subject_definitions:
<Subject 1> is Female Lead 1, the young adult Chinese woman whose facial identity comes from the frontal portrait in <Picture 3>. Preserve her softly tapered oval face, wide dark-brown almond eyes, delicate straight nose, full rose-colored lips, center-parted black hair, and elaborate antique-gold floral headpiece with red, amber, and blue hanging beads. Her pale taupe-grey silk layers and dense antique-gold embroidery come from <Picture 1> and <Picture 2>. She has no beauty mark beneath either eye.
<Subject 2> is Female Lead 2, the adult Chinese woman whose facial identity comes from the frontal portrait in <Picture 4>. Preserve her longer angular oval face, defined cheekbones, narrow intense dark eyes, straight brows, small beauty mark beneath her left eye, black hair pulled back beneath a dark antique-gold headpiece, long blue-red bead strands, and layered charcoal blue-grey embroidered wraps. Her on-set costume, pose, and position follow <Picture 1>.
<Picture 1> is the opening storyboard composition for [Shot 1], showing <Subject 1> in the screen-left foreground and <Subject 2> in the screen-right midground on the outdoor snowy woodland path.
<Picture 2> is the target composition for the end of [Shot 1], defining <Subject 1>'s close facial scale, head angle, rightward eyeline, framing, and depth of field.

summary:
[keyframe completion + reference generation] Generate one continuous 5.0-second silent camera move. Begin from the two-woman spatial relationship of <Picture 1>, truck right and push closer around <Subject 1>, allow <Subject 2> to leave the narrowing frame through camera parallax rather than physical movement, and settle on the close composition of <Picture 2>. Preserve both identities, costumes, the outdoor snowy setting, and the original storyboard texture. There is no dialogue, ambience, sound effect, or music.

retention_analysis:
<Subject 1> (appears throughout [Shot 1]): fully_preserved - retain her exact referenced face, no beauty mark, pale taupe-grey and antique-gold costume, red-gold bead accents, hairstyle, adult proportions, and controlled expression as the camera changes distance and viewing angle.
<Subject 2> (appears during the opening portion of [Shot 1]): fully_preserved - retain her sharper face, left-eye beauty mark, charcoal blue-grey costume, blue-red bead accents, stationary position, and guarded expression until the moving camera naturally excludes her.
<Picture 1> ([Shot 1] opening composition): fully_preserved - retain the initial left-foreground/right-midground blocking, snowy path, trees, rocks, winter daylight, and shared eyeline.
<Picture 2> ([Shot 1] ending composition): fully_preserved - use its facial scale, head angle, rightward gaze, costume crop, soft background and close framing as the camera's final destination.

detailed_description:
Preserve the exact live-action appearance of the storyboard images—cool winter shadows, natural softly lit skin, fine hair, and believable antique-gold embroidery—without redesigning or intensifying the visual style.

Use exactly four reference images: <Picture 1> controls the opening composition, <Picture 2> controls the ending composition, <Picture 3> fixes <Subject 1>'s facial identity, and <Picture 4> fixes <Subject 2>'s facial identity. The two frontal portraits are identity references only; their frontal poses and backgrounds must not replace either storyboard composition, become inserted shots, or create additional people.

[Shot 1] Begin from the spatial arrangement of <Picture 1>. <Subject 1> occupies the screen-left foreground in pale taupe-grey and antique-gold layers, seen from behind and in restrained three-quarter profile. <Subject 2> stands several steps away in the screen-right midground in charcoal blue-grey embroidered layers, facing <Subject 1>. The snowy forest path, irregular rocks, pine trunks, falling flakes, exposure, and depth remain fixed. Both women stay physically still. Their swords remain fully sheathed and untouched. Their mouths remain closed, and neither performs a fight pose, threat gesture, step, turn, or costume movement beyond minute natural breathing and a few visually moving loose hair strands.

From 0.00 to 0.55 seconds, hold the opening relationship long enough for the viewer to read the distance and opposing eyelines. From 0.55 to 3.90 seconds, the camera trucks to its own right at slow, even speed while simultaneously pushing in toward <Subject 1>. Add only a subtle rightward arc around her face, just enough to reveal more of her cheek, eye, and frontal structure. The camera movement creates natural foreground-to-background parallax: <Subject 2> appears to drift toward the right edge solely because the frame narrows, but she does not walk or change her stance. Keep the camera on the same side of the axis and preserve the two women's established left-right relationship.

From 3.90 to 5.00 seconds, decelerate smoothly. <Subject 2> leaves the image through the right edge as a consequence of the camera position. The camera finishes on <Subject 1> alone, matching <Picture 2>'s face size, head angle, rightward eyeline, shoulder crop, shallow depth of field, lighting and ornament placement as closely as possible. Do not dissolve, cut, morph, swap faces, change costume color, or turn <Picture 2> into a sudden still photograph. End with <Subject 1>'s lips fully closed and gaze held toward the unseen <Subject 2>, ready for the next independently generated clip.

overall_soundscape:
Absolute digital silence for the full 5.0 seconds. No dialogue, breathing, wind, snow, footsteps, fabric, beads, weapon sound, room tone, environmental ambience, or other sound effect is audible.

non_diegetic_music:
N/A
