Source study only: no video was generated in this round. This is the complete proposed shot-2 template, not an executed input. Picture 1 is the fixed v6 identity; Picture 2 supplies the existing room, outfit and notebook. Check sampling wiring before use.
AIMixer/ComfyUI_MiniMaxH3_Director
Shared prompt:
Shared reference and render instructions for each independently generated segment
Use <Picture 1> only to preserve Fang Zhiwei’s established adult face, facial proportions, base complexion, long dark wavy hair, and body proportions. Her 175 cm height is a character setting, not a measurement from the image.
Use <Picture 2> only for the existing room layout, daylight, notebook prop, and visible W066 clothing: an ivory-white short-sleeved collared button-front blouse with a chest pocket, pale-blue high-waisted loose trousers with turned-up cuffs, and beige flat sandals when visible. Do not use the woman in <Picture 2> as an identity, body, age, pose, or performance reference. Do not infer hidden garment construction or unseen details. The fixed v6 identity in <Picture 1> remains the sole face and body authority.
Render as a naturally photographed adult person in soft, scene-appropriate daylight. Keep skin and clothing colors lifelike, preserve natural hair and fabric texture, and integrate the subject with the room’s light and depth of field. Avoid waxy or porcelain skin, doll-like eyes, artificial whitening, heavy beauty smoothing, and global pink, magenta, or orange casts. Preserve the requested restrained performance and the visible room anchors. Use quiet in-room sound only: no dialogue, narration, or non-diegetic music.
The three shots are independent segment prompts intended to be edited together with hard cuts. Their requested durations are input targets, not measured outputs. No segment prompt claims that this plan has been generated or tested with ComfyUI_MiniMaxH3_Director.
Shot 2 / requested 5 seconds:
subject_definitions:
<Subject 1> is Fang Zhiwei, the adult woman whose face and body proportions are defined only by <Picture 1>.
<Subject 2> is the existing bright room in <Picture 2>, with a large window on the left, a long wooden table beneath it, a pale wall, a doorway on the right, and a wooden floor.
<Subject 3> is the visible W066 outfit from <Picture 2>: an ivory-white short-sleeved collared button-front blouse with a chest pocket, pale-blue high-waisted loose trousers with turned-up cuffs, and beige flat sandals when visible. This reference defines clothing only.
<Subject 4> is the single closed pale sage notebook shown on the floor in <Picture 2>; its position changes only through the visible action in this shot.
<Picture 1> is the fixed v6 identity reference for Fang Zhiwei’s face and body proportions.
<Picture 2> is a room, daylight, clothing, and prop reference only; the woman’s identity, body, age, and pose in it are not references.
summary:
[reference generation] A requested 5-second standalone locked wide shot of <Subject 1> picking up <Subject 4> with her right hand and placing it flat on <Subject 2>’s table. <Subject 3> follows <Picture 2> for clothing only.
retention_analysis:
<Subject 1> (appears in [Shot 1]): preserve the adult identity and proportions from <Picture 1>.
<Subject 2> (appears in [Shot 1]): retain the left window, wooden table, pale wall, right doorway, and daylight layout from <Picture 2>.
<Subject 3> (appears in [Shot 1]): retain the visible W066 clothing from <Picture 2>; do not infer hidden construction.
<Subject 4> (appears in [Shot 1]): transfer the same single notebook continuously from the floor to the tabletop; no duplication, disappearance, or teleportation.
detailed_description:
Requested segment duration: 5 seconds. [Shot 1] A locked wide shot shows <Subject 2> and enough floor, tabletop, and both of <Subject 1>’s hands to read the complete action. Start with the same closed pale sage notebook on the floor beside the table as in <Picture 2>. <Subject 1> wears <Subject 3>; her face, hair, age, and proportions follow <Picture 1> alone. The woman in <Picture 2> is not an identity or performance reference. Fang bends naturally, visibly contacts the notebook with her right hand, lifts that same notebook from the floor, straightens, and places it flat on the tabletop to the right of the lamp. Keep the notebook visible during the transfer and preserve believable hand contact. She releases it, withdraws her right hand, and remains upright beside the table. End with the notebook supported flat by the tabletop, her hand clear of it, and no notebook left on the floor. Keep the camera fixed and the room landmarks stable. This standalone segment ends cleanly for an editorial hard cut.
overall_soundscape:
Quiet indoor room tone with soft cloth movement, light footsteps, and a faint notebook contact against the floor and tabletop. No voice; no sound is copied from a reference.
non_diegetic_music:
N/A. No background music.
