FunAudioLLM / CosyVoice

Multi-lingual large voice generation model, providing inference, training and deployment full-stack ability.
https://funaudiollm.github.io/
Apache License 2.0
4.75k stars 479 forks source link

流式推理总感觉有爆破音如何解决 #341

Open lucasjinreal opened 2 weeks ago

lucasjinreal commented 2 weeks ago

流失推理,直接stream写入到pc扬声器,总感觉有破麦的那种声音,应该如何解决啊

https://github.com/user-attachments/assets/d675e682-b2bc-468f-b850-aa4bea0b396b

很大概率爆破音

aluminumbox commented 1 week ago

we may try to fix it, this is due to flow model discontinuity

lucasjinreal commented 1 week ago

@aluminumbox Thank you very much for your attention. Is there any estimated fix time for this?

luohao123 commented 5 days ago

@aluminumbox Hi, any updates for this? Same for stream