church’s performance platform, lined up before the historical slide screen; daylight filters through the church windows.\n\nSHOT TEXT (authoritative, Korean): 성당 단상 위 스크린 앞에 나란히 선 정승찬(승찬), 송영숙(영숙), 김광석(광석·광돌·김광돌), 박혜수(혜수)의 와이드 샷\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From the rear of the central aisle at eye level, hold a symmetrical wide view with the platform and projection screen centered beyond the seated audience. 정승찬, 송영숙, 김광석, and 박혜수 form a measured line across the mid-background, each given clear separation while audience shoulders create a low foreground frame that anticipates the forward move.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: projection screen (showing historical slide photographs) — The audience-facing surface is visible to camera and displays historical documentary photographs; used as Centered behind the performers as the principal visual and historical context; audience (filling the church and seated toward the platform); used as Low foreground layers establish the packed scale of the church and direct the eye toward the platform; platform (occupied by the four performers) — Its audience-facing edge runs horizontally across the centered background; used as Defines the full horizontal line occupied by the four performers.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L30B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "3a26f893-bc9e-44cb-9e17-2af0be910b7e" }, "S17sh8": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S17sh13": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S18sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S18sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L31B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside a moving train’s four-seat facing bay, beside a sunlit window overlooking fields.\n\nSHOT TEXT (authoritative, Korean): 햇살이 드는 기차 동반석에 마주 앉은 김광석(광석·광돌·김광돌), 정승찬(승찬), 송영숙(영숙), 박혜수(혜수)의 와이드 샷\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From the aisle opening at seated eye level, frame both facing benches and all four travelers in a wide composition without crossing their shared axis. 김광석 and 정승찬 occupy one side of the bay while 송영숙 and 박혜수 occupy the other; the window and passing fields remain visible behind the inward-facing arrangement as the camera pauses before its approach.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: train window (rice paddies and fields passing outside) — Its plane is seen obliquely from the aisle, with rice paddies and fields visible beyond; used as Provides the outward view anchoring Youngsook's attention and the train's movement; four-seat bay (occupied by four travelers) — The two seat rows face each other across the center of the frame; used as Defines the opposing arrangement and shared social space of the four travelers; poetry book (open and being read) — The open page faces Hyesu and is seen by camera at an oblique angle; used as Visible near Hyesu's lap as evidence of her private concentration.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L31B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "929f354a-1073-49c9-9bf2-bed59a282c95" }, "S18sh5": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S18sh5.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L31B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: At one seat of the train’s fixed four-person facing bay, diagonally across from the window-side passenger.\n\nSHOT TEXT (authoritative, Korean): 창밖을 보는 송영숙(영숙)을 향해 조심스럽게 곁눈질을 하는 김광석(광석·광돌·김광돌)의 애정 어린 시선\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Close to 김광석 on the aisle side at his seated eye level, angle across the bay so his three-quarter face occupies the left foreground while 송영숙 remains smaller in the right midground beside the window. Use shallow focus biased toward his cautious eyes while preserving enough definition in her window-facing profile to make the affectionate glance readable.\n- FRAMING SCALE: close-up\n- FRAME LAYOUT: Gwangseok in the middle-left of the frame, foreground, looks toward Youngsook across the train bay; Youngsook in the middle-right of the frame, midground.\n- KEY BACKGROUND ELEMENTS: train window (daytime landscape passing outside) — Seen behind Youngsook at an oblique angle, with the passing fields visible beyond; used as Frames Youngsook's outward attention and separates it from Gwangseok's inward attention toward her; opposing train seats (occupied) — The seat edges recede diagonally from Gwangseok toward Youngsook; used as Provides a restrained scale reference between the close foreground face and Youngsook across the bay.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L31B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "0097480a-0c16-41a7-86f9-3dad6ee4d9f4" }, "S19sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S19sh1.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: Outside on the station forecourt at dusk, where the travelers have gathered with their luggage.\n\nSHOT TEXT (authoritative, Korean): 어스름한 저녁 서울역 앞 거리에서 각자 커다란 짐을 양손에 든 채 동그랗게 모여 마주 보고 서 있는 정승찬(승찬), 송영숙(영숙), 김광석(광석·광돌·김광돌), 박혜수(혜수)의 와이드 샷\n\nCARRIED STATE (persist exactly): All four retain the performance luggage they are carrying outside the station.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At eye level across the open area before Seoul Station, hold a wide view of the four figures gathered in a loose circle, with space around them reserved for 정승찬's eventual departure. Their large loads hang from both hands and interrupt the circle at the lower frame, while the station frontage remains broad contextual background rather than a competing focal point.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: open area before Seoul Station (occupied by the four travelers); used as Provides open negative space around the group and a clear path for the coming exit; luggage carried by the group (held in both hands) — The bags and boxes angle inward around the four-person circle; used as Creates irregular lower-frame silhouettes while communicating the group's physical burden; station frontage (visible at dusk) — Its outward-facing side forms the distant background behind the group; used as Locates the gathering without drawing attention away from the group.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "132b57d9-7eba-4508-b10d-a46b8a5e186c" }, "S19sh6": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S20sh6": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S20sh6.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L33B02.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: In the lamplit residential alley directly outside the gate of a large two-story house.\n\nSHOT TEXT (authoritative, Korean): 어두운 골목 끝, 웅장한 2층 양옥집 대문 앞에 우뚝 멈춰 선 송영숙(영숙)과 김광석(광석·광돌·김광돌)의 뒷모습\n\nCARRIED STATE (persist exactly): Gwangseok is still carrying Youngsook's slide box as they reach her house.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From shoulder height behind 송영숙 and 김광석, center their backs in the lower foreground and hold the two-story house entrance ahead as the reveal. Their abruptly checked movement is expressed through stopped feet and slightly lifted shoulders, while the house and gate remain clearly readable above and between them.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: Youngsook in the lower-left of the frame, foreground, looks toward house entrance; Gwangseok in the lower-right of the frame, foreground, looks toward two-story house; two-story house entrance in the upper-center of the frame, background.\n- KEY BACKGROUND ELEMENTS: two-story Western-style house (visible at the end of the alley) — The entrance-facing facade is centered beyond the pair; used as Forms the principal reveal above the two figures and motivates Gwangseok's upward gaze; house gate (positioned before the house) — The entrance side faces the two characters and the camera behind them; used as Marks the precise stopping point and frames the approach to Youngsook's home; slide box (carried by Gwangseok) — Its outer side and top edge are partially visible beside Gwangseok; used as Remains carried at Gwangseok's side as the object connecting this reveal to the coming handoff.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L33B02.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "96ff53c2-73b9-41fd-910b-43c2026718ac" }, "S20sh10": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S20sh12": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S21sh1": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "bgonly", "error": null }, "S22sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S22sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L28B04.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Just inside the café entrance, looking into the dim performance room.\n\nSHOT TEXT (authoritative, Korean): 어두운 카페 문을 밀고 안으로 막 들어선 채 커다란 눈으로 주위를 응시하는 김선경(선경)의 상반신\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Just inside the entrance at chest height, the camera begins in a tight upper-body frame about 1.5 meters ahead of 김선경 and tracks backward along her inward path. She occupies the middle-right of the frame in three-quarter view, caught between pushing the door and taking her first step as the dim cafe opens behind her.\n- FRAMING SCALE: medium shot\n- FRAME LAYOUT: 김선경 in the middle-right of the frame, midground, moves toward cafe interior; cafe interior in the middle-center of the frame, background.\n- KEY BACKGROUND ELEMENTS: 카페 문 (김선경이 밀어 열고 들어오는 중) — Its interior-facing side is visible at an angle behind 김선경; used as Frames her entrance and marks the threshold she has just crossed; 어두운 카페 내부 (어두움); used as Provides the space she scans as the backward track begins.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L28B04.png" ], [ "char:김선경", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/f0ab63e0-2c21-4f58-a19c-12a0c735da38.png" ] ], "asset_id": "228cc92e-afe7-4319-a6e2-07383c84fd16" }, "S22sh8": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S22sh16": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S23sh5": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S23sh5.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: On a wooden bench outside the women’s dormitory entrance, under a solitary streetlamp.\n\nSHOT TEXT (authoritative, Korean): 노란 가로등 불빛이 떨어지는 낡은 나무 벤치 위에 약간의 거리를 두고 나란히 앉아 앞을 응시하는 김창기(창기)와 김선경(선경)\n\nCARRIED STATE (persist exactly): Seongyeong retains her bouquet and the book Changgi returned to her as they sit on the bench.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At seated eye level several meters in front of the bench, the camera holds a centered static wide two-shot with both figures and the small gap between them clearly legible. 김창기 sits on the left and 김선경 on the right, each looking ahead rather than at the lens, while the bench and surrounding exterior establish their shared pause.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: 김창기 in the middle-left of the frame, midground; 김선경 in the middle-right of the frame, midground; bench with a visible gap between the two figures in the lower-center of the frame, midground.\n- KEY BACKGROUND ELEMENTS: 낡은 나무 벤치 (낡은 상태) — The long front side faces the camera, with 김창기 and 김선경 seated apart along it; used as Shared horizontal anchor that makes the initial distance between the couple visible.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ], [ "char:김선경", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/f0ab63e0-2c21-4f58-a19c-12a0c735da38.png" ] ], "asset_id": "216838b7-0120-49fc-a107-5308e416c104" }, "S23sh7": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S24sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S24sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L35B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: On the second-floor terrace overlooking the house’s broad lawn-covered courtyard.\n\nSHOT TEXT (authoritative, Korean): 2층의 하얀 테라스 난간 밖으로 몸을 내민 채 아래를 향해 활짝 웃으며 한 팔을 크게 흔드는 송영숙(영숙)의 상반신\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From the yard at chest height several meters from the house, the upward tilt settles into a low-angle upper-body view of 송영숙 beyond the second-floor railing. She is centered above the railing, leaning outward with one arm crossing the upper portion of the frame in a broad wave toward 김광석 below.\n- FRAMING SCALE: medium shot\n- FRAME LAYOUT: 송영숙 in the upper-center of the frame, midground, looks toward 김광석 in the yard below; second-floor terrace railing in the middle-center of the frame, midground.\n- KEY BACKGROUND ELEMENTS: 2층 하얀 테라스 난간 (송영숙이 난간 밖으로 몸을 내밀고 있음) — Its outer side faces the yard camera, with 송영숙 leaning beyond its upper edge; used as Forms the lower compositional boundary and establishes 송영숙’s elevated position; 고급 주택 외관 (세련된 외관) — The yard-facing facade and second-floor terrace side are visible from below; used as Provides architectural context around the elevated greeting.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L35B01.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "d61d2e99-1cc5-4e0e-8d87-2c2625d4ebc9" }, "S24sh6": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S25sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S25sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L37B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the tidy bedroom at the upright piano, with daylight illuminating the room.\n\nSHOT TEXT (authoritative, Korean): 피아노 의자에 나란히 앉아 건반 위를 짚은 송영숙(영숙)과 그 옆에서 집중하는 김광석(광석·광돌·김광돌)의 모습\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At seated eye level in front of and slightly to one side of the piano bench, the static medium composition includes both players and the keyboard without breaking their shared axis. 송영숙 sits on the left with fingers demonstrating a chord, while 김광석 leans in from the right and concentrates on the same keys; the keyboard occupies less than the lower third.\n- FRAMING SCALE: medium shot\n- FRAME LAYOUT: 송영숙 in the middle-left of the frame, midground, reaches for piano keyboard; 김광석 in the middle-right of the frame, midground, looks toward piano keyboard; piano keyboard in the lower-center of the frame, foreground.\n- KEY BACKGROUND ELEMENTS: 업라이트 피아노와 건반 (송영숙이 코드를 짚는 중) — The keyboard side faces both players and is seen obliquely from the front-side camera position; used as Primary shared instrument and lower compositional anchor for the lesson; 책장 (영숙의 방에 놓여 있음) — Its shelf front faces generally toward the room and remains visible behind the players; used as Room context behind the seated pair; 단정하고 예쁜 침대 (단정하게 놓여 있음) — One long side is visible beyond the piano area; used as Provides orderly domestic context without competing with the lesson.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L37B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "0fdf4f99-0b10-44b4-b7ed-cfc2b4d22d54" }, "S25sh8": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S26sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S26sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_university_district_street_sel.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: On the sidewalk outside a record shop, looking through its display window into the store.\n\nSHOT TEXT (authoritative, Korean): 레코드 가게 안에서 앨범을 들고 있는 김광석(광석·광돌·김광돌)을 쇼윈도 밖에서 응시하는 송영숙(영숙)의 측면\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From outside the shop window at Song Youngsook’s shoulder height, hold a static medium composition with her right-facing profile occupying the left foreground and 김광석 visible beyond the window in the right background. Keep the album legible between them without enlarging it unnaturally, using the window plane to separate her quiet attention from his absorbed excitement inside the shop.\n- FRAMING SCALE: medium shot\n- FRAME LAYOUT: 송영숙 in the middle-left of the frame, foreground, looks toward 김광석 beyond the shop window; 김광석 in the middle-right of the frame, background.\n- KEY BACKGROUND ELEMENTS: record shop window (between 송영숙 and the shop interior) — The camera sees 송영숙 beside the surface and 김광석 holding the album inside the shop beyond it; used as A separating plane that preserves both characters at different depths; 신촌레코드 sign (visible) — The lettered face bearing ‘신촌레코드’ is angled toward the street camera; used as Background location identifier near the upper edge of the composition; LP album (held by 김광석) — The illustrated front cover is turned partly toward the window and remains readable from the camera position; used as Narrative focal detail held within the background figure's upper-body area.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_university_district_street_sel.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "d62bc18c-c76b-4605-89c4-1ac06de710d9" }, "S26sh3": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S26sh7": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S27sh8": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S27sh8.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L39B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the house living room near the kitchen doorway and the wall used for slide projection; a strong projector beam lights the room.\n\nSHOT TEXT (authoritative, Korean): 화들짝 놀란 표정으로 가위를 든 한국인 중년 남성의 팔목을 꽉 움켜잡은 영숙의 어머니의 양손\n\nCARRIED STATE (persist exactly): Youngsook's mother has both hands on the scissors, stopping Youngsook's father from using them.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At head height beside 송영숙’s position, continue the dolly-in into a close composition crossing 영숙의 어머니’s startled face and both hands locked around the raised wrist of 영숙 집의 한국인 중년 남성. Place her face in the upper left and the joined hands at center, while the scissors remain a small but unmistakable threat near the right edge as the camera reaches the end of its approach.\n- FRAMING SCALE: close-up\n- FRAME LAYOUT: 영숙의 어머니 in the middle-left of the frame, midground, reaches for the raised wrist holding the scissors; 영숙 집의 한국인 중년 남성 in the middle-right of the frame, midground.\n- KEY BACKGROUND ELEMENTS: scissors (held in the raised hand) — The scissors extend from the restrained raised hand toward 송영숙’s off-screen position; used as Threat detail held near the frame edge without overwhelming the human intervention; wall carrying the song-play slide image (slide images are projected on it) — The wall face toward the room carries the projected reference photographs; used as A blurred fragment of political and theatrical context behind the confrontation.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L39B01.png" ], [ "char:영숙의 어머니", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/ca920c37-9f55-4e70-abe7-c57dc52b36c5.png" ] ], "asset_id": "59b43025-25f8-41c6-bcca-d0b8ca72d331" }, "S27sh9": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S28sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S28sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_downtown_building_front_sel.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Outside a busy-district building, beside the low edge of a sidewalk planter.\n\nSHOT TEXT (authoritative, Korean): 김광석(광석·광돌·김광돌) 곁에 막걸리병을 두고 초췌한 몰골로 서 있는 김민기(민기)의 전신\n\nCARRIED STATE (persist exactly): Minki has a makgeolli bottle beside him and continues holding the paper cup.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From waist height and slightly beside the seated 김광석, finish the lateral reveal in a wide frame that shows 김민기 from head to foot beside the bottle and cup. Keep 김광석 low in the left foreground on the flowerbed ledge while 김민기 occupies the center-right midground, his worn coat drawn around his gaunt figure.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: 김광석 in the lower-left of the frame, foreground, looks toward 김민기 beside him; 김민기 in the middle-right of the frame, midground, looks toward 김광석 on the flowerbed ledge.\n- KEY BACKGROUND ELEMENTS: flowerbed ledge (김광석 is seated on it) — Its outer edge runs behind and beneath the seated 김광석; used as Supports 김광석’s seated foreground position and anchors the reveal spatially; 막걸리 bottle (set beside 김민기); used as Placed near 김민기’s feet as evidence of his condition; paper cup (held by 김민기); used as Handheld prop that completes 김민기’s disheveled introduction; building entrance (visible) — The street-facing entrance lies behind the two men; used as Background destination that will later receive 김광석.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_downtown_building_front_sel.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:김민기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/68a14492-591f-4f41-b82a-6184b8416a74.png" ] ], "asset_id": "72a6569b-eece-4a0d-a054-e91eed9567b6" }, "S28sh4": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S29sh7": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S29sh7.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L41B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: At the control-room side of a studio window, facing the compact recording booth and its microphones; studio lights are on.\n\nSHOT TEXT (authoritative, Korean): 부스 안의 김광석(광석·광돌·김광돌)을 향해 막걸리가 담긴 종이컵을 허공에 번쩍 들어 올린 김민기(민기)\n\nCARRIED STATE (persist exactly): Minki still has the makgeolli paper cup and raises it toward Gwangseok.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Outside the booth at eye level and medium distance, hold 김민기 in the right foreground as he lifts the 막걸리-filled paper cup high toward 김광석 inside. Preserve 김광석 as a smaller figure in the left background beyond the booth boundary, allowing their shared eyeline to cross the frame without turning either man into a direct-to-camera portrait.\n- FRAMING SCALE: medium shot\n- FRAME LAYOUT: 김민기 in the middle-right of the frame, foreground, looks toward 김광석 inside the recording booth; 김광석 in the middle-left of the frame, background, looks toward 김민기 outside the recording booth; recording booth boundary in the middle-center of the frame, midground.