🔊 Dasheng AudioGen Demo
Developed by SJTU X-LANCE & Xiaomi LLM Plus
支持结构化 Prompt 的混合音频生成，可用自然语言描述场景(Auto mode)或逐轨道填写(Manual mode)，一次生成包含音乐、可理解人声和音效的完整音频。。
Structured-prompt mixed audio generation that lets you describe a scene in natural language (Auto mode) or specify tracks manually (Manual mode), producing a complete audio clip with music, intelligible speech, and sound effects in one pass.
🤖 Auto Mode
If Auto Mode is unavailable, contact jiahaomei@sjtu.edu.cn.
输入整体音频描述，系统将调用 Prompt Refiner 自动转换为结构化 Prompt，再通过 DashengAudioGen 生成音频。
Enter an overall audio description. The system will call the Prompt Refiner to convert it into a structured prompt, then generate audio via DashengAudioGen.
Examples