58 views
本篇是源码研究，没有本轮实际生成输入。下面是第二镜的完整待测模板（尚未提交），使用前先核对采样接线。Picture 1为固定v6身份；Picture 2为已有房间、衣物和笔记本参考。
AIMixer/ComfyUI_MiniMaxH3_Director
Shared prompt:
Shared reference and render instructions for each independently generated segment
Use <Picture 1> only to preserve Fang Zhiwei’s established adult face, facial proportions, base complexion, long dark wavy hair, and body proportions. Her 175 cm height is a character setting, not a measurement from the image.
Use <Picture 2> only for the existing room layout, daylight, notebook prop, and visible W066 clothing: an ivory-white short-sleeved collared button-front blouse with a chest pocket, pale-blue high-waisted loose trousers with turned-up cuffs, and beige flat sandals when visible. Do not use the woman in <Picture 2> as an identity, body, age, pose, or performance reference. Do not infer hidden garment construction or unseen details. The fixed v6 identity in <Picture 1> remains the sole face and body authority.
Render as a naturally photographed adult person in soft, scene-appropriate daylight. Keep skin and clothing colors lifelike, preserve natural hair and fabric texture, and integrate the subject with the room’s light and depth of field. Avoid waxy or porcelain skin, doll-like eyes, artificial whitening, heavy beauty smoothing, and global pink, magenta, or orange casts. Preserve the requested restrained performance and the visible room anchors. Use quiet in-room sound only: no dialogue, narration, or non-diegetic music.
The three shots are independent segment prompts intended to be edited together with hard cuts. Their requested durations are input targets, not measured outputs. No segment prompt claims that this plan has been generated or tested with ComfyUI_MiniMaxH3_Director.
Shot 2 / requested 5 seconds:
subject_definitions:
<Subject 1> is Fang Zhiwei, the adult woman whose face and body proportions are defined only by <Picture 1>.
<Subject 2> is the existing bright room in <Picture 2>, with a large window on the left, a long wooden table beneath it, a pale wall, a doorway on the right, and a wooden floor.
<Subject 3> is the visible W066 outfit from <Picture 2>: an ivory-white short-sleeved collared button-front blouse with a chest pocket, pale-blue high-waisted loose trousers with turned-up cuffs, and beige flat sandals when visible. This reference defines clothing only.
<Subject 4> is the single closed pale sage notebook shown on the floor in <Picture 2>; its position changes only through the visible action in this shot.
<Picture 1> is the fixed v6 identity reference for Fang Zhiwei’s face and body proportions.
<Picture 2> is a room, daylight, clothing, and prop reference only; the woman’s identity, body, age, and pose in it are not references.
summary:
[reference generation] A requested 5-second standalone locked wide shot of <Subject 1> picking up <Subject 4> with her right hand and placing it flat on <Subject 2>’s table. <Subject 3> follows <Picture 2> for clothing only.
retention_analysis:
<Subject 1> (appears in [Shot 1]): preserve the adult identity and proportions from <Picture 1>.
<Subject 2> (appears in [Shot 1]): retain the left window, wooden table, pale wall, right doorway, and daylight layout from <Picture 2>.
<Subject 3> (appears in [Shot 1]): retain the visible W066 clothing from <Picture 2>; do not infer hidden construction.
<Subject 4> (appears in [Shot 1]): transfer the same single notebook continuously from the floor to the tabletop; no duplication, disappearance, or teleportation.
detailed_description:
Requested segment duration: 5 seconds. [Shot 1] A locked wide shot shows <Subject 2> and enough floor, tabletop, and both of <Subject 1>’s hands to read the complete action. Start with the same closed pale sage notebook on the floor beside the table as in <Picture 2>. <Subject 1> wears <Subject 3>; her face, hair, age, and proportions follow <Picture 1> alone. The woman in <Picture 2> is not an identity or performance reference. Fang bends naturally, visibly contacts the notebook with her right hand, lifts that same notebook from the floor, straightens, and places it flat on the tabletop to the right of the lamp. Keep the notebook visible during the transfer and preserve believable hand contact. She releases it, withdraws her right hand, and remains upright beside the table. End with the notebook supported flat by the tabletop, her hand clear of it, and no notebook left on the floor. Keep the camera fixed and the room landmarks stable. This standalone segment ends cleanly for an editorial hard cut.
overall_soundscape:
Quiet indoor room tone with soft cloth movement, light footsteps, and a faint notebook contact against the floor and tabletop. No voice; no sound is copied from a reference.
non_diegetic_music:
N/A. No background music.