\n- KEY BACKGROUND ELEMENTS: recording booth window (between the two men) — 김광석 is visible inside the booth beyond the surface while 김민기 remains outside; used as Separates the men spatially while allowing their acknowledgment to remain visible; paper cup containing 막걸리 (raised by 김민기); used as Raised signal and character detail kept below the size of 김민기’s head; recording booth (김광석 is inside) — The interior side is visible beyond the booth boundary; used as Defines 김광석’s background workspace.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L41B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:김민기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/68a14492-591f-4f41-b82a-6184b8416a74.png" ] ], "asset_id": "1d25dd3e-b2f6-4636-b63d-362670dc8cca" }, "S29sh15": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S29sh15.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L41B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the compact recording booth at the microphone position, visible through the control-room window.\n\nSHOT TEXT (authoritative, Korean): 한쪽 뺨이 붉게 부풀어 오른 채 밖의 김민기(민기)를 주시하는 김광석(광석·광돌·김광돌)의 얼굴 클로즈업\n\nCARRIED STATE (persist exactly): Gwangseok's repeatedly struck cheek remains red and swollen.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Inside the booth at face height, hold a nearly static close-up along 김광석’s eyeline toward 김민기 outside, angling slightly to make the reddened swollen cheek fully readable. His face occupies the central majority of the frame while the booth context falls soft at the edges, excluding other performers from the narrow emotional beat.\n- FRAMING SCALE: close-up\n- FRAME LAYOUT: 김광석 in the middle-center of the frame, midground, looks toward 김민기 outside the recording booth.\n- KEY BACKGROUND ELEMENTS: microphone (positioned in front of 김광석) — Its pickup end is angled toward 김광석 near the frame edge; used as Soft edge context indicating that 김광석 remains at his recording position; recording booth window (between 김광석 and 김민기) — The surface lies along 김광석’s gaze toward 김민기 outside, though the exterior is kept soft; used as Background division along 김광석’s outward eyeline.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L41B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "cdd49e28-0f1d-4170-a6b5-f8b312a2b8d3" }, "S29sh16": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S29sh16.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L41B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: In the studio control area immediately outside the recording booth, facing the booth window and audio equipment.\n\nSHOT TEXT (authoritative, Korean): 김광석(광석·광돌·김광돌)을 향해 한쪽 손의 엄지손가락을 곧게 치켜세운 김민기(민기)의 상반신\n\nCARRIED STATE (persist exactly): Gwangseok's cheek remains visibly red and swollen as Minki gives him a thumbs-up.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: On the reverse angle outside the booth at chest height and medium distance, hold 김민기’s upper body slightly off center as he taps one cheek and raises the other hand’s thumb toward 김광석 beyond the camera. Keep the thumb clearly separated from his torso but proportionate, while his amused three-quarter face remains the primary focus of the approval gesture.\n- FRAMING SCALE: medium shot\n- KEY BACKGROUND ELEMENTS: chair (positioned outside the booth) — The chair’s front and seat angle are visible behind and below 김민기; used as Places 김민기 in the control-side recording environment without competing with his gesture; recording booth boundary (separates 김민기 from 김광석) — Its outside-facing side lies in the direction of 김민기’s raised thumb toward 김광석; used as Provides restrained spatial context at the edge of the reverse composition.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L41B01.png" ], [ "char:김민기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/68a14492-591f-4f41-b82a-6184b8416a74.png" ] ], "asset_id": "0efef932-2e72-4bff-898f-fa2f843e734f" }, "S30sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S30sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L33B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Outside at the residential house gate, where the mother has opened the entrance.\n\nSHOT TEXT (authoritative, Korean): 영숙의 어머니를 향해 환하게 미소 짓는 박혜수(혜수)의 얼굴\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From eye level just inside the open entrance, the camera completes its restrained dolly-in toward 박혜수, facing outward along the established conversational axis. Her smiling face occupies the center-right of a close frame in three-quarter view, with the entrance edge retained along the left border as spatial context and 영숙의 어머니 remaining off-screen.\n- FRAMING SCALE: close-up\n- KEY BACKGROUND ELEMENTS: 영숙 집 대문 (open) — The inner side and thickness of the open gate edge face the camera; used as The near edge frames one side of 박혜수's close-up and preserves the inside-to-outside camera position.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L33B01.png" ], [ "char:박혜수", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/3b6dd6ee-4325-4523-9860-348e2cec3bd5.png" ] ], "asset_id": "18373efb-e434-4fb8-96d2-7f18890cb633" }, "S30sh3": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S31sh5": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S31sh5.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: On an old bench in a quiet corner of the outdoor playground.\n\nSHOT TEXT (authoritative, Korean): 낡은 벤치에 나란히 앉아 허공을 응시하는 송영숙(영숙)과 김광석(광석·광돌·김광돌)의 와이드 샷\n\nCARRIED STATE (persist exactly): Youngsook's unevenly cropped short hair is concealed beneath the hat while she sits with Gwangseok.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At seated eye level several meters in front of the bench, hold a static, centered wide composition that shows 송영숙 and 김광석 side by side without turning the image into a symmetrical portrait. The worn bench stretches across the lower middle of the frame; 송영숙 sits on the left and 김광석 on the right, each directing their gaze into separate areas of the playground rather than toward one another.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: 송영숙 in the middle-left of the frame, midground; 김광석 in the middle-right of the frame, midground; shared playground bench in the lower-center of the frame, midground.\n- KEY BACKGROUND ELEMENTS: 낡은 벤치 (worn and occupied by both characters) — The long front side faces the camera, with both figures seated along the same plane; used as It establishes the shared physical space while preserving emotional separation between the two seated figures; 놀이터 공간 (daytime exterior); used as Visible around the bench as broad environmental context without competing with the pair.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "f53e9abd-df87-487c-9426-1b189474c150" }, "S31sh8": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S31sh11": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S32sh8": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S32sh13": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S33sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S33sh1.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: At a secluded corner table inside the coffeehouse, beside an open newspaper.\n\nSHOT TEXT (authoritative, Korean): 다방 구석 테이블에 앉아 고개를 숙인 채 악보 위를 펜으로 짚고 있는 정승찬(승찬)의 상반신\n\nCARRIED STATE (persist exactly): The newspaper remains spread open beside Seungchan's composition work on the café table.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At the end of the approach, frame 정승찬 from slightly above table level on a three-quarter axis, with his bowed upper body occupying the center-right of a medium shot. His pen points into the open score near the lower center, while the unfolded newspaper remains beside his elbow at the left edge to prepare the camera's subsequent descent toward it.\n- FRAMING SCALE: medium shot\n- FRAME LAYOUT: 정승찬 in the middle-right of the frame, midground, looks toward open musical score; unfolded newspaper beside his elbow in the lower-left of the frame, foreground.\n- KEY BACKGROUND ELEMENTS: 구석 테이블 (occupied by score and unfolded newspaper) — Its top plane is visible at a shallow angle beneath 정승찬's upper body; used as The working surface supporting the score, pen action, and adjacent newspaper while defining the next downward camera move; 악보 (open on the table) — The written music side lies open toward 정승찬 and is partially readable to the camera; used as Its open pages receive the pen point and provide the immediate focal area beneath 정승찬's lowered gaze; 신문 (unfolded beside 정승찬) — The article side faces upward, showing the headline ‘대학생 전방입소 반대 시위 중 서울대 김세진군 운명, 이재호군 위독’; used as Placed beside his elbow as the visual destination of the next camera transition.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:정승찬", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/ece62b7c-1dd1-4871-8e5c-0ef3c926acc5.png" ] ], "asset_id": "a4fe367f-3ea0-450f-b5c1-237dd2440943" }, "S34sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S34sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L46B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: On the narrow exterior stairwell descending from the street to a basement office entrance.\n\nSHOT TEXT (authoritative, Korean): 지하로 통하는 좁은 계단 위쪽에서 무거운 책상을 쥔 채 힘겹게 버티고 선 김광석(광석·광돌·김광돌)의 전신\n\nCARRIED STATE (persist exactly): Gwangseok and the others remain supporting the same heavy desk as they lower it down the basement stairs.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From waist height on the upper stair landing, hold a static wide view directly down the narrow stair axis as 김광석 braces the desk at the head of the descent. His full body remains visible behind and below the desk's diagonal edge, occupying the middle of the frame, while the desk stays under roughly one third of the image and the descending stairs provide an unambiguous path into the basement.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: 김광석 supporting the desk in the middle-center of the frame, midground, moves toward basement stair descent; narrow stairs descending to basement in the lower-center of the frame, background.\n- KEY BACKGROUND ELEMENTS: 지하 계단 (narrow and leading underground) — The steps face the camera at the landing and recede downward through the center of the frame; used as The descending central axis organizes 김광석's strained posture and the desk's intended movement; 무거운 책상 (being supported at the top of the stairs) — One long side and part of the top face angle across the stair axis toward the camera; used as The carried load crosses the midground diagonally while leaving 김광석's full-body effort readable; 허름한 4층 건물 (fronted by the basement stair entrance) — The entrance-facing side frames the stair access; used as Visible beyond the stair area as exterior context for the move.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L46B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "78cf4834-d85b-466e-9b25-68cee0701325" }, "S34sh4": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S34sh6": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S35sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S35sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L48B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the spacious basement office, at a newly moved-in desk among sparsely placed furnishings; daylight reaches the basement room.\n\nSHOT TEXT (authoritative, Korean): 책상 위에 깔린 신문지 위에서 짜장면이 담긴 그릇을 앞에 두고 나란히 앉은 김광석(광석·광돌·김광돌), 송영숙(영숙), 정승찬(승찬)의 와이드 샷\n\nCARRIED STATE (persist exactly): The moved office furniture remains in the new basement space, and newspaper and jajangmyeon bowls remain spread across the desk.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At seated eye level and roughly four meters from the table, frame 김광석, 송영숙, and 정승찬 as a balanced wide group around the newspaper-covered desk, with their food in the lower middle and the sparsely placed office furnishings extending behind them. Begin from a settled, centered composition before the forward dolly gathers momentum toward one end of the table.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 책상 (covered with newspaper and used for the meal) — Its long side faces the camera, with all three seated figures arranged beyond and around it; used as The shared surface anchors the three-person composition and their informal meal; 신문지 (spread across the desk) — The printed side faces upward and is visible around the food bowls; used as Its printed face forms the visible layer beneath the bowls; 짜장면 그릇 (set in front of the seated group); used as The bowls form a modest foreground-center cluster linking the three figures; 피아노 (placed in the office) — Its keyboard side is angled away from the centered camera; used as It occupies one part of the otherwise sparse office background and establishes the music workspace; 책장 (placed in the office) — Its open front faces partly toward the camera; used as It adds vertical structure behind the group without competing for focus; 오디오 기기 (placed in the office) — Its control-facing side is partially visible from the wide frontal position; used as It identifies the room as a working music space.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L48B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "f64a5302-0ae1-4655-943c-77c39b9c92bd" }, "S35sh5": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S35sh6": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S36sh5": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S36sh5.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L28B03.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: At a drinking table inside the café, opposite the standing woman.\n\nSHOT TEXT (authoritative, Korean): 충격을 받은 듯 두 눈을 크게 뜬 채 굳어 있는 김창기(창기)의 클로즈업\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At 김창기's seated eye level, complete the dolly-in at the front of his three-quarter angle, letting his face occupy most of the frame while his body remains rigid against the lower edge. Hold his widened eyes toward 김선경 just outside the composition, with a shallow background separating his delayed shock from the café space.\n- FRAMING SCALE: close-up\n- KEY BACKGROUND ELEMENTS: 테이블 (positioned in front of 김창기) — Its near edge crosses below 김창기's seated torso; used as A thin lower-frame edge keeps 김창기 physically grounded in the confrontation.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L28B03.png" ], [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ] ], "asset_id": "3c0aba58-5d25-47e7-b022-57518d83047c" }, "S36sh7": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S37sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S37sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L48B02.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: In the basement office’s rehearsal area, beside the piano and egg-carton-covered acoustic wall; the room lights are on at night.\n\nSHOT TEXT (authoritative, Korean): 계란판이 붙은 벽면을 배경으로 피아노를 치고 있는 송영숙(영숙)과 그 옆에서 눈을 감고 열창하는 김광석(광석·광돌·김광돌)의 와이드 샷\n\nCARRIED STATE (persist exactly): The egg-carton soundproofing remains fixed across the office walls while Youngsook plays and Gwangseok sings.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At standing eye height and wide distance, frame 송영숙 seated at the piano on the left and 김광석 singing beside her on the right, both held against the egg-carton-covered wall. Keep the composition static and spacious, showing 김광석's closed-eyed vocal effort and 송영숙's active hands before the later horizontal pan leaves them.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 피아노 (being played by 송영숙) — The keyboard side faces 송영숙 and is visible obliquely to the camera; used as The piano anchors 송영숙 on one side and establishes the live rehearsal arrangement; 계란판이 붙은 벽면 (covered with egg cartons as sound absorption) — The attached egg-carton faces are visible behind both performers; used as It forms the principal background texture and identifies the improvised acoustic treatment; 사무실 가구 (placed in their proper positions) — Several furniture faces are visible around the edges of the performance area; used as The placed furnishings give the rehearsal area an established office context.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L48B02.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "902e21d3-5890-462a-817e-d586b53e0934" }, "S37sh4": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S37sh6": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S38sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S38sh1.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: Outside at the dark summit of the hillside neighborhood, above its steep stairways.\n\nSHOT TEXT (authoritative, Korean): 캄캄한 산동네 꼭대기에서 양손을 입가에 모으고 허공을 향해 소리치는 김광석(광석·광돌·김광돌)의 전신\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From a static standing eye-level position several meters in front of the hilltop, frame 김광석 full-body in a wide three-quarter view, his figure centered against the open space beyond the neighborhood. He is caught at the peak of the shout with both hands cupped around his mouth, elbows spread and weight pitched onto his front foot.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: hilltop neighborhood (seen at night); used as Provides the distant spatial context around and below the isolated figure; open space beyond the hilltop (dark); used as Forms the uncluttered background toward which the shout is projected.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "088e7126-8afe-4086-88dc-3c5c13745ae9" }, "S38sh3": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S38sh3.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: Inside the neighborhood electronics shop at the repair workbench, under the shop’s night lighting.\n\nSHOT TEXT (authoritative, Korean): 전파사 작업대 앞에서 입에 모나미 볼펜을 가로로 꽉 문 채 눈에 핏대를 세운 김광석(광석·광돌·김광돌)의 얼굴 클로즈업\n\nCARRIED STATE (persist exactly): Gwangseok keeps the ballpoint pen clamped horizontally in his mouth for the vocal exercise.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At 김광석's eye level across the work surface, complete the dolly-in and hold on a close three-quarter view of his strained face, with the horizontal pen crossing the lower center of the frame. His widened eyes and tightened jaw occupy most of the image while a narrow strip of the workbench remains below as spatial evidence.\n- FRAMING SCALE: close-up\n- FRAME LAYOUT: Kim Gwang-seok in the middle-center of the frame, foreground.\n- KEY BACKGROUND ELEMENTS: repair-shop workbench (in use) — Its near edge faces the camera beneath 김광석; used as Appears as a narrow lower-frame plane that anchors the repair-shop position without competing with the face; ballpoint pen (held tightly between his teeth) — Its long side is presented horizontally to the camera; used as Runs horizontally across 김광석's mouth as the central vocal-training prop.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "d4115197-9c97-46af-a6ce-c3f58837b8cf" }, "S38sh6": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S38sh6.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: Inside the small dark bedroom, on the bedding against the room wall.\n\nSHOT TEXT (authoritative, Korean): 어두운 방안 이불 위에서 상체를 벌떡 일으킨 채 허공을 향해 입을 크게 벌리고 있는 김광석(광석·광돌·김광돌)의 상반신\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Place the static camera beside the bedding at upper-torso height and frame 김광석 from the waist up in a medium three-quarter angle. Capture the instant his torso snaps upright into the center-right of frame, mouth open toward the unoccupied room while the disturbed bedding remains across the lower portion.\n- FRAMING SCALE: medium shot\n- KEY BACKGROUND ELEMENTS: bedding (in use for sleep); used as Anchors the abrupt rise and remains visible under and around his torso; dark room interior (dark); used as Provides negative space in front of his projected voice.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "5380f3c2-adfe-4e76-a040-3a1ccbe8590d" }, "S39sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S39sh2.png", "plate_path": null, "prompt": "Draw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: Inside the family living room beside the table holding the academic notice; daylight lights the room.\n\nSHOT TEXT (authoritative, Korean): 종이를 든 채 싸늘하고 경멸하는 눈빛으로 앞을 쏘아보는 한국인 중년 남성의 얼굴\n\nCARRIED STATE (persist exactly): Changgi's academic-failure notice is still in his father's hand after being taken from the table.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At chest height about a step in front of the father, aim slightly upward and hold a tight close view as his face turns three-quarters toward 김창기 off frame. Keep the notification low at the frame edge so the contemptuous eyes and compressed mouth carry the judgment without becoming a centered frontal portrait.\n- FRAMING SCALE: close-up\n- KEY BACKGROUND ELEMENTS: grade-retention notification (held in his hand) — The printed notification side is angled toward the man, with only an oblique portion visible to the camera; used as Remains partly visible at the lower edge as the evidence prompting the judgment.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [], "asset_id": "39c93cbb-1abd-478c-a38f-fa45ef3cb725" }, "S39sh4": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S40sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S40sh2.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: Under a tree at a distance from the university library exit and its forecourt.\n\nSHOT TEXT (authoritative, Korean): 먼발치 나무 밑에 가방을 멘 채 우두커니 서서 굳은 표정으로 도서관 쪽을 응시하는 김창기(창기)의 전신\n\nCARRIED STATE (persist exactly): Changgi keeps his bag slung on him as he watches from a distance.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From the library side at eye level, hold a distant wide frame of 김창기 full-body beneath the tree, leaving open space along the direction of his fixed eyeline. He occupies a modest portion of the frame with the bag on his shoulder, his weight stalled between steps rather than settled into a posed stance.