PROMPT RADAR / VIDEO视频叙事
Create your version on @budgetpixel using Seedance 2.0 Prompt: A dynamic
Create your version on @budgetpixel using Seedance 2.0 Prompt: A dynamic, cinematic action sequence inside a luxury jewelry boutique at night. A confident young woman with long dark hair, olive skin, and sharp green eyes
@mimu_ai1 962 热度 3.9 万 浏览
编辑先看
它为什么值得进入灵感库
先别急着照抄整段 Prompt:真正有用的是主体、空间和镜头之间的约束关系。 把一段描述拆成起始状态、动作变化、镜头响应和结束画面,连续性比形容词更重要。
01 / 原始方法
动作与镜头 Prompt
Create your version on @budgetpixel using Seedance 2.0 Prompt: A dynamic, cinematic action sequence inside a luxury jewelry boutique at night. A confident young woman with long dark hair, olive skin, and sharp green eyes is the hero. She wears an elegant emerald silk evening gown and gold earrings. Three masked thieves in black tactical gear attempt a robbery, and she single-handedly defeats them with fast, precise martial arts choreography. She vaults over glass display cases, shattering them in dramatic slow motion as diamonds, pearls, and glass shards scatter and sparkle through the air like rain, catching warm spotlight beams. She slides across a marble counter, throws one attacker into a shelf of watches, and disarms the last with a spinning kick. At the end, she stands calm amid the glittering wreckage, picks up a single diamond ring, examines it in the light, and looks directly into the camera with a faint smile. Dynamic handheld camera, whip pans, slow-motion inserts, moody warm lighting with deep shadows, realistic physics, high production value, cinematic style, 4K.
02 / 复刻路径
不要一次改完所有变量
用 Seedance 复刻时,建议先单独生成稳定首帧,再把动作拆成“开始、变化、收束”三段;每段只保留一个主要运动。
- 主体:Create your version on @budgetpixel using Seedance 2.0 Prompt: A dynamic, cinematic action sequence inside a luxury jewelry boutique at night. A confident young woman with long dark hair, olive skin, and sharp green eyes
- 视觉方向:实时 AI 视频案例;查看原帖媒体与作者说明
- 镜头与构图:按原帖镜头流程执行;查看原帖媒体与工作流
- 交付约束:auto,保留原作者方法与归属;复刻时避免水印、低清、畸变、主体漂移和无关文字。
负向约束
保留原作者方法与归属;复刻时避免水印、低清、畸变、主体漂移和无关文字。
模型建议
Seedance / Veo / Kling。先用原始 Prompt 建立基准结果,再根据模型对自然语言、镜头运动或文字排版的能力做局部改写。
来源与编辑原则