1 like · 50 views
Mentions
Source study only: no video was generated in this round. This is the complete proposed shot-2 template, not an executed input. Picture 1 is the fixed v6 identity; Picture 2 supplies the existing room, outfit and notebook. Check sampling wiring before use. AIMixer/ComfyUI_MiniMaxH3_Director Shared prompt: Shared reference and render instructions for each independently generated segment Use <Picture 1> only to preserve Fang Zhiwei’s established adult face, facial proportions, base complexion, long dark wavy hair, and body proportions. Her 175 cm height is a character setting, not a measurement from the image. Use <Picture 2> only for the existing room layout, daylight, notebook prop, and visible W066 clothing: an ivory-white short-sleeved collared button-front blouse with a chest pocket, pale-blue high-waisted loose trousers with turned-up cuffs, and beige flat sandals when visible. Do not use the woman in <Picture 2> as an identity, body, age, pose, or performance reference. Do not infer hidden garment construction or unseen details. The fixed v6 identity in <Picture 1> remains the sole face and body authority. Render as a naturally photographed adult person in soft, scene-appropriate daylight. Keep skin and clothing colors lifelike, preserve natural hair and fabric texture, and integrate the subject with the room’s light and depth of field. Avoid waxy or porcelain skin, doll-like eyes, artificial whitening, heavy beauty smoothing, and global pink, magenta, or orange casts. Preserve the requested restrained performance and the visible room anchors. Use quiet in-room sound only: no dialogue, narration, or non-diegetic music. The three shots are independent segment prompts intended to be edited together with hard cuts. Their requested durations are input targets, not measured outputs. No segment prompt claims that this plan has been generated or tested with ComfyUI_MiniMaxH3_Director. Shot 2 / requested 5 seconds: subject_definitions: <Subject 1> is Fang Zhiwei, the adult woman whose face and body proportions are defined only by <Picture 1>. <Subject 2> is the existing bright room in <Picture 2>, with a large window on the left, a long wooden table beneath it, a pale wall, a doorway on the right, and a wooden floor. <Subject 3> is the visible W066 outfit from <Picture 2>: an ivory-white short-sleeved collared button-front blouse with a chest pocket, pale-blue high-waisted loose trousers with turned-up cuffs, and beige flat sandals when visible. This reference defines clothing only. <Subject 4> is the single closed pale sage notebook shown on the floor in <Picture 2>; its position changes only through the visible action in this shot. <Picture 1> is the fixed v6 identity reference for Fang Zhiwei’s face and body proportions. <Picture 2> is a room, daylight, clothing, and prop reference only; the woman’s identity, body, age, and pose in it are not references. summary: [reference generation] A requested 5-second standalone locked wide shot of <Subject 1> picking up <Subject 4> with her right hand and placing it flat on <Subject 2>’s table. <Subject 3> follows <Picture 2> for clothing only. retention_analysis: <Subject 1> (appears in [Shot 1]): preserve the adult identity and proportions from <Picture 1>. <Subject 2> (appears in [Shot 1]): retain the left window, wooden table, pale wall, right doorway, and daylight layout from <Picture 2>. <Subject 3> (appears in [Shot 1]): retain the visible W066 clothing from <Picture 2>; do not infer hidden construction. <Subject 4> (appears in [Shot 1]): transfer the same single notebook continuously from the floor to the tabletop; no duplication, disappearance, or teleportation. detailed_description: Requested segment duration: 5 seconds. [Shot 1] A locked wide shot shows <Subject 2> and enough floor, tabletop, and both of <Subject 1>’s hands to read the complete action. Start with the same closed pale sage notebook on the floor beside the table as in <Picture 2>. <Subject 1> wears <Subject 3>; her face, hair, age, and proportions follow <Picture 1> alone. The woman in <Picture 2> is not an identity or performance reference. Fang bends naturally, visibly contacts the notebook with her right hand, lifts that same notebook from the floor, straightens, and places it flat on the tabletop to the right of the lamp. Keep the notebook visible during the transfer and preserve believable hand contact. She releases it, withdraws her right hand, and remains upright beside the table. End with the notebook supported flat by the tabletop, her hand clear of it, and no notebook left on the floor. Keep the camera fixed and the room landmarks stable. This standalone segment ends cleanly for an editorial hard cut. overall_soundscape: Quiet indoor room tone with soft cloth movement, light footsteps, and a faint notebook contact against the floor and tabletop. No voice; no sound is copied from a reference. non_diegetic_music: N/A. No background music.