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: tree (김창기 is positioned beneath it) — The side of the trunk facing the library-side camera is visible beside 김창기; used as Locates the distant figure and provides a vertical frame beside his body; bag (carried over his shoulder) — Its outer side is visible beside his torso; used as Hangs from his shoulder and clarifies that he has paused while on his way.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ] ], "asset_id": "bdb89837-ef00-4f82-9428-b29d6aa178a4" }, "S41sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S41sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L54B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the narrow vestibule between moving train cars, seated on the fixed steps by the exterior window.\n\nSHOT TEXT (authoritative, Korean): 기차 객차 사이의 좁은 출입구 계단에 쭈그려 앉아 창밖을 응시하는 김창기(창기)의 전신\n\nCARRIED STATE (persist exactly): Changgi has the soju and his notebook with him on the train vestibule steps.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From waist height on the opposite side of the narrow passage, frame 김창기 full-body on the entrance steps with the window and moving sky and fields visible beyond him. The static straight-on axis is offset by his profile as he sits low with knees drawn in, making the confined passage and passing exterior share the wide frame.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: Kim Chang-gi in the middle-center of the frame, midground, looks toward train window; train window and passing landscape in the upper-right of the frame, background.\n- KEY BACKGROUND ELEMENTS: train entrance steps (김창기 is seated on them) — The steps face the camera along the passage axis; used as Supports 김창기's compressed seated posture and emphasizes the narrow inter-car space; train window (the train is moving) — The camera sees the cloudy sky and fields passing beyond it; used as Places the moving exterior beyond 김창기 and carries his outward eyeline; cloudy sky and fields (passing outside the train); used as Provides continuous lateral movement behind the still figure.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L54B01.png" ], [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ] ], "asset_id": "bf8514ef-8d68-44ec-a404-f32805d6d05c" }, "S41sh3": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S42sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S42sh1.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: At the exposed outermost edge of a seaside cliff above breaking waves.\n\nSHOT TEXT (authoritative, Korean): 파도가 부서지는 아찔한 바닷가 절벽의 맨 끝자락에 위태롭게 서 있는 김창기(창기)의 전신\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From the inland side at standing eye level, hold a long oblique wide shot that places 김창기's full body close to one side of frame and the extreme cliff edge clearly beneath him. His body is angled away from the camera, feet held at the brink and head lowered toward the drop, while the sea and breaking waves remain visible beyond and below.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: Kim Chang-gi in the middle-right of the frame, midground, looks toward drop below the cliff edge; extreme cliff edge in the lower-right of the frame, midground; sea below in the lower-left of the frame, background.\n- KEY BACKGROUND ELEMENTS: cliff edge (김창기 is positioned at the extreme edge) — Its brink runs obliquely across the frame, with the inland side nearer the camera and the drop beyond 김창기; used as Defines the dangerous boundary directly beneath 김창기's feet; sea below (waves are breaking below); used as Establishes the height and downward direction of his gaze.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ] ], "asset_id": "979bc637-c91e-4100-9030-839408ae98ba" }, "S43sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S43sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L56B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside a narrow seaside lodging room, at a low desk lit only by a small lamp; the sea is visible through the window.\n\nSHOT TEXT (authoritative, Korean): 어두운 민박집 방 안, 작은 스탠드 불빛이 비치는 앉은뱅이책상 앞에 웅크리고 앉은 김창기(창기)의 뒷모습\n\nCARRIED STATE (persist exactly): The photograph of Changgi and Seongyeong, dozens of sleeping pills, and empty soju bottles remain arranged on the low desk.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Begin a slow dolly from seated eye level near the room entrance, holding a wide rear view of 김창기 hunched at the low desk with the window centered beyond him. His back occupies the lower middle of the frame while the photograph and clustered pills and bottles remain contained on the desk, and the narrow room closes around both sides.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: low desk (in use) — Its working side faces 김창기, while the camera sees it from behind him; used as Anchors 김창기's folded posture and contains the photograph, pills, bottles, and writing materials; small desk lamp (switched on) — Its illuminating side is directed toward the desk surface; used as Provides the localized illumination described in the room; photograph of 김창기 and 김선경 (placed on the desk) — The image side faces upward, showing 김창기 and 김선경 smiling together; used as Carries the emotional focal image beneath 김창기's lowered gaze; sleeping pills and empty soju bottles (dozens of pills and multiple empty bottles are on the desk); used as Forms a troubling cluster around the writing area without obscuring 김창기; window (the sea is visible outside at night) — The camera looks past 김창기 and through the window toward the sea; used as Provides the exterior endpoint beyond 김창기 and the desk.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L56B01.png" ], [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ] ], "asset_id": "b53a78f2-19b5-41e9-94ed-51f3c235c510" }, "S43sh4": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S43sh10": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S44sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S44sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L57B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the lodging owner’s room, on the floor around the collapsed woman and scattered pesticide bottles; the room is lit at night.\n\nSHOT TEXT (authoritative, Korean): 방바닥에 쓰러진 민박집 주인아주머니와 주변에 널브러진 농약병들을 내려다보는 김창기(창기)의 시선\n\nIMMOBILE CHARACTER POSE — CANONICAL (identical wherever this character appears in ANY panel; on any conflict THIS POSE WINS): She is collapsed on the room floor, with her torso and head down at floor level and her arms and legs slack where they fell.\n\nCARRIED STATE (persist exactly): The unconscious landlady remains collapsed on the floor with pesticide bottles scattered around her.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Align the lens precisely with 김창기's eye height and tilt sharply down in subjective point-of-view, placing the collapsed 민박집 주인아주머니 across the lower-middle frame with pesticide bottles dispersed around her. Use the room floor as negative space between the bottles so the dangerous layout reads immediately without showing any part of 김창기.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 방바닥 (민박집 주인아주머니 is collapsed on it); used as Establishes the woman's fallen position and the distribution of the bottles around her; 농약병들 (Scattered across the floor) — Several bottles lie at differing angles around the woman's body; used as Scattered evidence surrounding the fallen woman, kept individually small within the wide frame.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L57B01.png" ], [ "char:민박집 주인아주머니", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/296eedee-7dbd-4414-8f44-15b1f26b895b.png" ] ], "asset_id": "b6da04ce-c4bc-4b31-aff0-06b01e51c20c" }, "S44sh5": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S45sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S45sh1.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: On a dark coastal village road, running away from the lodging with the patient.\n\nSHOT TEXT (authoritative, Korean): 어두운 해변 도로에서 등에 민박집 주인아주머니를 업은 채 한쪽 발이 허공에 뜬 전력 질주의 mid-action 자세인 김창기(창기)의 전신\n\nIMMOBILE CHARACTER POSE — CANONICAL (identical wherever this character appears in ANY panel; on any conflict THIS POSE WINS): She is carried piggyback on Kim Chang-gi with her torso against his back, her head and arms hanging slack around his shoulders, and her legs hanging behind him.\n\nCARRIED STATE (persist exactly): The unconscious landlady is carried piggyback on Kim Chang-gi with her torso against his back, her head and arms hanging slack around his shoulders, and her legs hanging behind him as he runs, unable to support herself.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Track backward from a low waist-height position several strides ahead of 김창기, keeping his entire body centered as he sprints directly toward the lens with one foot suspended at the peak of the stride. 민박집 주인아주머니 hangs across his back as a compact second silhouette, while sufficient road remains beneath and around them to make the forward momentum unmistakable.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: 김창기 carrying 민박집 주인아주머니 in the middle-center of the frame, midground, moves toward retreating camera along the road.\n- KEY BACKGROUND ELEMENTS: 해변 도로 (Dark at night); used as Provides the receding travel axis beneath the centered runners.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ], [ "char:민박집 주인아주머니", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/296eedee-7dbd-4414-8f44-15b1f26b895b.png" ] ], "asset_id": "2e8d3d30-0816-4ec7-8b12-29ae75591f87" }, "S45sh5": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S46sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S46sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L59B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: At a treatment bed inside the rural hospital emergency room, beside the connected vital-sign equipment; clinical lights are on.\n\nSHOT TEXT (authoritative, Korean): 시골 병원 응급실 침대 위에 호흡기를 달고 고요히 누워있는 민박집 주인아주머니의 상반신\n\nIMMOBILE CHARACTER POSE — CANONICAL (identical wherever this character appears in ANY panel; on any conflict THIS POSE WINS): She is positioned flat on her back on the emergency-room bed, with her head and torso supported by the mattress, her arms and legs resting motionless, and a breathing apparatus attached to her.\n\nCARRIED STATE (persist exactly): The incapacitated landlady is positioned flat on her back on the emergency-room bed, with her head and torso supported by the mattress, her arms and legs resting motionless, and a breathing apparatus attached to her.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Dolly in beside the upper half of the emergency bed at patient chest height, settling on 민박집 주인아주머니's still upper body in a calm three-quarter view. Let her occupy most of the central frame while the attached breathing apparatus and a narrow edge of the normal vital-sign display remain visible without overtaking her face.\n- FRAMING SCALE: medium shot\n- KEY BACKGROUND ELEMENTS: 응급실 침대 (민박집 주인아주머니 is lying on it after emergency treatment) — The bed is seen obliquely from beside its upper half; used as Supports the patient and establishes the emergency-room setting around her upper body; 호흡기 (Attached to the patient) — The connected patient-facing components are visible beside and on the woman; used as Appears beside the patient as evidence of assisted breathing; 바이털 사인 기기 화면 (Connected and displaying normal vital signs) — The display face is partly turned toward the camera and shows normal vital signs; used as A peripheral status cue that prepares the following lateral move toward the equipment.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L59B01.png" ], [ "char:민박집 주인아주머니", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/296eedee-7dbd-4414-8f44-15b1f26b895b.png" ] ], "asset_id": "922ec200-39e9-449c-8be9-8f5df179f2f2" }, "S46sh7": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S47sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S47sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L60B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: In the rural hospital’s front yard at night, directly outside the entrance.\n\nSHOT TEXT (authoritative, Korean): 병원 마당 한가운데서 기타를 멘 채 서 있는 김광석(광석·광돌·김광돌)의 전신\n\nCARRIED STATE (persist exactly): Gwangseok has his guitar slung on him in the hospital yard.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Track laterally through the hospital yard at adult waist height until 김광석's full body is cleanly revealed near frame center, with the guitar strapped to him and his weight resting loosely on one leg. Keep him in a relaxed three-quarter angle toward 김창기 off-screen, leaving modest yard space around his figure to make the unexpected vigil readable.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 병원 마당 (김광석 is positioned in its middle); used as Provides open space around 김광석 for the lateral reveal; 병원 입구 (Visible at the edge of the yard) — The entrance is seen behind and to one side of 김광석; used as Anchors the reveal to the place from which 김창기 has emerged; 기타 (Strapped to 김광석) — Its body and neck are seen obliquely while strapped against 김광석; used as Defines 김광석's musician identity in silhouette and prepares the next scene.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L60B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "aeaa91fb-2558-4d9d-a636-48012d54934b" }, "S48sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S48sh1.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: On the dark sandy beach near the edge of the gentle surf.\n\nSHOT TEXT (authoritative, Korean): 잔잔한 파도가 들이치는 어두운 동해 해변 모래사장에 나란히 앉아, 새우깡과 소주병을 앞에 두고 입을 연 채 서로를 바라보는 김광석(광석·광돌·김광돌)과 김창기(창기)의 와이드 샷\n\nCARRIED STATE (persist exactly): The soju bottle and shrimp crackers remain on the sand in front of Gwangseok and Changgi; Gwangseok's guitar is with them.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Begin a slow dolly-in from seated eye level in a wide two-shot, placing 김광석 on the left and 김창기 on the right as they sit side by side but turn toward one another in conversation. Keep the snack and soju bottle small in the lower center foreground and the gently advancing shoreline behind both men, preserving ample beach around their intimate exchange.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 잔잔한 파도와 해안선 (Gentle waves are entering the shore); used as Creates a gently moving background line behind the seated pair; 모래사장 (The shoreline waves are wetting part of it); used as Supplies the broad ground plane and negative space surrounding the two men; 소주병 (Placed in front of them) — The bottle is seen from an oblique side angle between the seated figures; used as Small shared drinking object positioned between the men in the lower frame; 새우깡 (Placed in front of them); used as Small shared snack placed beside the bottle as part of their informal conversation.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ] ], "asset_id": "3c84ca1f-cb4f-499a-9c82-616f48fc3755" }, "S48sh5": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S48sh9": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S49sh3": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S49sh3.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L07B02.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: On the rain-soaked street immediately outside the neighborhood electronics shop.\n\nSHOT TEXT (authoritative, Korean): 한 손으로 우산을 받쳐 들고 다른 한 팔로 카세트 레코드를 가슴에 바짝 껴안은 채 빗속을 걷는 mid-stride 자세의 김광석(광석·광돌·김광돌)의 전신, 한 발이 지면에서 막 떨어진 찰나\n\nCARRIED STATE (persist exactly): Gwangseok carries the GoldStar cassette recorder tightly under one arm and holds the open umbrella over it.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Track parallel to 김광석 from waist-to-chest height in a full-body lateral view, timing the frame as his trailing foot has just left the ground. Keep him in the middle third moving screen-left to screen-right, with the open umbrella above him and the cassette recorder compressed between his free forearm and chest.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: 김광석 in the middle-center of the frame, midground, moves toward screen-right walking path.\n- KEY BACKGROUND ELEMENTS: open umbrella (Open and carried in one hand) — Its canopy is seen obliquely from the tracking camera's lateral position; used as Held over 김광석 and clearly outlines his protected walking posture; cassette recorder (Carried tightly under one arm) — One broad side is partially visible while its body is pressed against 김광석's chest; used as Held against the torso as the protected object motivating his posture; falling rain (Rain is falling); used as Visible throughout the lateral tracking frame and articulates the night exterior.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L07B02.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "9cd705bd-89f7-4120-827b-df4a946b8475" }, "S50sh9": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S50sh9.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L48B04.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the basement office’s seating area, beside the sofa and music table; the office lights are on at night.\n\nSHOT TEXT (authoritative, Korean): 눈을 감은 채 서로의 입술을 깊게 포개어 맞춘 송영숙(영숙)과 정승찬(승찬)의 측면\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Complete the lateral dolly-in at seated head height, settling into a close two-profile composition with 송영숙 on one side and 정승찬 on the other. Their closed eyes and joined lips occupy the central plane, while their shoulders remain at the lower edge to preserve the fact that both are seated rather than turning the image into an isolated facial detail.\n- FRAMING SCALE: close-up\n- KEY BACKGROUND ELEMENTS: office interior (Occupied at night); used as Soft, minimal spatial context behind the close profile pairing.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L48B04.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ], [ "char:정승찬", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/ece62b7c-1dd1-4871-8e5c-0ef3c926acc5.png" ] ], "asset_id": "fefb386e-5023-4b56-8e6d-c26507a3fbe9" }, "S50sh10": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S50sh12": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S51sh2": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "bgonly", "error": null }, "S52sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S52sh2.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: On a city street in heavy nighttime rain, away from the basement office.\n\nSHOT TEXT (authoritative, Korean): 쏟아지는 비를 온몸으로 맞으며 하늘을 향해 입을 크게 벌리고 처절하게 오열하는 김광석(광석·광돌·김광돌)의 일그러진 얼굴 클로즈업\n\nCARRIED STATE (persist exactly): Gwangseok remains soaked by the heavy rain after leaving the cassette recorder behind.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Finish the dolly-in directly before 김광석 at upper-chest height, angling slightly upward into a tight frontal close-up as he throws his face toward the sky. His distorted face occupies most of the frame, mouth open in a wail, while a narrow margin of falling rain remains around his head to keep the grief physically rooted in the street.\n- FRAMING SCALE: close-up\n- FRAME LAYOUT: 김광석 in the middle-center of the frame, foreground.\n- KEY BACKGROUND ELEMENTS: heavy rain (Falling more intensely); used as Remains visible around the face and across its surface, reinforcing the unprotected breakdown; Noryangjin street (Visible at night); used as A minimal contextual strip behind the close-up.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "c3b1a30c-4b10-45ec-a5b6-ea932ac7c82e" }, "S53sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S53sh1.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: At a secluded corner table inside a shabby drinking house, under its night lighting.\n\nSHOT TEXT (authoritative, Korean): 허름한 선술집 구석 테이블에 홀로 앉아 기울인 병 입구에서 투명한 소주가 빈 잔으로 쏟아져 내리는 순간의 김광석(광석·광돌·김광돌)의 손\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At seated waist height beside the corner table, hold a static oblique insert on 김광석's hand tipping the bottle over the empty cup. Keep the pouring hand and stream central while part of his seated torso and the table edge remain at the frame boundary as scale references, anticipating the upward tilt toward his face.\n- FRAMING SCALE: insert close-up on a detail\n- KEY BACKGROUND ELEMENTS: soju bottle (Tilted while pouring transparent soju) — The neck and opening angle downward toward the cup, with the bottle body seen obliquely; used as Primary action object held above the cup; soju cup (Empty at the start of the captured pour); used as Receives the visible stream at the center-lower part of the composition; corner table (김광석 is seated beside it); used as Provides a restrained spatial boundary and scale reference around the hand action.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "c57809f9-9353-4be2-b1a0-7852a9d2d915" }, "S53sh2": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S54sh11": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S54sh11.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L49B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: On the sofa inside the flooded basement office, with rising water reaching face level and rain leaking through the windows.\n\nSHOT TEXT (authoritative, Korean): 물이 턱 끝까지 찰랑이는 소파 위에서 차가운 촉감에 번쩍 눈을 뜬 김광석(광석·광돌·김광돌)의 얼굴\n\nCARRIED STATE (persist exactly): The basement office is flooded up to Gwangseok's face, with loose scores and objects floating around the sofa; the cassette recorder remains in the office.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From beside the sofa at waterline height, the dolly-in ends in an extreme close-up as 김광석's face occupies most of the frame, with the waterline cutting across his chin and the sofa edge barely visible behind him. Hold at eye level for the instant his eyes snap open, using shallow depth of field to isolate the abrupt return to consciousness.