Generate a 10-second ultra-photorealistic cinematic action video in vertical 9:16 format, resembling real live-action footage shot on a professional cinema camera.
SCENE: A powerful adult woman in practical black tactical clothing is trapped in a rain-soaked modern city street at dusk. A large crowd of adult attackers rushes toward her from multiple directions. She remains calm, intelligent, fearless, and highly agile.
0–3 seconds: Wide establishing shot. The crowd charges toward her through light rain. She studies their movements, takes a defensive stance, and dodges the first attacker with realistic timing.
3–7 seconds: Smooth handheld tracking shot. She performs believable martial arts movements, blocks attacks, sidesteps opponents, and uses controlled non-graphic takedowns. Every movement has realistic human weight, accurate body mechanics, natural foot contact, and consistent anatomy. The crowd moves independently, with no duplicated bodies or synchronized robotic movements.
7–10 seconds: She breaks through the crowd and reaches a brightly illuminated building entrance. The remaining attackers stop behind her. She turns toward the camera, breathes naturally, and raises one hand in a confident victory gesture. End on a sharp, powerful cinematic close-up.
VISUAL QUALITY: Photorealistic human faces, natural skin pores, realistic wet fabric, detailed eyes, physically accurate rain and reflections, balanced HDR lighting, sharp focus, natural motion blur, cinematic contrast, realistic depth of field, 4K-style detail, stable character identity, smooth continuous motion, professional action-film cinematography.
AUDIO: Generate original synchronized environmental audio: natural footsteps on wet pavement, realistic rain, clothing movement, subtle breathing, believable impacts, and distant crowd shouts. Include one original female dialogue line, spoken clearly and confidently: “You can stop me, but you can never break me!” Accurate lip synchronization, natural human voice, clean dialogue, balanced sound levels, no clipping, no artificial robotic voice, no background music overpowering the action.
STRICT NEGATIVE CONSTRAINTS: No animation, no cartoon appearance, no CGI-looking skin, no extra fingers, no deformed limbs, no floating bodies, no unnatural fighting, no teleportation, no flickering, no jitter, no excessive camera shake, no visual noise, no compression artifacts, no distorted faces, no subtitles, no text, no logos, no watermark, no gore.
Deliver a coherent 10-second live-action-style video with clean visuals, realistic movement, original synchronized audio, and premium cinematic quality.
生成一段10秒的超写实电影动作视频,采用竖屏9:16格式,呈现出类似专业电影摄像机拍摄的真实实拍画面质感。
场景:一名强壮的成年女性身穿实用的黑色战术服装,被困在黄昏时分被雨水浸透的现代城市街道上。一大群成年袭击者从多个方向向她冲来。她保持冷静、机智、无畏且身手敏捷。
0–3秒:广角建立镜头。人群穿过小雨向她冲来。她观察他们的动作,摆出防御姿态,并以真实的时机躲避第一个袭击者。
3–7秒:流畅的手持跟拍镜头。她执行可信的武术动作,格挡攻击,侧身避开对手,并使用受控的非血腥摔投技术。每个动作都具有真实的人体重量感、准确的身体力学、自然的脚部接触以及一致的解剖结构。人群独立移动,没有重复的身体或同步的机械式动作。
7–10秒:她突破人群,到达一个灯火通明的建筑物入口。剩余的袭击者在她身后停下。她转向镜头,自然地呼吸,并举起一只手做出自信的胜利手势。以锐利有力的电影特写结束。
视觉质量:照片级逼真的人脸,自然的皮肤毛孔,逼真的湿布料,细节丰富的眼睛,物理准确的雨水和反射,平衡的HDR光照,清晰的对焦,自然的运动模糊,电影般的对比度,真实的景深,4K级别的细节,稳定的角色身份,平滑连续的运动,专业的动作电影摄影。
音频:生成原创同步环境音频:湿滑路面上的自然脚步声,逼真的雨声,衣物摩擦声,微弱的呼吸声,可信的撞击声,以及远处人群的喊叫声。包含一句原创女性台词,清晰自信地说:“你可以阻止我,但你永远无法摧毁我!”准确的唇形同步,自然的人声,干净的对话,平衡的音量水平,无削波失真,无人工机器人声音,无背景音乐盖过动作音效。
严格负面约束:无动画,无卡通外观,无CGI感的皮肤,无多余手指,无畸形肢体,无漂浮身体,无不自然的打斗,无瞬移,无闪烁,无抖动,无过度相机晃动,无视觉噪点,无压缩伪影,无扭曲面孔,无字幕,无文字,无标志,无水印,无血腥内容。
交付一段连贯的10秒实拍风格视频,具有清晰的视觉效果、真实的动作、原创同步音频以及高端电影品质。
Comments & Questions (0)
No comments yet
Be the first to comment!