演示 — 输入与输出
Prompt 结构参考 Inspired by fal.ai H3 Guide
Voice Clone Dialogue Transfer
Replace spoken line using audio reference for voice timbre and delivery.
适用对象
ADR, localization, and dialogue revision workflows.
此 Prompt 如何构建
请先阅读 — 理解模式后，你可以从零撰写自己的 Prompt。
- Dialogue
Target line text.
- Audio 1
Voice reference — timbre and performance.
改编到你的项目
- Swap line text for your script.
- Shorter 5s clip for tests.
常见失误
- Ignoring breath pauses in audio ref.
- Lip sync longer than mouth movement.
完整 Prompt（复制并编辑）
The character says: "Follow the wind, live free. Leave worries behind, enjoy the moment." Match the voice in Audio 1 for timbre, pacing, and breath placement. Keep lip sync natural to the new line length. Preserve the original scene lighting and camera framing from the source video.
Ref2VA 新手？阅读 MiniMax H3 Ref2VA 指南
相关 Prompt
Multi-Asset Cinematic Remix
Maximum Ref2VA asset count demo (6 pictures)
Prompt 结构
Use Images 1–6 as assets. Match Reference Video 1 closely for shot rhythm, transition language,…
Street Dance Motion Transfer
Video motion ref + 2 character images
Prompt 结构
Use Video 1 as the motion reference for a street-dance performance. Use @image1 and @image2 as …
Live-Action Voxel Transformation
Video + style image selective replacement
Prompt 结构
Preserve the buildings, pedestrians, and overall environment in Video 1 as photoreal live actio…
Desert Highway Fashion Campaign
Fashion or accessories brands with model shots, product stills, and a logo end card — typical 15s campaign cut.
Prompt 结构
@image1 sets mood, desert light, and 35mm film grain. @image2 locks the female lead's face, hai…