\n- FRAMING SCALE: close-up\n- FRAME LAYOUT: 김광석 in the middle-center of the frame, foreground.\n- KEY BACKGROUND ELEMENTS: 소파 (Partially submerged) — Only the upper edge beside 김광석's head faces the camera; used as Forms a narrow support line behind his head without distracting from his awakening; 사무실을 채운 물 (Reaching 김광석's chin and rippling); used as Places the waterline directly against his chin as the immediate danger cue.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L49B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "9bd94aaa-1a09-4cdd-8b48-a9ebac3a8341" }, "S54sh14": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S54sh18": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S55sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S55sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L47B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Outside the flooded basement office, beside its broken low window as water is pumped out in the rain.\n\nSHOT TEXT (authoritative, Korean): 깨진 창문 틈으로 빗물이 흘러나오는 반지하를 바라보며 입을 벌린 채 황망하게 굳어있는 송영숙(영숙)의 측면\n\nCARRIED STATE (persist exactly): The basement window remains broken and floodwater is still being drained through it.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Hold statically at 송영숙's shoulder height from her side, placing her open-mouthed profile in the middle-left and the broken basement window beyond her sightline in the middle-right. A medium frame preserves both her arrested expression and the water coming through the damaged opening without turning the window into a separate insert.\n- FRAMING SCALE: medium shot\n- FRAME LAYOUT: 송영숙 in the middle-left of the frame, midground, looks toward broken basement window; broken basement window in the middle-right of the frame, background.\n- KEY BACKGROUND ELEMENTS: 깨진 반지하 창문 (Broken, with water coming through the opening) — The damaged opening faces the camera obliquely, revealing the flooded office area beyond; used as Remains visible beyond 송영숙's profile as the cause of her shock; 반지하 사무실 외부 (Water is being removed through the window) — Its exterior face and low window opening are angled toward the camera; used as Locates the damaged window below street level behind her sightline.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L47B01.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "ff5b973b-027e-4d12-9263-479fa173829f" }, "S55sh2": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S55sh7": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S56sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S56sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L65B01.png", "prompt": "The attached image is the LOCATION PHOTOGRAPH (loose spatial reference only).\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Outside the memorial hall entrance at night, along the street where the audience queue extends.\n\nSHOT TEXT (authoritative, Korean): 기념관 건물 앞 어두운 거리에서 끝없이 길게 줄을 늘어선 수많은 한국인 관객들의 와이드 샷\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From an elevated position above and farther back from the street, finish the crane move in a high-angle wide composition facing the memorial building. The queue runs across the lower frame and recedes toward the entrance, revealing its exceptional length while keeping the building façade as the spatial endpoint.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 기독교 100주년기념관 건물 (Standing at the end of the queue) — The front-facing building exterior is visible beyond the line; used as Acts as the destination toward which the long queue converges; 건물 앞 거리 (Filled by a long line of people); used as Provides the long ground plane occupied by the spectators.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L65B01.png" ] ], "asset_id": "dbf733d0-b1a4-43c0-95aa-a5cf023f4ddd" }, "S57sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S57sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L65B02.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: At center stage inside the memorial hall, isolated under a brightening spotlight.\n\nSHOT TEXT (authoritative, Korean): 어두운 무대 위 핀 조명을 받으며 홀로 서서 지그시 눈을 감은 김광석(광석·광돌·김광돌)의 전신\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Begin the forward dolly from audience-center eye height in a square frontal wide composition, with 김광석's full body isolated at center stage and generous darkness around him. His closed eyes and planted posture hold the frame still while the camera's gradual approach establishes the coming performance.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: 김광석 in the middle-center of the frame, background.\n- KEY BACKGROUND ELEMENTS: 무대 (Dark around 김광석); used as Creates open negative space around the solitary performer; 객석 (Positioned before the stage); used as Defines the camera's distant audience-center axis toward the stage.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L65B02.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "76beef75-aa7b-443f-9f36-3e9d94685e0b" }, "S57sh8": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S57sh15": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S58sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S58sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L23B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the cluttered small bedroom, on bedding amid discarded bottles and snack bags; afternoon sunlight enters through the curtain gap.\n\nSHOT TEXT (authoritative, Korean): 소주병과 과자봉지가 나뒹구는 지저분한 방안 이불 위에서 부스스한 머리로 상체를 일으켜 앉은 김광석(광석·광돌·김광돌)의 모습\n\nCARRIED STATE (persist exactly): Empty soju bottles and snack bags remain scattered around Gwangseok's room, and his guitar lies toppled in one corner.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From seated eye level near the room entrance, hold a static wide view across the bedding, with 김광석 occupying the middle-right of the frame as he pushes his torso upright. Scattered bottles and snack bags break up the lower foreground, preserving the untidy room around his isolated figure while his gaze begins to drift toward the opposite corner.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 이불 (Spread inside the room with 김광석 sitting on it); used as Extends across the lower midground beneath 김광석 and establishes where he has awakened; 소주병들 (Scattered around the room) — Several bottles lie at differing side angles across the room floor; used as Scattered through the lower foreground as evidence of the room's disorder; 과자봉지들 (Scattered around the room); used as Creates irregular foreground shapes among the bottles without obscuring the subject.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L23B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "e3e663b0-a2ee-4e0a-b7e5-7061e183de43" }, "S58sh5": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S59sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S59sh2.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: On a rain-soaked city street beneath a gray daytime sky, among umbrella-carrying pedestrians.\n\nSHOT TEXT (authoritative, Korean): 우산 하나 없이 쏟아지는 비를 온몸으로 맞으며, 젖은 담배를 입에 문 채 나사 빠진 듯 다리를 끌며 흐느적거리는 mid-stride 자세의 김광석(광석·광돌·김광돌)의 전신\n\nCARRIED STATE (persist exactly): Gwangseok remains soaked and keeps the wet cigarette in his mouth while walking without an umbrella.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Track beside 김광석 at waist height in a wide, perpendicular profile view, keeping his full body near the middle of the frame as one dragging step trails behind the other. His wet cigarette remains readable at his mouth while umbrella-carrying pedestrians pass in the deeper frame, providing stable scale and contrasting movement.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: 김광석 in the middle-center of the frame, midground, moves toward his walking direction at frame right.\n- KEY BACKGROUND ELEMENTS: 젖은 담배 (Wet and held between his lips) — Its length projects sideways from 김광석's mouth in profile; used as Held clearly at 김광석's mouth as a small but narratively important detail within the full-body view; 행인들의 우산 (Carried by people walking in the rain) — Carried above the pedestrians and seen from varied side and rear angles; used as Populate the background with steadier pedestrian movement around 김광석's loose gait; 회색 하늘 (Gray during the daytime rain); used as Forms the upper background and establishes the drained exterior atmosphere; 쏟아지는 비 (Falling heavily); used as Crosses the street space and 김광석's body, making his lack of an umbrella immediately legible.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "345924ef-45b7-4802-9617-ad7c98c7a84e" }, "S60sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S60sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L28B02.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Just inside the café doorway, facing the friends gathered around a table.\n\nSHOT TEXT (authoritative, Korean): 한 손으로 카페 문을 반쯤 민 채 우두커니 서서 친구들을 바라보는 김광석의 모습\n\nCARRIED STATE (persist exactly): Several liquor bottles remain on the café table where his friends are gathered.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From inside the café near the gathering area, hold a slightly widened medium view toward the entrance, placing 김광석 within the half-open doorway while keeping the hand pressing the door fully visible. The doorway supplies a firm vertical frame around him, and a low strip of the friends' table and bottles indicates the unseen group toward which he looks.\n- FRAMING SCALE: medium shot\n- FRAME LAYOUT: 김광석 in the middle-center of the frame, midground; half-open café door in the upper-center of the frame, background.\n- KEY BACKGROUND ELEMENTS: 카페 문 (Half-open and held by 김광석) — The half-open door presents an angled inner face toward the camera; used as Frames 김광석's entrance and keeps his door-holding hand readable; 탁자 (Several liquor bottles are placed on it); used as Appears as a narrow foreground strip linking the entrance to the gathering area; 술병들 (Several bottles rest on the table) — Seen in partial side views along the table edge; used as Provides foreground evidence of the gathering without blocking 김광석.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L28B02.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "60e7dd35-088b-41be-8d8c-ef6a0c4a70e0" }, "S60sh8": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S60sh11": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S61sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S61sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L67B02.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the lakeside villa living room, beside the musicians and a window overlooking the dark lake; interior lights are on.\n\nSHOT TEXT (authoritative, Korean): 창밖으로 어두운 호수가 보이는 거실에서 기타를 안은 채 입을 벌리고 열창하는 표정인 김창기(창기)의 상반신\n\nCARRIED STATE (persist exactly): The cassette recorder remains positioned in front of Gwangseok while Changgi performs with his guitar.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At eye level inside the living room, hold a close medium view of 김창기 near the window, with his singing upper body and guitar centered slightly left and the dark lake visible behind his shoulder. Keep enough of the instrument in frame to read the performance while the window creates a calm rectangular depth layer rather than overpowering him.\n- FRAMING SCALE: medium shot\n- FRAME LAYOUT: 김창기 in the middle-left of the frame, midground; window with dark lake beyond in the upper-center of the frame, background.\n- KEY BACKGROUND ELEMENTS: 거실 창문 (The lake is visible through it) — The camera looks through the window toward the dark nighttime lake; used as Creates the background depth directly behind the performance; 호수 풍경 (Dark and visible at night); used as Forms the distant exterior background behind 김창기; 기타 (Held and played by 김창기) — Its front and neck are seen obliquely across 김창기's torso; used as Occupies the lower central portion of the frame as 김창기's performance instrument.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L67B02.png" ], [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ] ], "asset_id": "ccb7ba64-f9a7-44a3-a90a-ee223025e9e2" }, "S61sh3": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S61sh6": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S62sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S62sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L68B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the villa’s narrow bathroom, facing the wall under the bathroom light.\n\nSHOT TEXT (authoritative, Korean): 좁은 화장실 안에서 벽을 향해 선 채 약간 풀린 눈으로 입술을 살짝 벌린 김광석(광석·광돌·김광돌)의 상반신\n\nCARRIED STATE (persist exactly): Gwangseok remains mildly intoxicated after drinking in the living room.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From chest height beside the wall, continue a gentle dolly-in on a medium three-quarter frontal view of 김광석's upper body as he faces the wall. Place him off-center with slack shoulders and slightly parted lips, leaving a narrow band of wall in his eyeline so the confined bathroom geometry remains legible without revealing his lower body.\n- FRAMING SCALE: medium shot\n- KEY BACKGROUND ELEMENTS: 화장실 벽 (Directly in front of 김광석); used as Defines the direction 김광석 faces and compresses the shallow space beside him; 좁은 화장실 내부 (Confined interior space); used as The tight boundaries restrict negative space around his upper body.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L68B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "f05fad60-d06c-40f1-92ae-7c3ae3a5a328" }, "S63sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S63sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L67B02.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Across the villa living-room floor, where cool early-morning light enters from the lake-facing windows.\n\nSHOT TEXT (authoritative, Korean): 푸르스름한 새벽빛이 스며드는 거실 바닥에 널브러져 잠든 김광석(광석·광돌·김광돌)과 김창기(창기)의 와이드 샷\n\nCARRIED STATE (persist exactly): The friends remain sprawled asleep around the living room among their instruments and recording equipment.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From one edge of the living room, hold a static wide frontal view at the flow’s elevated floor-level vantage, with 김광석 resting in the left midground and 김창기 offset in the right midground. Their irregular sleeping poses and the open floor path between them establish the route for the later low approach, while the frame keeps both bodies fully visible without favoring either man.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 거실 바닥 (광석과 창기가 여기저기 널브러져 잠들어 있음); used as Provides the broad spatial field around the two sleeping figures and preserves an open approach path through the composition.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L67B02.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ] ], "asset_id": "8e93bf60-81a8-4345-a478-013b4a6ac8f9" }, "S63sh4": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S63sh12": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S64sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S64sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L69B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: On a sofa in the bright record-company president’s office, facing the executives listening to a demo tape.\n\nSHOT TEXT (authoritative, Korean): 밝은 사장실 소파에 나란히 앉아 무릎 위에 두 손을 모은 채 바짝 굳어있는 김광석(광석·광돌·김광돌)과 김창기(창기)의 와이드 샷\n\nCARRIED STATE (persist exactly): The demo cassette containing Gwangseok's recording of “On the Street” is playing in the office.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At seated eye level several meters in front of the sofa, hold a static centered wide composition with 김광석 on the left and 김창기 on the right. Both sit compressed into the seating with hands gathered over their knees, while balanced negative space around them makes their rigidity conspicuous before the lateral track begins.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 사장실 소파 (두 사람이 나란히 앉아 있음) — The front of the sofa faces the camera, with 김광석 and 김창기 seated side by side; used as Shared seating anchor that keeps the two anxious men aligned in one plane; 사장실 내부 (밝게 보임); used as Establishes the office around the seated pair without distracting from their posture.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L69B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ] ], "asset_id": "07c1405c-fcd7-4f11-83aa-b0cb95b69c23" }, "S64sh4": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S64sh6": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S65sh6": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S65sh6.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L06B03.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the electronics repair shop at the fan-repair workbench, with the radio playing nearby.\n\nSHOT TEXT (authoritative, Korean): 전파사 안에서 라디오를 들으며 드라이버를 꽉 쥔 채 허공을 보며 대견하게 미소 짓는 광석의 아버지의 상반신\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Inside the repair shop at upper-torso height, briefly stabilize the lateral track on 광석의 아버지 in a medium three-quarter composition, placing him left of center and leaving the article wall available at frame right for the continuing move. He pauses over the fan with the screwdriver clenched in one hand, his distant gaze and restrained smile carrying the pride.\n- FRAMING SCALE: medium shot\n- KEY BACKGROUND ELEMENTS: 수리 중인 선풍기 (수리 중) — Its repair side is turned toward 광석의 아버지, with only part of the appliance visible below his torso; used as Foreground work context beneath the father’s paused hands; 라디오 (광석의 노래가 나오고 있음) — Its control and speaker face angle into the shop rather than squarely toward the camera; used as Visible source of the song he is listening to within the shop context; 액자에 스크랩된 신문 기사들 (전파사 벽에 걸려 있음) — The printed front faces are mounted toward the room, with their headlines partially visible from the camera’s oblique angle; used as Anticipates the next lateral reveal at the right side of the composition.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L06B03.png" ], [ "char:광석의 아버지", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/b6968f65-4381-4394-9823-7a248465e015.png" ] ], "asset_id": "e15ab8e2-d73e-4810-a4e9-1806329dc1ab" }, "S65sh7": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "bgonly", "error": null }, "S65sh10": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S65sh10.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L06B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: In an indoor promotional photo setup, before the camera and alongside the costumed band members.\n\nSHOT TEXT (authoritative, Korean): 조랑말 인형 탈을 머리에 쓴 채 카메라를 향해 장난스럽게 두 팔을 치켜든 김광석(광석·광돌·김광돌)의 전신\n\nCARRIED STATE (persist exactly): Gwangseok remains dressed in the pony promotional costume for the photo shoot.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From the centered promotional-photo position at chest height, hold a static full-body frame with 김광석 alone in the middle and clear space around his raised arms and feet. Time the image as both arms reach upward beneath the pony headpiece, turning the direct address into a playful publicity pose rather than a formal portrait.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: 김광석 in the middle-center of the frame, midground.\n- KEY BACKGROUND ELEMENTS: 앨범 홍보 사진 촬영 공간 (김광석이 혼자 포즈를 취하고 있음); used as Defines the dedicated area in which the full-body publicity pose is performed; 조랑말 인형 탈 (김광석이 머리에 쓰고 있음) — The pony face points toward the camera as 김광석 poses beneath it; used as Comic costume element framing 김광석’s head while leaving his entire body visible.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L06B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "3c0d74f5-b717-404d-8c1b-15def771cf44" }, "S66sh4": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S66sh4.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: Inside the compact radio studio at the broadcast microphone, surrounded by fixed microphones, headphones, seats, and console equipment.\n\nSHOT TEXT (authoritative, Korean): 마이크 쪽으로 몸을 숙인 채 뻔뻔하고 능청스러운 표정으로 입을 열고 있는 김광석(광석·광돌·김광돌)의 상반신\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: On the reverse axis at 김광석’s seated eye level, complete the dolly-out from close distance until his upper body and the microphone are established together. Place him slightly right of center in three-quarter profile as he leans toward the microphone, leaving the desk line to carry the next pan toward the group.\n- FRAMING SCALE: medium shot\n- FRAME LAYOUT: 김광석 in the middle-right of the frame, midground, reaches for studio microphone.\n- KEY BACKGROUND ELEMENTS: 마이크 (김광석이 몸을 숙여 답변 중) — The microphone head angles toward 김광석’s mouth, with its side visible to the camera; used as Immediate speaking target positioned between 김광석 and the interviewer’s side of the reverse axis; 스튜디오 책상 (마이크가 놓인 방송 진행 공간) — Its near edge crosses the lower frame and recedes toward the off-screen group; used as Creates the horizontal compositional line that will guide the subsequent pan.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "5555636f-013f-420b-a687-75a1d1cfbf8f" }, "S66sh6": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S66sh6.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: At a fixed seat inside the compact radio studio, beside the microphone and broadcast console.\n\nSHOT TEXT (authoritative, Korean): 헤드폰을 쓴 채 기타를 무릎에 올리고 여유로운 미소를 띠고 있는 김광석(광석·광돌·김광돌)의 얼굴 클로즈업\n\nCARRIED STATE (persist exactly): Gwangseok has the guitar on his lap and wears headphones in preparation for the radio performance.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At Kim Gwangseok’s seated eye level, the completed dolly-in holds a close three-quarter view along his established eyeline. His headphone-framed face occupies most of the center-right, while the upper edge of the guitar remains at the lower frame boundary to preserve his relaxed seated posture without distracting from his smile.\n- FRAMING SCALE: close-up\n- FRAME LAYOUT: 김광석 in the middle-center of the frame, foreground.