Galerie de prompts IA OpenZhenZhen · MiniMax H3 · video

Répétition d'attaque et défense à cinq au dojo, sans sang

Five-Person Dojo Non-Bloody Rehearsal Attack and Defense

Répétition d'attaque et défense à cinq au dojo, sans sang — découvrez cette création avec MiniMax H3 sur OpenZhenZhen, son prompt original et un exemple de génération à réutiliser.

Répétition d'attaque et défense à cinq au dojo, sans sang — AI video creation
Répétition d'attaque et défense à cinq au dojo, sans sang

Prompt

Prompt chinois

0.00秒完全参考 <Picture 1> [Shot 1]首帧。原图明亮道场的真人非血腥排练格斗。A为中央棕发女子、黑色无袖长上衣与黑裤；B为前左红衣男子；C为前右白衣马尾女子；D为后左绿衣胡须男子；E为后右紫衣男子。恰好五位不同成年角色，空手。角落地标和房间几何固定，短而可读路径，全身三分之四机位。 [Shot 1] 0–6秒：仅B对A出受控排练拳，A侧闪，C/D/E各向内一步但不攻击。仅C对A做受控腿攻，A蹲避，B停在A后、不占C位置。A起身可见掌触D肩，只有D向绿色地标退一步。 [Shot 2] 6.000秒在完整事件边界切同轴一侧连续匹配中广景，无时间跳跃。6–12秒：E握A前臂，A全身旋转解握，接触释放始终可见。A穿过B/C真实空隙，B/C转身跟随，D/E在另一侧绕行，仍同五人重新分布。B再推进，A/C沿分开短路径交叉，B步入C刚离开的位置。 [Shot 3] 12.000秒切更广连续空间关系。A受控推E肩一次，E退步，B/C/D分别停止，使五人构成互不重叠半圆。14–15秒保持不同最终姿势与持有状态。面孔服装布景一致，关键接触交接落地连续显示、不藏于切镜。实时物理编排，无慢动作填充、魔法闪光、能量束、字幕、额外演员或复制道具。 声音为安静道场的清晰脚步、衣料、呼气和轻排练肩部接触；无对白、歌唱、夸张冲击轰响或武器声。配乐：无。

Prompt anglais

For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced. integrated_multimodal_description: [Shot 1] Live-action choreographed, non-graphic rehearsal combat in the bright dojo from <Picture 1>. A is the central brown-haired woman in a black sleeveless tunic and black trousers. B is the red-clothed man at front left, C the white-clothed ponytailed woman at front right, D the green-clothed bearded man behind left, and E the purple-clothed man behind right. All five are distinct adults with empty hands. Keep the corner floor markers and room geometry fixed; use short, readable paths and a full-body three-quarter camera. From 0 to 6 seconds: B alone makes a controlled rehearsal punch toward A; A sidesteps it, while C, D and E each take one inward step without attacking. C alone makes a controlled leg attack toward A; A crouches below it, and B stops behind A without taking C's place. A rises and visibly palms D's shoulder; only D recoils one step toward the green floor marker. [Shot 2] At 00:06.000, cut at the completed event boundary to a medium-wide continuity-matched view on the same side of the action axis, without a time jump. From 6 to 12 seconds: E takes A's forearm, and A rotates her whole body to release that grip, with the contact and release continuously visible. A threads the real gap between B and C as they turn to follow; D and E circle on the other side, redistributing the same five people. B advances again; A and C cross on separate short paths, and B steps into the position C has just vacated. [Shot 3] At 00:12.000, cut to a wider continuity-matched view of the final spatial relationship. A gives E one controlled shoulder push; E steps back, while B, C and D stop separately, leaving all five in a non-overlapping half-circle. From 14 to 15 seconds, hold the distinct final poses and ownership states. Keep faces, clothing and fixed scenery consistent. Show every critical contact, transfer and landing continuously rather than hiding it across a cut. Real-time physical choreography, no slow-motion filler, magical flashes, energy beams, captions, extra actors or duplicated props. overall_soundscape: Clear footfalls, clothing movement, exhalations and soft staged shoulder contacts in a quiet dojo. No dialogue, singing, exaggerated impact booms or weapon sounds. non_diegetic_music: N/A

Thèmes et styles

Voir le modèle API

Découvrir plus de prompts image et vidéo · Générer des images · Générer des vidéos · Créer sur le canevas IA