视频 图片
登录

Subway Night Fashion Film

观看33,545
点赞586
评论105
分享36
收藏280
引用2
Photoreal cinematic ~9 seconds, high-end fashion-film look. A young woman with long wavy blonde hair, olive-tan skin, sharp brows, and glossy lips. A pale pink flower is pinned behind her right ear. She wears a light blue spaghetti-strap mini dress with small black polka dots and carries a woven straw shoulder bag. Soft glam makeup, gold jewelry, warm skin highlights. Location: underground subway platform at night. Yellow safety strip on the platform edge, tiled walls, overhead fluorescent lights. A silver metro train streaks behind her in heavy motion blur — windows, doors, and at one moment an Japanese flag graphic on the passing car. Shot sequence, continuous energy: Medium close-up, she looks over her shoulder into camera while the train screams past behind her. Profile: she brings a lip wand or finger to her mouth, applying gloss as hair lifts in the train wind. She turns, hair swinging, eyes on camera, slight smirk. Wider on the platform: she touches her collarbone / strap, dress and bag readable, train still moving in the background. Head tipped back, hair flying, chin up, train with the flag ripping through frame. She looks back and smiles. Fast hair flip — hand rakes through hair, back partly to camera. Tight close-up: she blows a kiss to lens, fingers near lips, flower still in hair. Camera: mixed handheld fashion-cam and locked beauty close-ups. Shallow depth of field on her face, background train smeared into silver-and-light streaks. Practical station lighting plus a warm key on her skin. Color: creamy highlights, rich dark hair, the pink flower and red lips pop against cool subway metal. Motion: hair and dress always moving from train rush / her own turns. No extra people in focus. Mood: confident, flirty, summer-night city girl, music-video energy. Audio direction: upbeat dance-pop / Latin-pop, lyrics in the vein of “she got that vibe… dance a

✨ 生成 原始帖子 在 Hitboard 打开

如何使用这个提示词

复制提示词粘贴到 Seedance 2.5。分镜结构、镜头运动与时间点请按原文保留,这是产出该结果的关键;只改主体、场景或风格用词让它成为你的,并附上提示词提到的参考图。第一次生成若跑偏,优先缩短提示词,而不是再加指令。