\n- KEY BACKGROUND ELEMENTS: headphones (worn by Kim Gwangseok) — The earcups are seen from a slight three-quarter angle while being worn; used as The earcups and headband closely frame his face; guitar (resting across his knees) — The body lies across his knees, angled slightly away from the camera; used as Its upper edge at the bottom of frame anchors his seated performance posture.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "a9a354c1-70f2-4923-82f2-9c12050e354b" }, "S67sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S67sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L74B02.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: In the theater’s dark backstage holding area immediately beside the stage entrance.\n\nSHOT TEXT (authoritative, Korean): 무대 뒤 어두운 공간에서 입술을 굳게 다문 채 긴장한 표정으로 서 있는 김광석(광석·광돌·김광돌)의 상반신\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From the prescribed chest-height position in front of Kim Gwangseok, hold a static medium frame with him in the center foreground and the waiting members receding in a compressed line behind him. Keep the framing steady on his sealed lips and braced shoulders, allowing the backstage depth rather than camera movement to carry the anticipation of the coming exit.\n- FRAMING SCALE: medium shot\n- FRAME LAYOUT: 김광석 in the middle-center of the frame, foreground; waiting members in the middle-right of the frame, background.\n- KEY BACKGROUND ELEMENTS: backstage waiting area (occupied by the waiting group); used as The confined background layers the waiting members closely behind Kim Gwangseok.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L74B02.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "fc92357f-70b6-4a34-b13d-ccb41640730a" }, "S67sh6": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S67sh6.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L74B01.png", "prompt": "The attached image is the LOCATION PHOTOGRAPH (loose spatial reference only).\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: In the packed auditorium seating facing the theater stage, under concert lighting.\n\nSHOT TEXT (authoritative, Korean): 무대 위로 등장하는 멤버들을 향해 글씨가 적힌 플래카드를 머리 위로 번쩍 든 채 입을 크게 벌려 환호하는 객석의 한국인 여학생들의 와이드 샷\n\nCARRIED STATE (persist exactly): The audience members retain their handwritten placards raised above their heads.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At eye level near the front of the stage, hold a static wide view across the front rows and the densely occupied seating area. The cheering students spread from edge to edge in layered depth, their raised placards forming an irregular band across the upper half while open mouths and lifted arms animate the lower half.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: written placards (held overhead) — The written front faces are turned toward the stage and therefore remain legible to the stage-side camera; used as The written faces create a rhythmic overhead layer above the cheering crowd; packed theater seating (filled with audience members); used as The tightly occupied rows establish the scale and intensity of the audience response.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L74B01.png" ] ], "asset_id": "548701cc-92d3-4981-af68-874df33d0c30" }, "S67sh10": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S68sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S68sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L75B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: At the applicant’s chair inside a bright corporate interview room, facing the interview panel.\n\nSHOT TEXT (authoritative, Korean): 밝은 면접장 의자에 허리를 꼿꼿이 세우고 두 손을 무릎 위에 모은 채 앉아있는 유준호(준호·준열)의 전신\n\nCARRIED STATE (persist exactly): Junho remains dressed in his suit for the interview.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From the interviewers’ side at seated eye level, hold a centered wide frame that shows 유준호’s full seated figure without crossing the conversation axis. He occupies the middle of the frame with his back held straight and both hands gathered on his knees, while the chair and surrounding interview-room space preserve the formal distance around him.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: interview chair (occupied by 유준호) — Its front is angled slightly toward the interviewers’ side of the room; used as The chair outlines his disciplined full-body posture and formal isolation; interview room interior (bright); used as Open space around the chair emphasizes the formal interview arrangement.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L75B01.png" ], [ "char:유준호", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/8c802930-f882-437a-93df-f3e345d96b6d.png" ] ], "asset_id": "f1368734-8ab1-4fd8-b7e0-ae1353bf29d1" }, "S68sh2": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S68sh6": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S69sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S69sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L76B01.png", "prompt": "The attached image is the LOCATION PHOTOGRAPH (loose spatial reference only).\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside a narrow television-stage waiting passage lined with stacked production equipment.\n\nSHOT TEXT (authoritative, Korean): 김광석(광석·광돌·김광돌)을 향해 굳은 표정으로 삿대질을 한 채 입을 크게 벌리고 있는 한국인 무대감독의 상반신\n\nCARRIED STATE (persist exactly): Gwangseok keeps his acoustic guitar slung over his shoulder in the backstage passage.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From chest height just behind and beside 김광석, frame the 한국인 무대감독 in a tight medium view while retaining 김광석’s shoulder as a foreground reference. The director occupies center-right with his pointing arm extending diagonally toward the left foreground, his open mouth and rigid expression carrying the confrontation through the cramped equipment-lined passage.\n- FRAMING SCALE: medium shot\n- FRAME LAYOUT: 김광석 in the middle-left of the frame, foreground; 한국인 무대감독 in the middle-right of the frame, midground, points to 김광석.\n- KEY BACKGROUND ELEMENTS: broadcast equipment (stacked) — Mixed sides and edges face the camera along both sides of the passage; used as The stacked forms compress the frame edges and reinforce the restricted working passage; backstage waiting passage (narrow and occupied); used as Provides the narrow depth line behind the confrontation.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L76B01.png" ] ], "asset_id": "8d4186e1-a17c-4978-90ba-99a0f6bc14da" }, "S69sh4": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S69sh11": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S70sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S70sh2.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: Along the busy hospital emergency-room aisle leading toward a treatment bed, under bright clinical lighting.\n\nSHOT TEXT (authoritative, Korean): 응급실 복도 저편에서 다급한 표정으로 한쪽 발이 바닥에서 막 떨어진 전력 질주의 mid-stride 자세인 김창기(창기)의 전신\n\nCARRIED STATE (persist exactly): Changgi remains in his doctor's coat as he runs through the emergency room.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From chest height at the bed-corridor junction, hold a long full-body view down the corridor as 김창기 runs toward camera, his figure centered deep in the frame with one foot just leaving the floor. Begin the backward track only as he closes in, keeping the patient's bed at the near frame edge as his destination and preserving the corridor's strong approach axis.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: 김창기 in the upper-center of the frame, background, moves toward patient's bed; patient's bed in the lower-right of the frame, foreground.\n- KEY BACKGROUND ELEMENTS: emergency-room corridor (open for his running approach) — Its length recedes from the camera position toward the far end behind 김창기; used as Creates the deep approach axis behind 김창기; patient's bed (occupied by the fracture patient) — Its corridor-facing side enters the lower-right edge near the camera; used as Marks the near destination at the edge of the composition.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ] ], "asset_id": "6725a605-1380-460f-9800-e0f5dfa51741" }, "S70sh5": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S71sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S71sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L78B02.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Around a soju-covered table inside the late-night soup restaurant.\n\nSHOT TEXT (authoritative, Korean): 테이블에 소주병이 놓인 해장국집에서 각자 굳은 표정으로 앉아 침묵을 지키는 김광석(광석·광돌·김광돌), 김창기(창기), 유준호(준호·준열), 박경찬(경찬)의 와이드 샷\n\nCARRIED STATE (persist exactly): The soju bottles, glasses, and dishes remain on the table between the four friends.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At seated eye height and several meters away, compose a static wide frame centered on all four men and the complete table, using the table axis to separate two subdued pairs without privileging a speaker. 김광석 and 김창기 occupy opposing thirds while 유준호 and 박경찬 close the remaining sides, leaving their rigid silence and the soju bottle plainly readable before the push-in begins.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: restaurant table (upright and in use) — Its full width and seating sides are visible from the centered camera position; used as Organizes the four-way composition and establishes the axis for the coming dolly; soju bottle (placed on the table) — Seen upright near the middle of the table between the four men; used as Provides the central visual evidence of their silent drinking; haejangguk restaurant interior (occupied at night); used as Supplies the wider nighttime restaurant context without distracting from the group.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L78B02.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ] ], "asset_id": "1f1abee2-143c-42d7-8241-db602e30c151" }, "S71sh21": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S71sh32": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "bgonly", "error": null }, "S72sh3": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S72sh3.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: In the crowded concourse of the subway station, amid passing commuters and a newspaper stand.\n\nSHOT TEXT (authoritative, Korean): 김창기(창기)를 올려다보며 깜짝 놀라 두 눈을 크게 뜬 김선경(선경)의 얼굴\n\nCARRIED STATE (persist exactly): Changgi still has his shoulder bag and the newspaper he just bought while facing Seongyeong.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At close facial distance near 김선경's eye height, angle slightly upward along her natural sightline toward the taller 김창기, using his softly focused shoulder at one edge as an over-the-shoulder anchor. Her widened eyes and startled expression occupy most of the frame while the station crowd remains indistinct behind her, ready for the backward dolly that will open the conversation.\n- FRAMING SCALE: close-up\n- FRAME LAYOUT: 김선경 in the middle-center of the frame, foreground, looks toward 김창기 at the frame edge; 김창기 in the middle-right of the frame, foreground, looks toward 김선경.\n- KEY BACKGROUND ELEMENTS: subway-station crowd (busy with passing commuters); used as Supplies soft moving depth behind 김선경 without competing with her recognition; subway-station interior (occupied during daytime) — Seen obliquely behind 김선경 along the crowd's path; used as Locates the encounter within the station while remaining subordinate in shallow focus.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김선경", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/f0ab63e0-2c21-4f58-a19c-12a0c735da38.png" ] ], "asset_id": "9fca4e15-7b94-40b2-a6a6-d80aa40db275" }, "S72sh7": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S73sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S73sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L48B03.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: On the floor of the dark basement music office, beside stacks of cassette-filled cardboard boxes; office lighting is low at night.\n\nSHOT TEXT (authoritative, Korean): 어두운 사무실 바닥에 쪼그려 앉아 카세트테이프가 담긴 종이 상자 위로 다른 상자를 쥔 채 강하게 내려놓아 두 상자가 막 맞닿은 순간의 송영숙(영숙)의 측면\n\nCARRIED STATE (persist exactly): The boxes of activist-song tapes remain stacked as Youngsook lowers another box onto the pile.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Place the camera 0.45 meters above the floor and 0.9 meters from 송영숙, perpendicular to her side and level with the boxes, then hold static. In a tightened medium profile, keep her hands and the exact contact line between the descending paper box and the cassette-filled box centered, while her crouched face remains visible in the upper side of frame.\n- FRAMING SCALE: medium shot\n- KEY BACKGROUND ELEMENTS: paper box containing cassette tapes (resting on the office floor) — Its open upper side and camera-facing side are visible beneath 송영숙's hands; used as Provides the lower support and reveals the cassette-tape contents beside the point of contact; upper paper box (just making contact) — Its lower face meets the upper edge of the box below in strict side view; used as Captures the decisive downward action without overwhelming 송영숙's profile; office floor (supporting the boxes); used as Forms the low visual plane beneath the stacked boxes.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L48B03.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "6b39fad7-d0e4-4b0d-b256-2e8838efffde" }, "S74sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S74sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L06B06.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: At a small table in one corner of the electronics repair shop, beside the shelves and work area under night lighting.\n\nSHOT TEXT (authoritative, Korean): 전파사 한구석 탁자에 마주 앉아 소주잔을 입술로 기울인 채 정지한 김광석(광석·광돌·김광돌)과 광석의 아버지의 와이드 샷\n\nCARRIED STATE (persist exactly): The soju bottle and glasses remain on the small table between Gwangseok and his father.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At seated eye level from an open corner of the radio shop, begin a slow dolly toward a wide view of 김광석 and 광석의 아버지 facing one another across the table. Keep both full seated figures on opposing thirds with their raised soju glasses paused at their lips, allowing the table and quiet room around them to carry the familial distance before the camera arcs closer.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: table in the radio shop (set for their private drinking) — Its width runs between father and son and faces the camera obliquely; used as Establishes the face-to-face axis and supports the drinking gesture; soju glasses (raised to their lips); used as Keeps the reciprocal drinking action readable without becoming oversized; radio-shop interior (occupied at night) — Viewed from an open corner toward the table area; used as Frames the intimate meeting within the business space.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L06B06.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:광석의 아버지", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/b6968f65-4381-4394-9823-7a248465e015.png" ] ], "asset_id": "f90b1ed0-6456-43e6-a85b-18f42562166e" }, "S74sh7": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S74sh10": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S75sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S75sh1.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: At a roadside bus stop on a quiet nighttime street.\n\nSHOT TEXT (authoritative, Korean): 밤거리 버스 정류장에서 굳은 표정으로 나란히 선 김광석(광석·광돌·김광돌)과 송영숙(영숙)의 전신\n\nCARRIED STATE (persist exactly): Gwangseok is already holding the paper shopping bag containing Youngsook's damaged Freewheelin' album.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From eye level several meters ahead, hold a static wide two-shot with Gwangseok on the left and Youngsook on the right, both visible head to foot with an uncomfortable strip of empty space between them. Their torsos turn only slightly toward one another, while the paper shopping bag hangs near the lower center as the future destination of the camera move.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 버스 정류장 (광석과 영숙이 버스를 기다리고 있음) — Its waiting area is presented frontally enough to hold both figures within the same spatial plane; used as Establishes the location around and behind the silent pair; 종이 쇼핑백 (광석이 들고 있음) — The bag hangs upright with its opening above and its broad side angled toward camera; used as Placed low between the figures as a small compositional bridge and the target of the coming push-in.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "17789954-825e-405f-9a46-333d29523787" }, "S75sh5": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S76sh4": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S76sh4.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: Inside the nighttime city bus, at a fixed passenger seat opposite another rider.\n\nSHOT TEXT (authoritative, Korean): 물에 불어 쭈글쭈글해진 밥 딜런 앨범 커버를 쥔 송영숙(영숙)의 손 클로즈업\n\nCARRIED STATE (persist exactly): Youngsook holds the water-swollen, dried Freewheelin' album taken from the shopping bag.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Complete the lap-height push beside Youngsook’s seat with a downward insert of both hands holding the warped album cover above the open shopping bag. The cover occupies less than two fifths of the frame, with her lap, forearms, and the bag providing clear scale around its wrinkled edges.\n- FRAMING SCALE: insert close-up on a detail\n- KEY BACKGROUND ELEMENTS: 프리휠링 앨범 커버 (물에 퉁퉁 불었다가 말라 쭈글쭈글해짐) — The illustrated front face identifying the Freewheelin' album is visible to camera, distorted by the warped cover; used as Primary detailed subject held between Youngsook's hands; 종이 쇼핑백 (열려 있음) — Its opening faces upward below Youngsook's hands; used as Open beneath the album cover to establish where it was hidden; 버스 좌석 (영숙이 앉아 있음) — The seat surface runs beneath and behind Youngsook's body; used as Provides the immediate interior spatial support around Youngsook's lap.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "fa0a6bba-85c2-447d-b96c-5cc24da6623e" }, "S76sh7": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S76sh7.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: From a window seat inside the departing bus, looking back toward the illuminated bus stop.\n\nSHOT TEXT (authoritative, Korean): 창밖 정류장에 우두커니 서서 버스 쪽을 응시하는 김광석(광석·광돌·김광돌)의 모습\n\nCARRIED STATE (persist exactly): Youngsook retains the damaged album inside the bus while Gwangseok remains outside watching her.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: The pan from Youngsook’s side settles inside the bus at window height, holding Gwangseok centered beyond the window in a medium environmental frame. The window edges form a restrained foreground border while his still body remains fully readable at the stop, his face directed back toward Youngsook’s position inside the bus.\n- FRAMING SCALE: medium shot\n- FRAME LAYOUT: bus window edges in the middle-center of the frame, foreground; 김광석 in the middle-center of the frame, background, looks toward Youngsook inside the bus.\n- KEY BACKGROUND ELEMENTS: 버스 창문 (버스 안과 정류장 사이에 놓여 있음) — The camera looks outward through the window toward Gwangseok at the bus stop; used as A transparent foreground layer that separates the moving interior viewpoint from Gwangseok outside; 버스 정류장 (광석이 남아 있음) — The waiting area faces the bus and supports Gwangseok's isolated figure; used as Keeps Gwangseok spatially fixed in the exterior location after the bus departs.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "803a317f-1bd8-44ce-8c14-84ce29b475ba" }, "S77sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S77sh1.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: On the grass immediately beside a military gravestone in the cemetery.\n\nSHOT TEXT (authoritative, Korean): 비석 옆 잔디밭에 앉아 기타를 안고 노트를 굽어보는 김광석(광석·광돌·김광돌)의 전신\n\nCARRIED STATE (persist exactly): Gwangseok has his guitar and lyric notebook with him beside his brother's gravestone.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At seated eye level on the grass, hold a static wide three-quarter view with Gwangseok seated in the lower center-left and the adjacent gravestone rising beside him. His full body, guitar, and open notebook form a compact working triangle, while his bowed head directs attention from his face to the written lyrics.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 형의 묘비 (광석의 바로 옆에 있음) — Its principal face is angled partly toward camera beside Gwangseok; used as Shares the frame with Gwangseok as the emotional spatial anchor of his solitary work; 기타 (광석이 안고 있음) — The body faces partly toward camera while the neck extends across the frame; used as Held across Gwangseok's body while he tests the lyrics musically; 노트 (가사와 기타 코드가 적혀 있음) — The open page facing upward shows written lyrics and added guitar chords; used as Open within Gwangseok's downward sightline as the immediate work surface; 잔디밭 (광석이 앉아 있음); used as Supports the seated figure and establishes the cemetery exterior.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "d377e7a4-457f-45bf-943f-f64fa54cf1a9" }, "S77sh5": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S78sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S78sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L07B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: On the street directly outside the neighborhood electronics shop, beside the newly uncovered compact car.\n\nSHOT TEXT (authoritative, Korean): 모습을 드러낸 낡은 프라이드 승용차를 보며 깜짝 놀라 입을 떡 벌린 김광석(광석·광돌·김광돌)의 전신\n\nCARRIED STATE (persist exactly): The plastic cover has been removed, leaving the old Pride car fully exposed.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At eye level several meters from the front three-quarter side of the car, hold a static wide frame with Gwangseok recoiling on the left and the revealed Pride occupying no more than the right third. His full body is caught at the instant his mouth opens in surprise, with his weight thrown back from the car rather than posed frontally.