subject_definitions:
<Video 1> is the source reference video for the target video edit. It supplies the original dance performance, body movement, pose timing, stance changes, arm gestures, the beat where the two pistols are drawn, the camera framing and the shot rhythm, but it supplies no identity appearance and no background.
<Subject 1> is the original performer in <Video 1>: a smooth featureless red humanoid mannequin with a glossy jointed body, an egg-shaped head with no facial features, bare feet, and two small silver pistols that it draws from its waist. <Subject 1> is a source entity that must be completely replaced; its red mannequin surface, its featureless head, its visible joint lines and its pistols must not survive in the target video.
<Picture 1> is the three-view character turnaround sheet of the target character <Subject 2>: one single young woman shown as a front view, a side view and a back view on a plain light grey background.
<Subject 2> is the target character whose identity and complete visual appearance come from <Picture 1>: a young Chinese woman with long straight black hair, two high buns decorated with pink flowers and long trailing ribbons, wearing a light mint-green and white short hanfu-style dress with wide short sleeves over a white underlayer, a golden phoenix waist ornament with a green sash and a white bow at the front, white wrist wraps, white opaque tights and black platform Mary-Jane shoes.
<Subject 3> is the new environment for the whole shot: an ancient Chinese courtyard at golden hour - grey stone paving, a carved wooden gallery with dark red lacquered pillars and hanging bamboo blinds on the left, potted peonies and a stone lantern along the right wall, a low stone balustrade, and warm low sunlight with a few drifting petals. The courtyard is empty of other people, clean and evenly lit.

summary:
[video editing + character replacement] three-view identity reference
The target video is an edited version of <Video 1>. Completely replace <Subject 1> with <Subject 2>, whose identity and complete visual appearance are defined by <Picture 1>. Replace the flat green chroma-key background of <Video 1> with the ancient courtyard of <Subject 3>. Preserve the dance performance, the pose timing, the body rhythm and the camera work of <Video 1> one-to-one for the whole 5.17 seconds. The red mannequin body and the two pistols must never appear; the target character performs the same dance with her empty hands.

retention_analysis:
<Video 1>: fully_preserved for body movement, dance poses, action timing, stance and weight shifts, the beat where the pistols are drawn, camera framing and shot rhythm; the glossy red mannequin identity, the two silver pistols and the flat green chroma-key backdrop are explicitly excluded and must never be reproduced.
<Subject 1>: identity_not_preserved. Completely remove the glossy red mannequin surface, the featureless egg-shaped head, the visible joint lines, the bare feet and the two silver pistols.
<Picture 1>: fully_referenced as the identity and complete visual-appearance source for <Subject 2>; the three views are one single person, never three people.
<Subject 2>: identity_fully_preserved. Keep her face, long black hair, two high buns, pink flower ornaments, mint-green and white hanfu dress, golden waist ornament, green sash with white bow, white wrist wraps, white tights and black shoes consistent from every visible angle in every frame.
<Subject 3>: fully_preserved as the only environment; the courtyard, its light and its drifting petals stay stable in every frame.

reference_interpretation:
<Picture 1> is a three-view character sheet and the three figures inside it are the front view, the side view and the back view of the same single woman, not three people and not a group. The front view defines her face and the front of the dress, the side view defines her profile, her hair buns and the dress silhouette, and the back view defines the back of her hair and the back of the dress. All three views describe the same body, the same hairstyle and the same costume, so the target video must contain exactly one such woman.

detailed_description:
[Shot 1] Photorealistic cinematic footage of a real young woman, vertical 9:16, one unbroken take for the full 5.17 seconds. Medium-full shot at chest height, the dancer centred and framed from head to feet exactly as in <Video 1>, standing on the stone paving of the ancient courtyard of <Subject 3>. The camera is almost static with only a very slight slow drift; the sunlight stays warm and low, and a few petals drift through the frame.
00:00.000-00:01.200: At 0.00 seconds <Subject 2> is already fully present in the courtyard, standing in a wide stance facing the camera, both knees bent outward, the left arm bent and raised with the hand near her shoulder and the right arm extended out to the side, the torso twisted a little; she sways once on the beat, shoulders rolling, her long hair and ribbons swinging.
00:01.200-00:02.600: she shifts her weight from one leg to the other, the raised left hand curls inward, the right arm folds down and comes back up, the hips move in a slow figure; her gaze stays on the camera and her expression is calm and a little playful.
00:02.600-00:03.800: she brings her right hand up to her temple in a salute-like gesture and holds it while the left arm stretches out to the side, then dips lower into a deeper stance, the chest turning slightly away from the camera and back again.
00:03.800-00:05.167: she reaches down to her waist and lifts both hands up in front of her chest in the same double-pistol gesture as <Video 1>, but with empty hands; she holds the crouched dance pose with one heel lifting off the ground, then settles with a small bounce as the shot ends.
She is alone in the frame from the first frame to the last; nothing enters or leaves the shot. No green screen, no mannequin, no visible joints, no pistol, no weapon, no text, no subtitle, no watermark, no logo and no UI overlay anywhere in the frame. No camera shake, no jump cut, no extra person, no duplicated person.

overall_soundscape:
A steady mid-tempo electronic dance beat with a soft kick on every beat and a light synth pulse, consistent from start to finish, over a faint outdoor courtyard ambience of light wind and rustling leaves. No dialogue and no footsteps.

non_diegetic_music:
The same steady electronic dance beat continues as the non-diegetic score at a constant tempo, with no change of instrumentation from the beginning to the end.