Veo 3.1 原生对白与口型同步写法（120ms lip-sync）

STORYBOARD PREVIEW

A four-frame storyboard of the prompt’s visual direction. Actual video results depend on generation.

Model: veo-3
Duration: 5s
Aspect ratio: 16:9
Resolution: 720p
Images: 0
Media refs: — / —

Make the shot betterVeo 3.1 的 48kHz native audio 对白标准写法：角色描述+says+精确引语+语气+lip-sync 时延；与已有 Veo 3 对白条目不同，这是 3.1 专属的同步语法。演示视频：https://www.youtube.com/embed/48jtRE42Bw0

What to look for本站尚未实测。检查主体一致性、运镜连续性与动作完整性；参考原始来源核对效果。

Community source · not tested on Genwww.atlascloud.ai ↗ · cclank/lanshu-awesome-ai-video-kit
MIT (repository)

The complete prompt v1

Dialogue Structure: "Character [Description] says '[Exact Quote],' [Tone Adjective], [Sync Timing]." Example: A middle-aged detective says 'I told you not to come here,' weary and gravelly tone, 120ms lip-sync delay. Use quotation marks for spoken lines: putting the exact dialogue in quotes (says, "...") gives the model a clear start and stop point. List ambient sounds (like rain tapping or cafe murmur) at the end of the prompt so the engine does not blend environmental noise into the main speech track.

Adapt in the studio

Original model: Veo · veo-3
Original parameters: {}
A model category does not mean it is available on Gen. Check the studio for currently available models; review settings when adapting a prompt across models.

Reuse carries the prompt, model, duration, ratio and published references into the studio. Add any missing files and review the price before submitting.

Read the video guide