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: 김광석 in the middle-left of the frame, midground, looks toward revealed Pride car; revealed Pride car in the middle-right of the frame, midground.\n- KEY BACKGROUND ELEMENTS: 낡아 보이는 프라이드 승용차 (비닐 커버가 벗겨져 모습을 드러냄) — Its front and passenger side are both visible in a front three-quarter presentation; used as Shares the reveal with Gwangseok while remaining secondary to his full-body reaction; 비닐 커버 (차에서 벗겨짐) — Its collapsed form lies away from the car after being pulled off; used as May remain low at the edge of the reveal as evidence of the uncovering action; 광돌전파사 앞 (낮의 외부 공간) — The frontage sits behind the revealed car and Gwangseok; used as Establishes the exterior directly in front of the repair shop.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L07B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "adff1850-e6e6-4169-82fa-9d2576b65246" }, "S78sh5": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S78sh7": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S79sh1": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "bgonly", "error": null }, "S80sh3": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S80sh3.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L84B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: At the open doorway of the rural clinic’s examination room, facing the patient and doctor; daylight enters the room.\n\nSHOT TEXT (authoritative, Korean): 진료실 열린 문간에 삐딱하게 기대어 서서 씩 웃고 있는 김광석(광석·광돌·김광돌)의 전신\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From chest height inside the examination room, the pan settles on the open doorway at full-body distance, placing 김광석 slightly off-center within the door opening. He leans asymmetrically into the frame with one shoulder supported by the jamb, smiling toward 김창기 deeper inside the room.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: 김광석 in the middle-right of the frame, midground, looks toward 김창기 inside the examination room; open examination-room doorway in the middle-center of the frame, background.\n- KEY BACKGROUND ELEMENTS: 진료실 문과 문틀 (Open) — The doorway is seen from inside the examination room, with its opening facing the camera obliquely; used as Frames 김광석's entrance and separates him from the examination-room interior.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L84B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "380fb7eb-ea19-46c6-a434-d068ebf3d54d" }, "S80sh4": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S81sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S81sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L85B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: On a raised wooden platform in the clinic residence courtyard, around a nighttime meal and drinking spread.\n\nSHOT TEXT (authoritative, Korean): 닭백숙과 술병이 어지럽게 놓인 평상 위에 러닝 차림으로 마주 앉은 김광석(광석·광돌·김광돌)과 김창기(창기)의 와이드 샷\n\nCARRIED STATE (persist exactly): The chicken soup, sweet potatoes, corn, kimchi, and accumulated liquor bottles remain spread across the platform table.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At seated eye height from the yard's outer edge, the static wide composition places 김광석 and 김창기 on opposite sides of the platform with the crowded drinking table between them. Their relaxed, alcohol-heavy postures form an open diagonal across the frame, while the full platform and the spread of food and bottles establish the social space before the camera advances.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 평상 (Occupied by the two men and the drinking spread) — Its broad top and front edge are visible from the yard side; used as Defines the seating area and supports the two-person composition; 닭백숙, 고구마, 옥수수, 포기김치가 놓인 술상 (Food is spread across the table); used as Forms the central divider and narrative evidence of their long, drunken conversation; 여러 개의 술병 (Accumulated on the table) — Mixed bottle sides face the camera across the tabletop; used as Adds an irregular vertical rhythm among the food.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L85B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:김창기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/6be5f966-3590-4958-b4c4-851c031c5ab5.png" ] ], "asset_id": "0bd9c0bc-b902-49ec-93d5-96b4bad70bad" }, "S81sh6": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S81sh8": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S82sh4": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "bgonly", "error": null }, "S82sh5": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S83sh2": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S83sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L21B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: In the small house courtyard at night, directly beside the porch and a metal washbasin.\n\nSHOT TEXT (authoritative, Korean): 마당 바닥에 놓인 양철 대야 위로 발을 꾹 내디디는 찰나의 김광석(광석·광돌·김광돌)의 구두 클로즈업\n\nCARRIED STATE (persist exactly): The metal washbasin remains on the courtyard floor as Gwangseok steps directly onto it.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At ground level beside the basin, the dolly-in catches 김광석's shoe descending into the upper face of the metal basin, with the compressed rim and foot centered in the lower-middle of the frame. The yard ground remains visible around the impact as a scale reference, and the low angle is held on the instant before the camera rises toward his body.\n- FRAMING SCALE: insert close-up on a detail\n- FRAME LAYOUT: 김광석's descending foot in the lower-center of the frame, foreground, moves toward metal basin on the yard ground; metal basin on yard ground in the middle-center of the frame, foreground.\n- KEY BACKGROUND ELEMENTS: 양철 대야 (Placed on the yard ground and being stepped on) — Its open upper side and rim face the low camera obliquely as the shoe presses down; used as Primary impact point beneath the descending shoe; 마당 바닥 (The basin rests on it); used as Surrounds the impact point and prevents the insert from losing spatial scale.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L21B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "6ea6a822-bb3a-455e-94de-e260e39a91c1" }, "S83sh7": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S84sh2": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "bgonly", "error": null }, "S85sh6": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S85sh6.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L88B02.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: At the backstage entrance to the university auditorium stage, where outgoing and incoming performers cross under spill from the stage lights.\n\nSHOT TEXT (authoritative, Korean): 무대 입구에서 서로를 등진 채 한 발을 뻗어 엇갈리는 mid-stride 자세의 송영숙(영숙)과 김광석(광석·광돌·김광돌)의 측면\n\nCARRIED STATE (persist exactly): Gwangseok carries his acoustic guitar and continues wearing the harmonica holder around his neck as he passes Youngsook.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From waist height beside the stage entrance, track perpendicular to the crossing path and hold both figures in a strict full-body profile as their extended strides overlap for an instant. 송영숙 crosses one side of the frame while 김광석 advances toward the other, their backs turned toward one another at the exact midpoint of passing.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 무대 입구 (Open for performers entering and leaving the stage) — The entrance is seen from the side so both opposite travel paths remain readable; used as Side-on spatial threshold separating the descending performers from the ascending soloist; 무대 (In use during the performance) — Its near side recedes away from the crossing point toward the performance area; used as Background context beyond 김광석's movement path.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L88B02.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:송영숙", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/5cff28ce-b352-45b2-83b1-c6b78deba757.png" ] ], "asset_id": "f835be3b-48f7-4d40-92c7-786adeff028f" }, "S85sh12": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S85sh18": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S85sh18.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L88B03.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Inside the auditorium’s backstage waiting room, seated opposite the senior musician after the performance.\n\nSHOT TEXT (authoritative, Korean): 김민기(민기)의 말에 깨달음을 얻은 듯 눈동자가 크게 반짝이는 김광석(광석·광돌·김광돌)의 클로즈업\n\nCARRIED STATE (persist exactly): The makgeolli cup remains in front of Gwangseok after Minki has poured it for him.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Dolly in along 김민기’s side of the conversational axis until 김광석’s face becomes a close frame at eye level, preserving 김민기’s near shoulder as a soft foreground edge. Place 김광석’s newly brightened eyes near the upper third and hold the final instant as he leans forward, absorbing the words rather than addressing the lens.\n- FRAMING SCALE: close-up\n- KEY BACKGROUND ELEMENTS: 막걸리잔 (Placed in front of 김광석); used as Soft lower-edge context from the ongoing backstage conversation.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L88B03.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "f9458d8f-ccc3-4ca2-bea9-9b56eb9946f1" }, "S86sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S86sh1.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: Inside a multipurpose music workroom equipped for rehearsal and recording, beneath a large wall logo; daylight lights the room.\n\nSHOT TEXT (authoritative, Korean): 벽면에 커다란 'SM' 로고가 박힌 음악 작업실을 흥미로운 듯 바라보며 서 있는 김광석과 한동준의 와이드 샷\n\nCARRIED STATE (persist exactly): The large “SM” logo remains fixed on the music-workroom wall.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At standing eye level from one side of the workspace, hold a static wide composition with 김광석 and 한동준 separated across the midground and the large SM logo clearly visible behind them. Capture both men paused mid-survey—one turning toward the logo and the other scanning across the room—while leaving open foreground space for the subsequent forward dolly.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: SM 로고가 있는 벽면 (Logo fixed prominently on the wall) — The logo-bearing face of the wall is visible behind the two men, with the large letters 'SM' readable; used as Prominent background identifier sharing the frame with both men; 음악 작업실 (Being inspected by 김광석 and 한동준) — The room opens across the frame around them, allowing both practice and recording use to register spatially; used as Environmental context around the two men.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:한동준", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/1176211b-735e-46dd-b952-f4d014f9221e.png" ] ], "asset_id": "fc947c60-a39c-4ad6-8314-039bd9bea7a8" }, "S86sh6": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S87sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S87sh1.png", "plate_path": null, "prompt": "Every attached image is a CHARACTER REFERENCE for proportion only. There is no location photograph — stage the location purely from THE LOCATION and SHOT TEXT below.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nTHE LOCATION: At a dining table inside the sashimi restaurant, with a large fish platter and soju between the two men.\n\nSHOT TEXT (authoritative, Korean): 테이블 위에 커다란 광어회 접시와 소주잔을 두고 마주 앉은 김광석(광석·광돌·김광돌)과 한동준(동준)의 와이드 샷\n\nCARRIED STATE (persist exactly): The platter of flounder sashimi, soy sauce, wasabi, and soju glasses remain on the table between them.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At seated eye level beside the table, hold a static wide two-shot centered on the space between 김광석 and 한동준, placing the plate of flatfish sashimi and soju glasses in the lower middle without allowing any single object to overpower the men. Their bodies angle inward across the table, establishing a relaxed conversational axis and preserving the restaurant surroundings at the frame edges.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 광어회 접시 (Placed between the two men); used as Shared central object anchoring the conversation; 소주잔 (Set on the table); used as Small-scale table context beside each diner; 테이블 (In use for the meal) — Its long plane extends between 김광석 and 한동준 and is viewed slightly from the side; used as Horizontal spatial divider between the seated men.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "no_plate": true, "ref_paths": [ [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ], [ "char:한동준", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/1176211b-735e-46dd-b952-f4d014f9221e.png" ] ], "asset_id": "3fdf3582-c3da-4b3e-9871-a48ead9febcb" }, "S87sh6": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S88sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S88sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L04B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: In the temporary seating area of a small theater auditorium, overlooking an unfinished stage where workers handle materials; work lights are on.\n\nSHOT TEXT (authoritative, Korean): 무대 자재를 쥐고 서 있는 한국인 인부들 너머로 임시 객석에 나란히 앉아있는 김민기와 김광석의 와이드 샷\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: From standing eye level behind the working 인부들, hold a static wide view through their irregular foreground spacing toward 김민기 and 김광석 seated side by side in the temporary audience area. The workers and stage materials form a partial foreground frame, while the two seated men remain clearly readable beyond them as they observe the ongoing stage work.\n- FRAMING SCALE: wide shot\n- KEY BACKGROUND ELEMENTS: 무대 자재 (Being handled during stage work) — The pieces cross portions of the foreground at varied side angles while leaving sightlines toward the temporary seating; used as Foreground framing elements held by the workers without obscuring the seated pair; 무대 (Under construction) — Its working side faces partly toward the camera and the temporary audience area; used as Midground construction context between the camera and the seated observers; 임시 객석 (Temporarily installed) — The seating faces the stage, with the two men viewed obliquely from behind the workers; used as Background seating anchor occupied by 김민기 and 김광석.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L04B01.png" ], [ "char:김민기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/68a14492-591f-4f41-b82a-6184b8416a74.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "ea748e4b-d160-4911-a213-28d9649f7567" }, "S88sh5": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S89sh3": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S89sh3.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L05B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Outside the small theater entrance, directly before a weather-beaten concert poster on the wall.\n\nSHOT TEXT (authoritative, Korean): 벽면의 포스터를 물끄러미 응시하며 굳은 의지가 담긴 눈빛으로 서 있는 김광석(광석·광돌·김광돌)의 옆얼굴\n\nCARRIED STATE (persist exactly): The concert poster remains fixed to the wall with the word “Zoo” still visible before Gwangseok crosses it out.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At shoulder height beside 김광석, finish the inward move on a close side profile so his eyes command the upper half while the poster remains along his natural eyeline at the opposite edge. Hold the poster face sufficiently legible to establish what he is studying, without yet following his arm downward.\n- FRAMING SCALE: close-up\n- FRAME LAYOUT: 김광석 in the middle-right of the frame, foreground, looks toward concert poster; concert poster in the middle-left of the frame, midground.\n- KEY BACKGROUND ELEMENTS: concert poster (Flapping in wind and rain) — The printed face is angled toward camera, showing ‘동물원 김광석의 첫 단독콘서트.’; used as Eyeline anchor retained beside 김광석’s profile; audience queue (Waiting to enter); used as Soft background indication of the waiting public.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L05B01.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "e28c83ac-6575-4df8-9a54-7aa6477679a0" }, "S89sh5": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S90sh1": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S90sh7": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S90sh7.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L01B01.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: In the dark backstage passage connecting the waiting room to the stage entrance.\n\nSHOT TEXT (authoritative, Korean): 무대를 향해 뻗은 어두운 복도를 나란히 걷고 있는 mid-stride 순간, 한 발이 앞으로 나간 김민기(민기)와 김광석(광석·광돌·김광돌)의 뒷모습\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: Track at waist height behind 김민기 and 김광석, centered on their shared route through the dark corridor, catching both men mid-stride with their backs occupying the middle third. Let the corridor’s forward convergence carry the eye toward the stage entrance while maintaining equal scale between the two figures.\n- FRAMING SCALE: wide shot\n- FRAME LAYOUT: 김민기 in the middle-left of the frame, midground, moves toward stage entrance; 김광석 in the middle-right of the frame, midground, moves toward stage entrance; stage entrance in the upper-center of the frame, background.\n- KEY BACKGROUND ELEMENTS: backstage corridor (Dark) — Its length recedes away from camera toward the stage route; used as Creates the converging passage that guides both figures forward; stage entrance (Ahead of the two men) — The opening faces back toward the approaching men and camera; used as Destination point on the tracking axis.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L01B01.png" ], [ "char:김민기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/68a14492-591f-4f41-b82a-6184b8416a74.png" ], [ "char:김광석", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/bb534045-35e7-460e-a874-555d8e0643c9.png" ] ], "asset_id": "20132607-6629-4d5f-a560-aec9a683274e" }, "S90sh9": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S91sh4": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S91sh5": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S92sh1": { "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S92sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L05B02.png", "prompt": "The FIRST attached image is the LOCATION PHOTOGRAPH (loose spatial reference only). Every following attached image is a CHARACTER REFERENCE for proportion.\n\nDraw ONE storyboard layout frame, 16:9, filling the entire image edge-to-edge.\n\nSTYLE — LAYOUT-ONLY THUMBNAIL SKETCH: the lightest possible\ndirector's layout drawing. THIN, CLEAN, UNIFORM PENCIL CONTOUR\nLINES ONLY on pure white paper. Absolutely NO shading, NO tone, NO\nhatching, NO texture, NO grayscale fill, NO light-and-shadow, NO\nfacial features beyond a bare head outline, NO clothing detail, NO\nmaterial rendering. The frame contains ONLY: the camera framing,\neach figure as a simple correct-proportion human contour in its\nexact pose and screen position/size, and the few major set masses\nas plain outline shapes. It must read as a quick layout guide,\nnever as a finished drawing.\n\nThe attached photograph is the LOCATION where this shot happens.\nUse it ONLY as a LOOSE spatial reference — the rough set masses\nand where things sit relative to each other. You do NOT need to\nfollow it exactly: do not copy its camera, its lighting or its\ndetails. Choose the camera and framing the SHOT TEXT needs, and\ndraw only the few set masses that matter as plain outline shapes.\n\nTHE LOCATION: Outside the small theater entrance at night, against the wall beside the solo-concert poster.\n\nSHOT TEXT (authoritative, Korean): '첫 단독콘서트' 포스터 앞 벽면에 기대어 입에 담배를 문 김민기(민기)의 상반신\n\nCARRIED STATE (persist exactly): The poster remains on the exterior wall with “Zoo” visibly crossed out in marker.\n\nPROPS FACE THE RIGHT WAY (also in this sketch): every handheld or used\nobject must be oriented as its real-world use requires. A person\nreading, watching or operating something (a phone, a photograph, a\npaper, any device) has its functional side — screen, front, page —\nfacing THEIR OWN eyes; the camera then sees whatever side the staging\ngeometry implies (often its back). Show the functional side to the\ncamera ONLY when the shot text itself stages it toward the viewer.\nNever flip, mirror or reverse an object's front and back.\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, gaze on the target the SHOT TEXT implies; avoid\na stiff attention stance (feet together, arms hanging straight down)\nand a blank stare into the lens. If the SHOT TEXT or any POSE contract\nabove stages stillness, death, sleep, unconsciousness, restraint,\ndrill or an explicit direct-to-camera look, draw THAT exactly — this\nclause never overrides it.\n\nKEYSHOT LAYOUT VALUE: this storyboard frame is a KEY FRAME extracted\nfrom a finished film — its layout alone must convey why this moment\nmatters in the scene. Compose the frame so the subject's pose, gaze\ntarget and spatial relationships read clearly. Do not fill the frame\nwith a single face and do not draw a subject standing square to the\ncamera in a passport-photo composition unless the SHOT TEXT itself\nexplicitly stages that framing for a narrative reason. Even in an\nextreme close-up, keep at least one meaning-bearing element (the gaze\ntarget, the object reacted to, the relevant surroundings) inside or\nat the edge of the frame.\n\nABSOLUTELY NO TEXT anywhere: no captions, no numbers, no labels of any kind.\n\nCAMERA & FRAME (follow exactly — this is the composition authority for this still; the reference images supply identity and place, never the framing):\n- CAMERA: At chest height and medium distance on the established frontal axis, hold 김민기’s upper body against the wall with the altered concert poster clearly visible behind one shoulder. Turn his head slightly away from the lens and catch the cigarette held in his mouth, keeping his weight visibly settled into the wall before the camera begins its later advance.