You are looking at a top-down floor-plan diagram of an enclosed,
control-heavy interior (a cabin, cockpit, wheelhouse, crew bay — as the
texts below describe). FIRST read it literally: which station each
primary control is attached to, where any mirrors/reflective surfaces
are and which way their faces point, where the camera stands and which
way it looks, and which seats/stations are marked as occupied. Report
those in `reads`.

THEN compare your reading against the LOCATION / CAMERA / SHOT TEXT
below. List every disagreement in `mismatches` (empty if none).

FINALLY, write `scene_description_en`: 4-7 sentences describing exactly
what this camera sees, derived ONLY from the diagram — for each element
(each person, each primary control, each mirror/reflective surface,
windows/openings), say WHERE it appears on screen (left / right /
center, near / far) and which way it faces. Name who occupies which
seat/station. Be explicit about any mirror: where it sits on screen and
what its face can physically reflect from this camera. This text will
drive image generation, so every left/right must be from the CAMERA's
point of view.