Cinematic Somersault Performance Video Prompt

Style: 8K cinematic texture. Real photography—not 3D rendering, not game engines, not game cutscenes. 
Photography: Naturalist master-level photography. 
Lighting: Artificial studio light sources only—softbox as key light from the left side of the frame, Fresnel lamp as rim light, a spotlight on the background wall, a warm lamp in the corner for environmental accent, and thin airborne dust/mist. 
Color: 60:30:10—Main color / Secondary color / Accent color. 
Lens: Physical cinematic lens. 180° shutter motion blur. 
Skin: Pore-level realism—vellus hair, asymmetrical small moles, capillary redness, pore shadows matching the site lighting. 
Performance: Top-tier cinematic—micro-pauses before reactions, precise eye contact, moist and lifelike eyes with catchlights, visible breathing and chest rise/fall. 
Physics: Respect gravity and inertia—mass has real weight, correct contact shadows, skirt and hair strands swing realistically with movement. No floating props. 
Composition: Rule of thirds + Golden ratio. The subject is vivid and active from the first frame. 
Continuity: Person, clothing, hairstyle, stockings, and leather shoes are 100% consistent across all shots. No identity drift. 
Tech: 24fps smooth motion. 8K detail. No jitter. 
Audio: Clean vocals + action SFX (stepping, landing, fabric sounds) + room ambient noise. No background music. No subtitles. 

SUBJECT —— @ wenyou (100% consistent with input image): An academic-style girl-next-door type young woman, long brown wavy hair with air bangs, cream-white academic knit cardigan (navy stripes on cuffs) + beige vest, plaid pleated miniskirt, thin black stockings, brown loafers. An audition self-tape including a gymnastics talent segment. Goal: Full body in frame, cheering with fresh energy and a clean cartwheel/front flip, making the audience feel her vitality and like her within 10 seconds. Emotional beats: Bright greeting → flip explosion → landing with a finger heart. White balance 5600K. MULTISHOT. 

LOCATION —— @ Image 

ACTION —— She brightly introduces herself, then runs and performs a clean cartwheel (transitioning to a front flip) as a talent, landing steadily and cheering with a finger heart. Full body always in frame, showing the run-up, mid-air, and landing cushion. Mouth movements and breathing strictly lip-sync to the following dialogue. 
Dialogue + Talent (Chinese, approx. 10s): 
"Hi, I'm Wen You! (smiling eyes) Words aren't enough—watch this!" (light run-up, mid-air flip) (steady landing, double finger heart) "If you like it, please vote for me, let's go!" 
Shot 1 (0:00–0:03) —— Medium wide shot: @ wenyou stands still, tilts her head with a bright smile, brushes her hair, greets warmly and introduces herself. Hard cut. 
Shot 2 (0:03–0:08) —— Full body wide shot: She takes a few light run-up steps, executes a clean cartwheel into a front flip, skirt and curls swing realistically with the rotation, body stretched, axis stable, lands firmly with a slight knee cushion and a realistic weight sound. Hard cut. 
Shot 3 (0:08–0:10) —— Medium close-up: After landing, she straightens up with a post-exercise flush and a radiant smile, makes a large heart with her hands in front of her chest, looks at the camera and says "If you like it, please vote for me, let's go!". Freeze on smiling eyes. Hard cut. 

CAMERA —— Shot 1: Medium shot from the chest up, eye level, 50mm, locked position with slight handheld breathing; Motive = bright first impression. Shot 2: Zoom out to full body shot, slightly low eye level, 28mm wide angle, smooth horizontal tracking pan following the flip, fully capturing the airborne arc; Motive = displaying gymnastic tension. Shot 3: Push back to medium close-up, eye level, 65mm, slow zoom-in on the heart gesture; Motive = magnifying energetic infection. 

STYLE —— Main color: dark studio set + light gray background wall 60% / Secondary color: cream-white cardigan + warm skin tone 30% / Accent color: plaid skirt + navy stripes + background spotlight 10%. White balance 5600K. Softbox key light from left, Fresnel rim light, thin mist. 

CONSTRAINTS —— 16:9. Keep normal speed for the flip, no slow motion or wire floating, mid-air and landing constrained by real gravity and inertia, landing has cushion weight. Full body in frame, run-up, mid-air, and landing all visible. Stable camera without motion sickness. Natural and precise lip-sync. No markings, stickers, or tracking points on the ground or in the frame. Maintain 100% identity, clothing, hair, stockings, shoes, and studio set consistency across all shots. No glowing eyes.