\n- FRAMING SCALE: medium shot\n- KEY BACKGROUND ELEMENTS: first solo concert poster (Displayed on the wall) — The printed face is visible behind 김민기, reading ‘김광석의 첫 단독콘서트’ with ‘동물원’ crossed out; used as Narrative backdrop identifying the concert and the earlier erasure; exterior wall (Supporting 김민기); used as Physical support for 김민기’s leaning posture.\nCompose the frame exactly as specified above — camera angle, subject scale and screen placement. Keep true physical scale between people and background elements: fixtures and distant objects occupy only the small screen area their real size and distance dictate; never enlarge a background object into a foreground presence.\n\nPERSPECTIVE & SCALE GUIDES (draw these as part of the sketch — they are the whole point of this layout):\n- Draw the HORIZON LINE at the camera's eye height, and 2-3 thin PERSPECTIVE GUIDE LINES converging toward the vanishing point, so the depth of the space is explicit.\n- TRUE RELATIVE SIZE: draw every person and every set mass at the exact screen size that its real-world size and its distance from the camera dictate. A high wall-mounted device or any distant object stays SMALL against the architecture; people scale with their depth along the guide lines. NEVER enlarge a background object beyond its true perspective size.\n- For EVERY person draw one thin EYELINE ARROW from the eyes to the exact thing they look at; if something moves, one thin MOTION ARROW along its path.\n- All guides are simple thin pencil lines — layout language only, never objects in the scene, and still NO text, numbers or letters anywhere.", "skipped_reason": null, "error": null, "ref_paths": [ [ "plate", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L05B02.png" ], [ "char:김민기", "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/reference/68a14492-591f-4f41-b82a-6184b8416a74.png" ] ], "asset_id": "b50da4ce-9190-42b5-a3ab-571c2fc24ce0" }, "S92sh3": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S93sh5": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S93sh12": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S93sh16": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null }, "S94sh1": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "bgonly", "error": null }, "S94sh6": { "image_path": null, "plate_path": null, "prompt": null, "skipped_reason": "prev", "error": null } }, "map_conti": {}, "conti_pack": "5.202607222353", "lane_conti": { "S3sh1": { "lane": "map_marker", "group_id": "bg_university_campus", "segment_id": "SEG_01", "status": "ok", "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/lane_sketch_S3sh1.png", "base_map_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/canon_bg_university_campus_map.png", "base_map_asset_id": "c957a028-363a-4e48-a567-ca6a55e8d045", "marker_map_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/lane_marker_map_S3sh1.png", "marker_check": { "markers_ok": true, "details_ko": "정상" }, "marker_prompt": "Edit the attached TOP-DOWN SITE PLAN image: draw STAGING MARKERS on it\nat the positions described below. This is a planning aid for a film\ncrew, drawn over the plan.\n\nPLACEMENT AUTHORITY: each marker line below places its marker relative\nto a fixed feature, zone or boundary VISIBLE on the attached plan —\nthat landmark-relative description is the placement authority. The\nnormalized fractions in parentheses are coarse estimates of the same\nspot (x runs left→right 0..1, y runs top→bottom 0..1 of the full\ncanvas) — use them only to confirm the rough region; when in doubt,\nthe landmark description wins. Do NOT relocate a marker onto a\ndifferent salient line or feature than the one its description names.\n\nNEVER move a marker to make the drawing fit. The camera circle and\nevery subject circle go exactly where their descriptions place them,\neven when that leaves the view wedge or an arrow running off the edge\nof the image — in that case just let the wedge or arrow continue to\nthe image border and stop there. Shifting the camera outward, pulling\na subject aside, or shrinking the wedge to keep it inside the canvas\nall break the plan this map exists to record.\n\nDraw EXACTLY these markers, flat and high-contrast, small enough not\nto hide the plan under them:\n- CAMERA: one filled circle at the camera position, labelled CAM, and\n out of that circle a FIELD-OF-VIEW WEDGE showing what this shot\n sees — two straight thin lines, one running toward each of the two\n frame-edge positions given below, and a third thinner line to the\n look-target position between them. Fill the area between the two\n edge lines with a light translucent tint so the wedge reads as the\n camera's view; the plan underneath must stay fully readable through\n the tint. Every subject listed below stands inside this wedge —\n if your drawing leaves one outside it, the wedge is drawn wrong,\n not the subject.\n- SUBJECTS: one filled circle per subject at its position, labelled\n with its slot code (E1, E2, ...) followed by its short description\n in parentheses, and one short ARROW leaving that circle toward the\n spot the subject faces, so the direction each subject is turned is\n readable at a glance.\n\nChange NOTHING else: keep every zone, landmark, line, label and color\nof the plan exactly as attached — do not redraw, recolor, move or\nremove anything, do not add scenery, shadows or any other symbols or\ntext beyond the markers, wedge, arrows and labels specified above. The\noutput is the SAME plan image with only these staging markers added.\n\nMARKERS TO DRAW:\n- CAM (camera): circle At the southern entrance edge of the Outdoor Campus Ground. (coarse estimate x=0.50, y=0.88); its sight line ends at LOOK TARGET: In the open north-central portion of the Outdoor Campus Ground. (coarse estimate x=0.50, y=0.48); the frame's LEFT edge cuts through At the western boundary of the Outdoor Campus Ground beside the long planted strip. (coarse estimate x=0.20, y=0.43) and its RIGHT edge through At the eastern boundary of the Outdoor Campus Ground where the Campus Approach joins it. (coarse estimate x=0.79, y=0.43)\n- E1 (university student holding a raised sign): circle On the southwestern side of the Outdoor Campus Ground, north of its southern entrance. (coarse estimate x=0.41, y=0.72), facing In the open left-center portion of the Outdoor Campus Ground. (coarse estimate x=0.45, y=0.58)\n- E2 (university student carrying a placard): circle On the southeastern side of the Outdoor Campus Ground, north of its southern entrance. (coarse estimate x=0.59, y=0.71), facing In the open right-center portion of the Outdoor Campus Ground. (coarse estimate x=0.55, y=0.58)\n- E3 (university student calling through a megaphone): circle On the west side of the Outdoor Campus Ground beside the open central corridor. (coarse estimate x=0.35, y=0.59), facing In the open central corridor of the Outdoor Campus Ground. (coarse estimate x=0.49, y=0.59)\n- E4 (university student holding a raised sign): circle On the east side of the Outdoor Campus Ground, south of the Campus Approach junction. (coarse estimate x=0.65, y=0.57), facing In the open central corridor of the Outdoor Campus Ground. (coarse estimate x=0.51, y=0.59)\n- E5 (university student speaking into a megaphone): circle On the left side of the north-central gathering area of the Outdoor Campus Ground. (coarse estimate x=0.43, y=0.49), facing In the open central corridor leading from the southern entrance of the Outdoor Campus Ground. (coarse estimate x=0.49, y=0.65)\n- E6 (university student displaying a placard): circle On the right side of the north-central gathering area of the Outdoor Campus Ground. (coarse estimate x=0.57, y=0.47), facing In the open central corridor leading from the southern entrance of the Outdoor Campus Ground. (coarse estimate x=0.51, y=0.64)", "marker_attempts": 1, "geometry": { "camera": { "look_target": { "x": 0.5, "y": 0.48 }, "look_target_anchor_en": "In the open north-central portion of the Outdoor Campus Ground.", "origin": { "x": 0.5, "y": 0.875 }, "origin_anchor_en": "At the southern entrance edge of the Outdoor Campus Ground.", "view_left": { "x": 0.2, "y": 0.43 }, "view_left_anchor_en": "At the western boundary of the Outdoor Campus Ground beside the long planted strip.", "view_right": { "x": 0.79, "y": 0.43 }, "view_right_anchor_en": "At the eastern boundary of the Outdoor Campus Ground where the Campus Approach joins it." }, "entity_placements": [ { "anchor_en": "On the southwestern side of the Outdoor Campus Ground, north of its southern entrance.", "camera_facing_evidence": "", "camera_facing_relation": "unspecified", "facing": { "x": 0.45, "y": 0.58 }, "facing_anchor_en": "In the open left-center portion of the Outdoor Campus Ground.", "slot": "E1", "subject_en": "university student holding a raised sign", "x": 0.405, "y": 0.72 }, { "anchor_en": "On the southeastern side of the Outdoor Campus Ground, north of its southern entrance.", "camera_facing_evidence": "", "camera_facing_relation": "unspecified", "facing": { "x": 0.55, "y": 0.58 }, "facing_anchor_en": "In the open right-center portion of the Outdoor Campus Ground.", "slot": "E2", "subject_en": "university student carrying a placard", "x": 0.595, "y": 0.71 }, { "anchor_en": "On the west side of the Outdoor Campus Ground beside the open central corridor.", "camera_facing_evidence": "", "camera_facing_relation": "unspecified", "facing": { "x": 0.49, "y": 0.59 }, "facing_anchor_en": "In the open central corridor of the Outdoor Campus Ground.", "slot": "E3", "subject_en": "university student calling through a megaphone", "x": 0.35, "y": 0.59 }, { "anchor_en": "On the east side of the Outdoor Campus Ground, south of the Campus Approach junction.", "camera_facing_evidence": "", "camera_facing_relation": "unspecified", "facing": { "x": 0.51, "y": 0.59 }, "facing_anchor_en": "In the open central corridor of the Outdoor Campus Ground.", "slot": "E4", "subject_en": "university student holding a raised sign", "x": 0.65, "y": 0.57 }, { "anchor_en": "On the left side of the north-central gathering area of the Outdoor Campus Ground.", "camera_facing_evidence": "", "camera_facing_relation": "unspecified", "facing": { "x": 0.49, "y": 0.65 }, "facing_anchor_en": "In the open central corridor leading from the southern entrance of the Outdoor Campus Ground.", "slot": "E5", "subject_en": "university student speaking into a megaphone", "x": 0.43, "y": 0.49 }, { "anchor_en": "On the right side of the north-central gathering area of the Outdoor Campus Ground.", "camera_facing_evidence": "", "camera_facing_relation": "unspecified", "facing": { "x": 0.51, "y": 0.64 }, "facing_anchor_en": "In the open central corridor leading from the southern entrance of the Outdoor Campus Ground.", "slot": "E6", "subject_en": "university student displaying a placard", "x": 0.57, "y": 0.47 } ], "rationale_ko": "카메라는 광장 남쪽 입구 가장자리에서 북쪽 군중을 향하도록 배치했다. 전경 학생들은 좌우에서 프레임에 들어오며, 중앙에는 전진 트래킹이 통과할 수 있는 열린 통로를 남겼다. 확성기와 팻말을 든 학생들을 좌우와 북쪽에 분산해 붐비는 와이드 샷의 깊이와 밀도를 만들었다." }, "leakage": { "has_marker_leakage": false, "details_ko": "도면 마커나 텍스트 같은 유출 흔적이 없는 투시도 렌더링입니다." }, "prompt": "Draw a single CAMERA-VIEW storyboard frame (16:9) of the scene staged\non the attached MARKER-ANNOTATED TOP-DOWN SITE PLAN, seen from the\ncamera position marked on it.\n\nThe attached plan carries STAGING MARKERS drawn over it: a\ncamera-position marker with a field-of-view wedge (where the camera\nstands, which way it looks and how wide it sees) and one labelled\ncircle per subject with an arrow (where each subject stands and which\nway that subject is turned). Jurisdiction of sources — follow all three, they do not\nconflict:\n- The SITE PLAN under the markers is the SOT for what exists on the\n site and how its zones and landmarks are arranged relative to each\n other. The markers themselves are staging aids, NOT site content.\n- The markers on the plan are the SOT for XY POSITIONS, horizontal\n heading and body facing ONLY: where each subject stands on that\n plan, which way each one is turned, where the camera stands, which\n way it looks and how wide its wedge opens. Distances, directions\n and relative depths between subjects come from the marked\n positions; what the wedge covers is what this frame contains, and\n subjects outside the wedge are out of frame. The STAGING GEOMETRY\n text below restates the same staging with coarse normalized\n estimates — it never overrides the marked plan.\n- The CAMERA DIRECTION text below (when present) is the SOT for the\n VERTICAL angle, elevation, tilt and framing (eye-level, high angle,\n overhead crane, low angle, ...). If it asks for a high or overhead\n angle, draw that angle — the site plan being top-down does NOT mean\n the frame must be eye-level. It does NOT govern horizontal\n placement: which side of the frame a subject occupies and which way\n that subject is turned come from the marked plan and the staging\n text, even when that memo also mentions left, right, foreground or\n background.\n\nSTYLE — MANNEQUIN BLOCKING SKETCH: thin uniform dark outlines on plain\npaper, nothing else. Draw every person as a PLAIN ARTIST'S MANNEQUIN:\na featureless oval head with NO face, NO hair, NO clothing and no\ncostume detail — simple tapered limbs and torso, like a wooden posing\nfigure. Because the head has no face and no hair, the ONLY thing that\nshows which way a figure is turned is the BODY ITSELF: the direction\nthe chest, hips and feet point, and the way the head-oval sits on the\nshoulders. Aim each mannequin's chest and feet at the spot the staging\ntext says that person faces, and let the arms carry the action. A\nmannequin turned to the wrong side is the single worst error in this\ndrawing. Environment is reduced to a few placement lines (ground line,\nkey landmark silhouettes) — just enough to read positions and depth.\n\nSTAGING GEOMETRY — PLANNING DATA ONLY, NEVER RENDER IT: the lines\nbelow restate, marker by marker, the staging already marked on the\nattached plan: each landmark-relative description restates the\nmarker-map authority in words, and the normalized fractions in\nparentheses (x left→right 0..1, y top→bottom 0..1) are coarse audit\nestimates only — neither ever overrides the marked plan.\n\nThose fractions are TOP-DOWN bookkeeping. Never read them as left and\nright of your frame: on a plan seen from above, a subject \"to the\nright\" can sit on either side of a camera that looks any direction.\nEach subject line therefore ends with the SAME staging expressed from\nTHIS camera — which part of the frame the subject occupies, and which\nway that subject faces in screen terms (toward the left or right of\nthe frame, toward or away from the camera or side-on to it in\nprofile, and whom they are looking at). Those screen-space statements\nare what your drawing must match:\na figure said to face the right of the frame is turned to the right in\nyour drawing, never mirrored.\n\nDo NOT draw any of this data in the image — no numbers, no\ncoordinates, no labels, no arrows, no cones, no dots, no diagram\nsymbols of any kind.\n\nWhere the CAMERA DIRECTION text below also describes horizontal\nplacement (which side of the frame someone occupies, foreground left,\nbackground right), treat that as background reading only: for anything\nabout left, right and where a subject sits in this frame, the staging\nabove wins. CAMERA DIRECTION governs the vertical angle, elevation,\ntilt and framing.\n\nFor whether a subject faces the camera or is turned away from it, each\nsubject line above already states it in screen terms and the SHOT TEXT\nis the source it was derived from — draw exactly what the subject line\nsays, and never mirror it.\n- the camera stands At the southern entrance edge of the Outdoor Campus Ground. (coarse estimate x=0.50, y=0.88); its sight line ends at LOOK TARGET: In the open north-central portion of the Outdoor Campus Ground. (coarse estimate x=0.50, y=0.48); the frame's LEFT edge cuts through At the western boundary of the Outdoor Campus Ground beside the long planted strip. (coarse estimate x=0.20, y=0.43) and its RIGHT edge through At the eastern boundary of the Outdoor Campus Ground where the Campus Approach joins it. (coarse estimate x=0.79, y=0.43) — everything between those two edges is in frame, anything outside them is out of frame\n- university student holding a raised sign stands On the southwestern side of the Outdoor Campus Ground, north of its southern entrance. (coarse estimate x=0.41, y=0.72), facing In the open left-center portion of the Outdoor Campus Ground. (coarse estimate x=0.45, y=0.58) — from this camera the subject is in the LEFT part of the frame and faces toward the RIGHT of the frame and away from the camera, into the depth of the shot, so the camera sees this subject from behind\n- university student carrying a placard stands On the southeastern side of the Outdoor Campus Ground, north of its southern entrance. (coarse estimate x=0.59, y=0.71), facing In the open right-center portion of the Outdoor Campus Ground. (coarse estimate x=0.55, y=0.58) — from this camera the subject is in the RIGHT part of the frame and faces toward the LEFT of the frame and away from the camera, into the depth of the shot, so the camera sees this subject from behind\n- university student calling through a megaphone stands On the west side of the Outdoor Campus Ground beside the open central corridor. (coarse estimate x=0.35, y=0.59), facing In the open central corridor of the Outdoor Campus Ground. (coarse estimate x=0.49, y=0.59) — from this camera the subject is in the LEFT part of the frame and faces toward the RIGHT of the frame\n- university student holding a raised sign stands On the east side of the Outdoor Campus Ground, south of the Campus Approach junction. (coarse estimate x=0.65, y=0.57), facing In the open central corridor of the Outdoor Campus Ground. (coarse estimate x=0.51, y=0.59) — from this camera the subject is in the RIGHT part of the frame and faces toward the LEFT of the frame\n- university student speaking into a megaphone stands On the left side of the north-central gathering area of the Outdoor Campus Ground. (coarse estimate x=0.43, y=0.49), facing In the open central corridor leading from the southern entrance of the Outdoor Campus Ground. (coarse estimate x=0.49, y=0.65) — from this camera the subject is in the LEFT part of the frame and faces toward the RIGHT of the frame and toward the camera, so the camera sees this subject from the front\n- university student displaying a placard stands On the right side of the north-central gathering area of the Outdoor Campus Ground. (coarse estimate x=0.57, y=0.47), facing In the open central corridor leading from the southern entrance of the Outdoor Campus Ground. (coarse estimate x=0.51, y=0.64) — from this camera the subject is near the CENTER of the frame and faces toward the LEFT of the frame and toward the camera, so the camera sees this subject from the front\n\nCAMERA DIRECTION (vertical angle / elevation / framing SOT): At eye level along one edge of the plaza, the camera begins a broad forward track into the dense campus gathering. Foreground students enter from both sides while a central corridor remains open into the crowd; raised signs and megaphones punctuate the upper half without overwhelming the human scale.\n\nTHE LOCATION: 명지대 캠퍼스 광장\n\nSHOT TEXT (authoritative, Korean): 햇살이 내리쬐는 캠퍼스 광장에서 확성기를 들거나 팻말을 든 한국인 대학생들이 붐비는 와이드 샷\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, head-oval and chest turned toward the target the\nSHOT TEXT implies; avoid a stiff attention stance (feet together, arms\nhanging straight down) and a head-oval squared flat to the lens. If the\nSHOT TEXT or any POSE contract above stages stillness, death, sleep,\nunconsciousness, restraint, drill or an explicit direct-to-camera look,\ndraw THAT exactly — this clause never overrides it.\n\nCRITICAL — THE SITE PLAN IS A STAGING REFERENCE, NOT CONTENT: do NOT\ndraw the plan itself, and do NOT draw any diagram elements in the\nframe — no circles, dots, wedges, cones, arrows, callouts, codes or\nany other overlay symbols. The output is a clean storyboard drawing\nof the scene as seen by the camera only.\n\nNo text, no letters, no numbers, no captions anywhere in the image.", "reused": true, "lane_sketch_pack": "14.202607270344", "lane_geometry_pack": "4.202607270248", "mannequin_chain_contract": "1", "asset_id": "30e5437c-7a2a-48e0-b412-9e1bf2ccec46" }, "S8sh8": { "lane": "structure_plate", "group_id": "bg_university_library_exterior", "segment_id": "SEG_01", "status": "ab_select_ready", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_university_library_exterior_sel.png", "seed_asset_id": "5ba93641-4859-4b4c-a21d-f114f004c31d", "seed_status": "ok", "bg_source": "seed", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_university_library_exterior_sel.png", "bg_reason": "", "conti_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S8sh8.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_university_library_exterior_sel.png" }, "S20sh6": { "lane": "structure_plate", "group_id": "bg_affluent_house", "segment_id": "seg_front_approach", "status": "ab_select_ready", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_affluent_house_sel.png", "seed_asset_id": "039b239d-6fbe-4620-83b2-7b5f9e25f450", "seed_status": "ok", "bg_source": "plate", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L33B02.png", "bg_reason": "", "conti_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S20sh6.