PROMPT RADAR / VIDEO视频叙事
llevo un rato dándole vueltas a esto y no me lo creo Seedance 2.0 on Hig
llevo un rato dándole vueltas a esto y no me lo creo Seedance 2.0 on Higgsfield AI Prompts y assets open-source completos abajo: o sea el prompt que usaron ELLOS para ESTE video, te lo ponen ahí para que lo copies, así d
@robiartec 568 热度 2.5 万 浏览
编辑先看
它为什么值得进入灵感库
它最有价值的部分,是把一个模糊想法推进到了可以被模型执行的程度。 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
llevo un rato dándole vueltas a esto y no me lo creo Seedance 2.0 on Higgsfield AI Prompts y assets open-source completos abajo: o sea el prompt que usaron ELLOS para ESTE video, te lo ponen ahí para que lo copies, así de fácil antes tocabas adivinar cómo lo habían hecho ahora te lo dan todo hecho @Image1 is the lead performer — preserve her exact identity in every shot: cornrow braids, septum ring, statement earrings, sculptural glossy white wrapped top, dark indigo denim. @Audio1 is the music track — entire edit cut to it; no invented music, no voiceover, no lip-sync. Music-video route: dance/performance, rooftop golden-hour editorial. Track: industrial hip-hop, 95 BPM half-time, distorted 808s, luxury-noir. Director thesis: a skyline coronation — the lead and her four dancers own an empty rooftop at last light, silhouettes against the city. Visual world: flat concrete rooftop of a brutalist tower at golden hour, low sun raking long shadows across the roof, city skyline hazy in the background, wind moving fabric. Background dancers: 4 dancers in tonal rust and camel utilitarian looks, sculptural streetwear. Palette: golden amber, rust, concrete, bone white top cutting through, deep indigo. Natural light only, no neon, no particles. Cut map synced to @Audio1 : 0.0–2.0s THE BREATH — static ultra-wide from across the rooftop: five figures spread far apart in a pentagon, all facing the sun, long shadows pointing at camera. Wind and stillness. 2.0s THE DROP — 808 kick: all five pivot to camera in unison on the hit; hard cut to low-angle medium of the lead, sun flaring behind her braids. 2.0–4.0s she struts along the roof edge parapet line (safe, well inside the roof), the four dancers converging on her diagonally from the corners in lockstep, shadows sweeping like clock hands; camera dollying backward. 4.0s snare — whip-pan: full freeze in a tight cluster, she hits a profile vogue arm frame, dancers frozen in four different angular silhouettes against the skyline. 4.0–8.0s golden-hour choreography in medium-wide: her sharp head isolations and New Way vogue arm lines land on the snares; the four dancers frame her in rotating formations — box, diagonal, split pairs — changing formation exactly on each snare hit; torso rolls ripple through the group on the 808 slides. Camera orbits with the sun cross-lighting faces. 8.0s hi-hat triple roll — 3 rapid macro inserts: her earring flaring in backlight, four hands hitting the same vogue frame stacked in depth, boots pivoting on rooftop gravel. 8.5–12.0s escalation: accelerating orbital around the cluster as the dancers whip around her in a fast carousel, her solo arm work speeding up at the axis, long shadows strobing across the lens; speed-ramp slow-motion stretch — hair and fabric floating in wind — then snap back on the kick. 12.0s bass slide — the whole group drops level together into deep sliding lunges across the concrete, camera dutch and low, the sun skimming the roof surface. 13.5–15.0s FINAL BUTTON — hard cut to the opening ultra-wide: pentagon reformed but now all facing camera, the lead center in a frozen vogue pose one arm high, four dancers kneeling on one knee, five long shadows converging at the lens; micro push-in, hold. Loopable. Performance rules: lead stoic, owns the horizon; dancers precise, faces neutral. Continuity: same five people, same wardrobe, same rooftop and light. Audio intent: every cut and formation change on a hit of @Audio1 ; faint foley only (wind, gravel, fabric). Quality bar: expensive editorial music video, natural cinematic grade, no AI gloss, no glow, no lens-flare abuse.
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:llevo un rato dándole vueltas a esto y no me lo creo Seedance 2.0 on Higgsfield AI Prompts y assets open-source completos abajo: o sea el prompt que usaron ELLOS para ESTE video, te lo ponen ahí para que lo copies, así d
- 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