风格：8K 电影质感。真实摄影——非 3D 渲染，非游戏引擎，非游戏过场画面。
摄影：自然主义大师级摄影。
布光：仅人工片场光源——柔光箱作主光自画面左侧，菲涅尔灯作边缘轮廓光，背景墙一束聚光，角落暖光台灯作环境点缀，淡薄空气尘雾。
色彩：60:30:10——主色 / 辅色 / 点缀色。
镜头：物理电影镜头。180° 快门运动模糊。
皮肤：毛孔级真实——汗毛、不对称小痣、毛细血管泛红、毛孔阴影与现场光匹配。
表演：顶级电影级——反应前的微停顿、精准视线、湿润有生命力的眼睛带眼神光、可见的呼吸与胸口起伏。
物理：尊重重力与惯性——质量有真实重量，correct 接触阴影，裙摆与发丝随动作真实甩动。无悬浮道具。
构图：三分法 + 黄金比例。主体从第一帧起就鲜活、有动作。
连续性：人物、服装、发型、丝袜、皮鞋在每个镜头间完全一致。无身份漂移。
技术：24fps 流畅运动。8K 细节。无抖动。
音频：干净的人声＋动作 SFX（踏地、落地、衣料声）＋房间环境底噪。无配乐。无字幕。

主体 SUBJECT —— @ wenyou（与输入图 100% 一致）：学院风邻家学姐型年轻女生，长棕色波浪卷发配空气刘海，米白学院针织开衫（袖口藏青条纹）＋米色背心，格纹百褶短裙，黑色薄丝袜，棕色乐福鞋。一段试镜自拍带，含一段体操才艺。目标：全身入镜，用清爽元气＋一个利落的侧手翻/前空翻应援，在 10 秒内让观众被她的活力感染、喜欢上她。情绪节拍：明亮打招呼 → 翻身爆发 → 落地比心。白平衡 5600K。MULTISHOT 多镜头。

场景 LOCATION —— @ 图片

动作 ACTION —— 她明亮地自报家门，随即助跑翻一个利落的侧手翻（接前空翻）作为才艺，稳稳落地后双手比心应援。全身始终入镜，能看到助跑、腾空与落地缓冲。嘴型与呼吸严格对口型于以下对白。
对白＋才艺（中文，约 10 秒）：
"嗨，我是温柚！（弯眼一笑）光说不练可不行——看好咯！"（轻快助跑，腾空翻身）（稳稳落地，双手比心）"喜欢的话，给我投一票呀，加油！"
镜头 1（0:00–0:03）—— 中景偏全：@ wenyou 站定，歪头弯眼一个明亮甜笑，撩了下垂发，亲切打招呼并报名"嗨，我是温柚！光说不练可不行——看好咯！"。硬切。
镜头 2（0:03–0:08）—— 全景全身入镜：她两三步轻快助跑，一记干净的侧手翻接前空翻腾空，裙摆与卷发随翻转真实甩动，身体舒展、轴线稳定，落地双脚踏实带轻微屈膝缓冲，踏地有真实重量声。硬切。
镜头 3（0:08–0:10）—— 中近景：她落地后顺势直起身，脸上带着运动后的红晕与灿烂笑容，双手在胸前比一个大大的心，看着镜头说"喜欢的话，给我投一票呀，加油！"。定格在弯弯笑眼上。硬切。

镜头运动 CAMERA —— 镜头1：胸部以上中景，平视，50mm，锁定机位轻呼吸手持；动机＝明亮第一眼。镜头2：拉远全景全身，平视略低，28mm 广角，随翻身做平滑横移跟摇，完整收住腾空弧线；动机＝展示体操张力。镜头3：推回中近景，平视，65mm，微缓推近比心；动机＝放大元气感染力。

风格配比 STYLE —— 主色 深色棚景＋浅灰背景墙 60% / 辅色 米白开衫＋暖肤色 30% / 点缀色 格纹裙＋藏青条纹＋背景聚光 10%。白平衡 5600K。柔光箱主光自画面左侧，菲涅尔边缘光，淡薄尘雾。

约束 CONSTRAINTS —— 16:9。翻身保持正常速度，禁止慢动作与威亚漂浮，腾空与落地受真实重力与惯性约束、落地有缓冲重量。全身入镜，助跑—腾空—落地全程可见。镜头稳定不晕。嘴型自然精准。地面与画面中不得出现任何标记符号、贴纸或定位点。在所有镜头间 100% 保持身份、服装、发型、丝袜、皮鞋与棚内陈设一致。禁止眼睛发光。