本篇是源码研究，没有本轮实际生成输入。下面是第二镜的完整待测模板（尚未提交），使用前先核对采样接线。Picture 1为固定v6身份；Picture 2为已有房间、衣物和笔记本参考。 AIMixer/ComfyUI_MiniMaxH3_Director Shared prompt: Shared reference and render instructions for each independently generated segment Use <Picture 1> only to preserve Fang Zhiwei’s established adult face, facial proportions, base complexion, long dark wavy hair, and body proportions. Her 175 cm height is a character setting, not a measurement from the image. Use <Picture 2> only for the existing room layout, daylight, notebook prop, and visible W066 clothing: an ivory-white short-sleeved collared button-front blouse with a chest pocket, pale-blue high-waisted loose trousers with turned-up cuffs, and beige flat sandals when visible. Do not use the woman in <Picture 2> as an identity, body, age, pose, or performance reference. Do not infer hidden garment construction or unseen details. The fixed v6 identity in <Picture 1> remains the sole face and body authority. Render as a naturally photographed adult person in soft, scene-appropriate daylight. Keep skin and clothing colors lifelike, preserve natural hair and fabric texture, and integrate the subject with the room’s light and depth of field. Avoid waxy or porcelain skin, doll-like eyes, artificial whitening, heavy beauty smoothing, and global pink, magenta, or orange casts. Preserve the requested restrained performance and the visible room anchors. Use quiet in-room sound only: no dialogue, narration, or non-diegetic music. The three shots are independent segment prompts intended to be edited together with hard cuts. Their requested durations are input targets, not measured outputs. No segment prompt claims that this plan has been generated or tested with ComfyUI_MiniMaxH3_Director. Shot 2 / requested 5 seconds: subject_definitions: <Subject 1> is Fang Zhiwei, the adult woman whose face and body proportions are defined only by <Picture 1>. <Subject 2> is the existing bright room in <Picture 2>, with a large window on the left, a long wooden table beneath it, a pale wall, a doorway on the right, and a wooden floor. <Subject 3> is the visible W066 outfit from <Picture 2>: an ivory-white short-sleeved collared button-front blouse with a chest pocket, pale-blue high-waisted loose trousers with turned-up cuffs, and beige flat sandals when visible. This reference defines clothing only. <Subject 4> is the single closed pale sage notebook shown on the floor in <Picture 2>; its position changes only through the visible action in this shot. <Picture 1> is the fixed v6 identity reference for Fang Zhiwei’s face and body proportions. <Picture 2> is a room, daylight, clothing, and prop reference only; the woman’s identity, body, age, and pose in it are not references. summary: [reference generation] A requested 5-second standalone locked wide shot of <Subject 1> picking up <Subject 4> with her right hand and placing it flat on <Subject 2>’s table. <Subject 3> follows <Picture 2> for clothing only. retention_analysis: <Subject 1> (appears in [Shot 1]): preserve the adult identity and proportions from <Picture 1>. <Subject 2> (appears in [Shot 1]): retain the left window, wooden table, pale wall, right doorway, and daylight layout from <Picture 2>. <Subject 3> (appears in [Shot 1]): retain the visible W066 clothing from <Picture 2>; do not infer hidden construction. <Subject 4> (appears in [Shot 1]): transfer the same single notebook continuously from the floor to the tabletop; no duplication, disappearance, or teleportation. detailed_description: Requested segment duration: 5 seconds. [Shot 1] A locked wide shot shows <Subject 2> and enough floor, tabletop, and both of <Subject 1>’s hands to read the complete action. Start with the same closed pale sage notebook on the floor beside the table as in <Picture 2>. <Subject 1> wears <Subject 3>; her face, hair, age, and proportions follow <Picture 1> alone. The woman in <Picture 2> is not an identity or performance reference. Fang bends naturally, visibly contacts the notebook with her right hand, lifts that same notebook from the floor, straightens, and places it flat on the tabletop to the right of the lamp. Keep the notebook visible during the transfer and preserve believable hand contact. She releases it, withdraws her right hand, and remains upright beside the table. End with the notebook supported flat by the tabletop, her hand clear of it, and no notebook left on the floor. Keep the camera fixed and the room landmarks stable. This standalone segment ends cleanly for an editorial hard cut. overall_soundscape: Quiet indoor room tone with soft cloth movement, light footsteps, and a faint notebook contact against the floor and tabletop. No voice; no sound is copied from a reference. non_diegetic_music: N/A. No background music.