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L33B02.png" }, "S21sh1": { "lane": "structure_plate", "group_id": "bg_basement_cafe", "segment_id": "SEG_01", "status": "ab_select_bypass", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_basement_cafe_sel.png", "seed_asset_id": "bf16d642-a79a-4c66-a038-f27d0812282f", "seed_status": "ok", "bg_source": "plate", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L29B01.png", "bg_reason": "", "bypass_reason": "bg_only" }, "S24sh2": { "lane": "structure_plate", "group_id": "bg_affluent_house", "segment_id": "seg_yard_terrace", "status": "ab_select_ready", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_affluent_house_sel.png", "seed_asset_id": "039b239d-6fbe-4620-83b2-7b5f9e25f450", "seed_status": "ok", "bg_source": "plate", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L35B01.png", "bg_reason": "", "conti_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S24sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L35B01.png" }, "S26sh2": { "lane": "structure_plate", "group_id": "bg_university_district_street", "segment_id": "SEG_STRUCTURE_FRONTAGE", "status": "ab_select_ready", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_university_district_street_sel.png", "seed_asset_id": "7ea812c9-5faa-4ed4-a85b-5e84a54d9e7e", "seed_status": "ok", "bg_source": "seed", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_university_district_street_sel.png", "bg_reason": "", "conti_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S26sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_university_district_street_sel.png" }, "S26sh3": { "lane": "structure_plate", "group_id": "bg_university_district_street", "segment_id": "SEG_STRUCTURE_FRONTAGE", "status": "ab_select_bypass", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_university_district_street_sel.png", "seed_asset_id": "7ea812c9-5faa-4ed4-a85b-5e84a54d9e7e", "seed_status": "ok", "bg_source": "seed", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_university_district_street_sel.png", "bg_reason": "", "bypass_reason": "prev" }, "S28sh2": { "lane": "structure_plate", "group_id": "bg_downtown_building_front", "segment_id": "SEG_01", "status": "ab_select_ready", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_downtown_building_front_sel.png", "seed_asset_id": "33b23b3a-8cdb-49d5-a07f-257eff773ea5", "seed_status": "ok", "bg_source": "seed", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_downtown_building_front_sel.png", "bg_reason": "", "conti_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S28sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_downtown_building_front_sel.png" }, "S30sh2": { "lane": "structure_plate", "group_id": "bg_affluent_house", "segment_id": "seg_front_approach", "status": "ab_select_ready", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_affluent_house_sel.png", "seed_asset_id": "039b239d-6fbe-4620-83b2-7b5f9e25f450", "seed_status": "ok", "bg_source": "plate", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L33B01.png", "bg_reason": "", "conti_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S30sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L33B01.png" }, "S30sh3": { "lane": "structure_plate", "group_id": "bg_affluent_house", "segment_id": "seg_front_approach", "status": "ab_select_bypass", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_affluent_house_sel.png", "seed_asset_id": "039b239d-6fbe-4620-83b2-7b5f9e25f450", "seed_status": "ok", "bg_source": "plate", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L33B01.png", "bg_reason": "", "bypass_reason": "prev" }, "S32sh2": { "lane": "map_marker", "group_id": "bg_memorial_cemetery", "segment_id": "seg_approach_path", "status": "ok", "image_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/lane_sketch_S32sh2.png", "base_map_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/canon_bg_memorial_cemetery_map.png", "base_map_asset_id": "02639f68-8226-476a-a822-0480235781a5", "marker_map_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/lane_marker_map_S32sh2.png", "marker_check": { "markers_ok": true, "details_ko": "정상" }, "marker_prompt": "Edit the attached TOP-DOWN SITE PLAN image: draw STAGING MARKERS on it\nat the positions described below. This is a planning aid for a film\ncrew, drawn over the plan.\n\nPLACEMENT AUTHORITY: each marker line below places its marker relative\nto a fixed feature, zone or boundary VISIBLE on the attached plan —\nthat landmark-relative description is the placement authority. The\nnormalized fractions in parentheses are coarse estimates of the same\nspot (x runs left→right 0..1, y runs top→bottom 0..1 of the full\ncanvas) — use them only to confirm the rough region; when in doubt,\nthe landmark description wins. Do NOT relocate a marker onto a\ndifferent salient line or feature than the one its description names.\n\nNEVER move a marker to make the drawing fit. The camera circle and\nevery subject circle go exactly where their descriptions place them,\neven when that leaves the view wedge or an arrow running off the edge\nof the image — in that case just let the wedge or arrow continue to\nthe image border and stop there. Shifting the camera outward, pulling\na subject aside, or shrinking the wedge to keep it inside the canvas\nall break the plan this map exists to record.\n\nDraw EXACTLY these markers, flat and high-contrast, small enough not\nto hide the plan under them:\n- CAMERA: one filled circle at the camera position, labelled CAM, and\n out of that circle a FIELD-OF-VIEW WEDGE showing what this shot\n sees — two straight thin lines, one running toward each of the two\n frame-edge positions given below, and a third thinner line to the\n look-target position between them. Fill the area between the two\n edge lines with a light translucent tint so the wedge reads as the\n camera's view; the plan underneath must stay fully readable through\n the tint. Every subject listed below stands inside this wedge —\n if your drawing leaves one outside it, the wedge is drawn wrong,\n not the subject.\n- SUBJECTS: one filled circle per subject at its position, labelled\n with its slot code (E1, E2, ...) followed by its short description\n in parentheses, and one short ARROW leaving that circle toward the\n spot the subject faces, so the direction each subject is turned is\n readable at a glance.\n\nChange NOTHING else: keep every zone, landmark, line, label and color\nof the plan exactly as attached — do not redraw, recolor, move or\nremove anything, do not add scenery, shadows or any other symbols or\ntext beyond the markers, wedge, arrows and labels specified above. The\noutput is the SAME plan image with only these staging markers added.\n\nMARKERS TO DRAW:\n- CAM (camera): circle On the grass between the second and third grave-marker rows south of the inter-row cemetery path. (coarse estimate x=0.35, y=0.76); its sight line ends at LOOK TARGET: On the grass directly in front of the inscribed grave marker beside the inter-row cemetery path. (coarse estimate x=0.42, y=0.61); the frame's LEFT edge cuts through On the grass along the south edge of the inter-row cemetery path. (coarse estimate x=0.20, y=0.53) and its RIGHT edge through On the grass between neighboring markers in the first grave-marker row south of the inter-row cemetery path. (coarse estimate x=0.70, y=0.58)\n- E1 (standing young man): circle On the grass immediately in front and slightly left of the inscribed grave marker beside the inter-row cemetery path. (coarse estimate x=0.39, y=0.65), facing At the inscribed grave marker beside the inter-row cemetery path. (coarse estimate x=0.42, y=0.59)\n- E2 (standing young woman): circle On the grass immediately in front and slightly right of the inscribed grave marker beside the inter-row cemetery path. (coarse estimate x=0.45, y=0.65), facing At the inscribed grave marker beside the inter-row cemetery path. (coarse estimate x=0.42, y=0.59)", "marker_attempts": 1, "geometry": { "camera": { "look_target": { "x": 0.42, "y": 0.61 }, "look_target_anchor_en": "On the grass directly in front of the inscribed grave marker beside the inter-row cemetery path.", "origin": { "x": 0.35, "y": 0.76 }, "origin_anchor_en": "On the grass between the second and third grave-marker rows south of the inter-row cemetery path.", "view_left": { "x": 0.2, "y": 0.53 }, "view_left_anchor_en": "On the grass along the south edge of the inter-row cemetery path.", "view_right": { "x": 0.7, "y": 0.58 }, "view_right_anchor_en": "On the grass between neighboring markers in the first grave-marker row south of the inter-row cemetery path." }, "entity_placements": [ { "anchor_en": "On the grass immediately in front and slightly left of the inscribed grave marker beside the inter-row cemetery path.", "camera_facing_evidence": "", "camera_facing_relation": "unspecified", "facing": { "x": 0.421, "y": 0.592 }, "facing_anchor_en": "At the inscribed grave marker beside the inter-row cemetery path.", "slot": "E1", "subject_en": "standing young man", "x": 0.39, "y": 0.65 }, { "anchor_en": "On the grass immediately in front and slightly right of the inscribed grave marker beside the inter-row cemetery path.", "camera_facing_evidence": "", "camera_facing_relation": "unspecified", "facing": { "x": 0.421, "y": 0.592 }, "facing_anchor_en": "At the inscribed grave marker beside the inter-row cemetery path.", "slot": "E2", "subject_en": "standing young woman", "x": 0.45, "y": 0.65 } ], "rationale_ko": "두 사람을 김광동 비석 바로 앞 잔디에 함께 멈춰 세우고 모두 비석을 바라보게 했다. 남성을 화면 중심에 가깝게, 여성을 옆으로 치우쳐 배치했으며, 전방 좌측 사분면의 카메라에서 비석이 두 사람 사이에 보이도록 구성했다. 넓은 화각에는 뒤로 이어지는 묘비 행과 묘지 통로가 함께 들어온다." }, "leakage": { "has_marker_leakage": false, "details_ko": "스케치에 마커나 다이어그램 기호가 보이지 않음." }, "prompt": "Draw a single CAMERA-VIEW storyboard frame (16:9) of the scene staged\non the attached MARKER-ANNOTATED TOP-DOWN SITE PLAN, seen from the\ncamera position marked on it.\n\nThe attached plan carries STAGING MARKERS drawn over it: a\ncamera-position marker with a field-of-view wedge (where the camera\nstands, which way it looks and how wide it sees) and one labelled\ncircle per subject with an arrow (where each subject stands and which\nway that subject is turned). Jurisdiction of sources — follow all three, they do not\nconflict:\n- The SITE PLAN under the markers is the SOT for what exists on the\n site and how its zones and landmarks are arranged relative to each\n other. The markers themselves are staging aids, NOT site content.\n- The markers on the plan are the SOT for XY POSITIONS, horizontal\n heading and body facing ONLY: where each subject stands on that\n plan, which way each one is turned, where the camera stands, which\n way it looks and how wide its wedge opens. Distances, directions\n and relative depths between subjects come from the marked\n positions; what the wedge covers is what this frame contains, and\n subjects outside the wedge are out of frame. The STAGING GEOMETRY\n text below restates the same staging with coarse normalized\n estimates — it never overrides the marked plan.\n- The CAMERA DIRECTION text below (when present) is the SOT for the\n VERTICAL angle, elevation, tilt and framing (eye-level, high angle,\n overhead crane, low angle, ...). If it asks for a high or overhead\n angle, draw that angle — the site plan being top-down does NOT mean\n the frame must be eye-level. It does NOT govern horizontal\n placement: which side of the frame a subject occupies and which way\n that subject is turned come from the marked plan and the staging\n text, even when that memo also mentions left, right, foreground or\n background.\n\nSTYLE — MANNEQUIN BLOCKING SKETCH: thin uniform dark outlines on plain\npaper, nothing else. Draw every person as a PLAIN ARTIST'S MANNEQUIN:\na featureless oval head with NO face, NO hair, NO clothing and no\ncostume detail — simple tapered limbs and torso, like a wooden posing\nfigure. Because the head has no face and no hair, the ONLY thing that\nshows which way a figure is turned is the BODY ITSELF: the direction\nthe chest, hips and feet point, and the way the head-oval sits on the\nshoulders. Aim each mannequin's chest and feet at the spot the staging\ntext says that person faces, and let the arms carry the action. A\nmannequin turned to the wrong side is the single worst error in this\ndrawing. Environment is reduced to a few placement lines (ground line,\nkey landmark silhouettes) — just enough to read positions and depth.\n\nSTAGING GEOMETRY — PLANNING DATA ONLY, NEVER RENDER IT: the lines\nbelow restate, marker by marker, the staging already marked on the\nattached plan: each landmark-relative description restates the\nmarker-map authority in words, and the normalized fractions in\nparentheses (x left→right 0..1, y top→bottom 0..1) are coarse audit\nestimates only — neither ever overrides the marked plan.\n\nThose fractions are TOP-DOWN bookkeeping. Never read them as left and\nright of your frame: on a plan seen from above, a subject \"to the\nright\" can sit on either side of a camera that looks any direction.\nEach subject line therefore ends with the SAME staging expressed from\nTHIS camera — which part of the frame the subject occupies, and which\nway that subject faces in screen terms (toward the left or right of\nthe frame, toward or away from the camera or side-on to it in\nprofile, and whom they are looking at). Those screen-space statements\nare what your drawing must match:\na figure said to face the right of the frame is turned to the right in\nyour drawing, never mirrored.\n\nDo NOT draw any of this data in the image — no numbers, no\ncoordinates, no labels, no arrows, no cones, no dots, no diagram\nsymbols of any kind.\n\nWhere the CAMERA DIRECTION text below also describes horizontal\nplacement (which side of the frame someone occupies, foreground left,\nbackground right), treat that as background reading only: for anything\nabout left, right and where a subject sits in this frame, the staging\nabove wins. CAMERA DIRECTION governs the vertical angle, elevation,\ntilt and framing.\n\nFor whether a subject faces the camera or is turned away from it, each\nsubject line above already states it in screen terms and the SHOT TEXT\nis the source it was derived from — draw exactly what the subject line\nsays, and never mirror it.\n- the camera stands On the grass between the second and third grave-marker rows south of the inter-row cemetery path. (coarse estimate x=0.35, y=0.76); its sight line ends at LOOK TARGET: On the grass directly in front of the inscribed grave marker beside the inter-row cemetery path. (coarse estimate x=0.42, y=0.61); the frame's LEFT edge cuts through On the grass along the south edge of the inter-row cemetery path. (coarse estimate x=0.20, y=0.53) and its RIGHT edge through On the grass between neighboring markers in the first grave-marker row south of the inter-row cemetery path. (coarse estimate x=0.70, y=0.58) — everything between those two edges is in frame, anything outside them is out of frame\n- standing young man stands On the grass immediately in front and slightly left of the inscribed grave marker beside the inter-row cemetery path. (coarse estimate x=0.39, y=0.65), facing At the inscribed grave marker beside the inter-row cemetery path. (coarse estimate x=0.42, y=0.59) — from this camera the subject is near the CENTER of the frame and faces away from the camera, into the depth of the shot, so the camera sees this subject from behind\n- standing young woman stands On the grass immediately in front and slightly right of the inscribed grave marker beside the inter-row cemetery path. (coarse estimate x=0.45, y=0.65), facing At the inscribed grave marker beside the inter-row cemetery path. (coarse estimate x=0.42, y=0.59) — from this camera the subject is in the RIGHT part of the frame and faces toward the LEFT of the frame and away from the camera, into the depth of the shot, so the camera sees this subject from behind\n\nCAMERA DIRECTION (vertical angle / elevation / framing SOT): From standing eye level on the grave marker's front quarter, hold a static wide composition with 김광석 and 송영숙 stopped together in the midground. Place the inscribed marker between and slightly below their sightlines, with 김광석 nearer the center and 송영숙 offset beside him; surrounding grave markers recede behind them to establish the cemetery layout.\n\nTHE LOCATION: 국립묘지\n\nSHOT TEXT (authoritative, Korean): '육군 대위 김광동'이라 적힌 비석 앞에 우두커니 멈춰 선 김광석(광석·광돌·김광돌)과 송영숙(영숙)\n\nNATURAL PERFORMANCE (default only — every explicit direction above\nwins): when no pose contract or shot-text direction says otherwise,\nsketched figures read as alive in mid-moment — believable weight\nshift, hands engaged, head-oval and chest turned toward the target the\nSHOT TEXT implies; avoid a stiff attention stance (feet together, arms\nhanging straight down) and a head-oval squared flat to the lens. If the\nSHOT TEXT or any POSE contract above stages stillness, death, sleep,\nunconsciousness, restraint, drill or an explicit direct-to-camera look,\ndraw THAT exactly — this clause never overrides it.\n\nCRITICAL — THE SITE PLAN IS A STAGING REFERENCE, NOT CONTENT: do NOT\ndraw the plan itself, and do NOT draw any diagram elements in the\nframe — no circles, dots, wedges, cones, arrows, callouts, codes or\nany other overlay symbols. The output is a clean storyboard drawing\nof the scene as seen by the camera only.\n\nNo text, no letters, no numbers, no captions anywhere in the image.", "reused": true, "lane_sketch_pack": "14.202607270344", "lane_geometry_pack": "4.202607270248", "mannequin_chain_contract": "1", "asset_id": "5dc1c8a4-231a-44ad-8e8b-765b44135121" }, "S34sh2": { "lane": "structure_plate", "group_id": "bg_basement_office", "segment_id": "SEG_STAIR_THRESHOLD", "status": "ab_select_ready", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_basement_office_sel.png", "seed_asset_id": "ca2e4a9c-9e0a-4154-b333-889e1c877528", "seed_status": "ok", "bg_source": "plate", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L46B01.png", "bg_reason": "", "conti_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S34sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L46B01.png" }, "S51sh2": { "lane": "structure_plate", "group_id": "bg_basement_office", "segment_id": "SEG_STAIR_THRESHOLD", "status": "ab_select_bypass", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_basement_office_sel.png", "seed_asset_id": "ca2e4a9c-9e0a-4154-b333-889e1c877528", "seed_status": "ok", "bg_source": "plate", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L46B02.png", "bg_reason": "", "bypass_reason": "bg_only" }, "S55sh1": { "lane": "structure_plate", "group_id": "bg_basement_office", "segment_id": "SEG_WINDOW_FRONTAGE", "status": "ab_select_ready", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_basement_office_sel.png", "seed_asset_id": "ca2e4a9c-9e0a-4154-b333-889e1c877528", "seed_status": "ok", "bg_source": "plate", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L47B01.png", "bg_reason": "", "conti_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S55sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L47B01.png" }, "S75sh7": { "lane": "map_marker", "group_id": "bg_bus_stop", "segment_id": "SEG_02", "marker_attempts": 3, "status": "failed", "error": "marker map check 3회 소진: 지정된 마커 외에 요청하지 않은 추가 텍스트('(motionless young man)', '(departing city bus)')가 도면에 기재되어 있습니다.", "marker_check": { "markers_ok": false, "details_ko": "지정된 마커 외에 요청하지 않은 추가 텍스트('(motionless young man)', '(departing city bus)')가 도면에 기재되어 있습니다." } }, "S78sh2": { "lane": "structure_plate", "group_id": "bg_electronics_repair_shop", "segment_id": "SEG_02", "status": "ab_select_ready", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_electronics_repair_shop_sel.png", "seed_asset_id": "1b07e7d2-173a-4738-8d14-371b1463ea3e", "seed_status": "ok", "bg_source": "plate", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L07B01.png", "bg_reason": "", "conti_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S78sh2.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L07B01.png" }, "S78sh5": { "lane": "structure_plate", "group_id": "bg_electronics_repair_shop", "segment_id": "SEG_02", "status": "ab_select_bypass", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_electronics_repair_shop_sel.png", "seed_asset_id": "1b07e7d2-173a-4738-8d14-371b1463ea3e", "seed_status": "ok", "bg_source": "plate", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L07B01.png", "bg_reason": "", "bypass_reason": "prev" }, "S78sh7": { "lane": "structure_plate", "group_id": "bg_electronics_repair_shop", "segment_id": "SEG_02", "status": "ab_select_bypass", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_electronics_repair_shop_sel.png", "seed_asset_id": "1b07e7d2-173a-4738-8d14-371b1463ea3e", "seed_status": "ok", "bg_source": "plate", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L07B01.png", "bg_reason": "", "bypass_reason": "prev" }, "S89sh3": { "lane": "structure_plate", "group_id": "bg_small_theater", "segment_id": "SEG_01", "status": "ab_select_ready", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_small_theater_sel.png", "seed_asset_id": "f7d94261-acb1-4890-bd3a-9bf8bf558ca8", "seed_status": "ok", "bg_source": "plate", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L05B01.png", "bg_reason": "", "conti_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S89sh3.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L05B01.png" }, "S92sh1": { "lane": "structure_plate", "group_id": "bg_small_theater", "segment_id": "SEG_01", "status": "ab_select_ready", "seed_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/background_chain/seed_bg_small_theater_sel.png", "seed_asset_id": "f7d94261-acb1-4890-bd3a-9bf8bf558ca8", "seed_status": "ok", "bg_source": "plate", "bg_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L05B02.png", "bg_reason": "", "conti_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/images/7c902020-4451-4967-9eb6-1e53c2b9b717/conti/conti_S92sh1.png", "plate_path": "/Users/manta/Documents/Projects/TheRoad-I1/projects/c7e3b2e7-c545-4516-93b2-62a51a74d794/episodes/7c902020-4451-4967-9eb6-1e53c2b9b717/images/background_chain/L05B02.png" } } }, "step_id": "shot_conti_light", "run_id": "7bd50129-f04d-4d6a-a1ff-5bb9136da4f8", "resolved_model": "gpt-image-2", "updated_at": "2026-08-11T09:08:59.639287+00:00", "project_config_snapshot": {} }