subject_definitions:
<Subject 1> is one coherent adult character assembled from the ordered reference pictures. <Picture 1> defines the character's identity, face, hairstyle, body proportions, established body coverage, and overall visual style while contributing no background or pose. <Picture 2> supplies complementary front and side evidence for anatomy, proportions, colors, accessories, and surface detail while contributing no background or pose. <Picture 3> supplies complementary rear and facial evidence while contributing no background or pose. Preserve one stable identity, one stable anatomy, one stable set of proportions, and the same established coverage across the complete target.
<Picture 1> is a native Qwen+VAE reference. Use its detailed visual appearance only for the assigned role; it is not a frame or composition anchor.
<Picture 2> is a native Qwen+VAE reference. Use its detailed visual appearance only for the assigned role; it is not a frame or composition anchor.
<Picture 3> is a native Qwen+VAE reference. Use its detailed visual appearance only for the assigned role; it is not a frame or composition anchor.

summary:
[reference generation] The 171-frame target preserves the canonical four-view full-body orbit, inserts an approximately two-second waist-to-knee phase during one uninterrupted clockwise detail move, then presents the original facial sequence shifted later.

retention_analysis:
<Subject 1> (appears throughout the target video): fully_preserved - preserve the requested identity, facial structure, anatomy, hairstyle, body proportions, surface detail, colors, accessories, and established body coverage without drift between captures; exclude every source background, source pose, and unrequested person.

detailed_description:
The target uses the requested visual style with sharp anatomical and facial detail against a solid uniform light-gray seamless backdrop under soft, even, color-neutral form lighting. <Subject 1> stands centered in a relaxed neutral A-pose, arms slightly clear of the torso, palms toward the thighs, head level, eyes open, and expression neutral. The character remains rigidly motionless. Hair, surface details, accessories, and all established body coverage remain locked in exactly the same configuration; do not introduce nudity or explicit anatomical exposure. Camera roll, exposure, focus, lighting, and subject proportions remain fixed. There is no body motion, breathing, blinking, wind, secondary motion, camera shake, or motion blur. Use one continuous untagged timeline.

Begin with exactly the canonical six-view full-body phase, unchanged. From 00:00.000 to 00:03.000, use a fixed long-telephoto near-orthographic full-body composition while the camera performs one smooth constant-speed 360-degree clockwise orbit, beginning square front, passing the left profile at one quarter turn, reaching square back at halfway, passing the right profile at three quarters, and returning square front. Keep framing, scale, lighting, proportions, and surface details identical throughout this phase. The decoder preserves the original canonical body captures at frames 2, 21, 42, and 63.

At 00:03.000, rapidly push in and lower the optical target to a waist-to-knee composition whose upper edge sits slightly above the waist and whose lower edge sits immediately below both knees. The camera never stops during this detail orbit. Continue rotating clockwise without any pause or capture hold. Pass the exact left-profile waist-to-knee view at 00:03.500, decoder frame 84, and the exact back-right three-quarter waist-to-knee view at 00:04.500, decoder frame 108. Keep both moving pass-by frames crisp with no motion blur; changing background parallax must prove the camera is still rotating. Complete this inserted detail phase square front by 00:04.958.

Immediately redirect upward and rapidly push into the original facial sequence shifted later by the insert: a locked square-front head-and-shoulders view centered at 00:05.458, decoder frame 131, followed by a quick reposition to a clean front-left three-quarter head-and-shoulders view centered at 00:06.667, decoder frame 160. Retain the final view unchanged through 00:07.083, the last frame. Every transition is a deliberate physical camera move, never a morph, and every extracted capture depicts the same unchanged person.

overall_soundscape:
Silence; no dialogue, ambience, Foley, room tone, or movement sounds.

non_diegetic_music:
N/A
