seedance-audio
SKILL.md
seedance-audio
Audio design, lip-sync, and multi-character dialogue for Seedance 2.0 video generation.
Source intelligence: ByteDance official Seedance 2.0 release blog (seed.bytedance.com), Douyin/抖音 creator community, CSDN practitioner tutorials, Q1 2026. Western sources have minimal real-world data.
⚠️ Critical Distinction: Two Separate Tools on the Jimeng Platform
The Jimeng platform hosts two completely different tools that both involve lip-sync. Mixing them up is the most common documentation error.
| Tool | Model | Where to find | What it does |
|---|---|---|---|
| 视频生成 (Video Generation) | Seedance 2.0 | Jimeng → Video Generation → Seedance 2.0 | Generates full video clips (4–15 s) with native audio-video joint generation. Audio is part of the generated output. |
| 数字人 (Digital Human) | OmniHuman-1 | Jimeng → Digital Human | Portrait animation tool — uploads a face image + audio → generates a talking head with precise lip-sync. Has Master/Quick/Standard modes. |
This skill covers Seedance 2.0 video generation only. The Master mode (大师模式), Quick mode (快速模式), Standard mode (标准模式), and OmniHuman-1 engine all belong to the Digital Human tool — not Seedance 2.0. Do not import those concepts here.