KwaiVGI / LivePortrait

Bring portraits to life!
https://liveportrait.github.io
Other
11.45k stars 1.17k forks source link

v2v bug: the audio of the Driving Video is not concatenated #179

Open henjicc opened 1 month ago

henjicc commented 1 month ago

我用的是 webui (app.py),v2i模式下,音频是正常的,但是v2v模式下,音频并没有被拼接

I am using webui (app. py). In v2i mode, the audio is normal, but in v2v mode, the audio is not concatenated

cleardusk commented 1 month ago

用的是source video的音频 @henjicc

The audio is from source video by default in v2v mode.

henjicc commented 1 month ago

用 source video 的音频感觉有点奇怪,因为 v2v 的时候,口型会随 Driving Video 而改变,所以用 Driving Video 的音频应该会更合理点

cleardusk commented 1 month ago

感谢反馈。是个好问题,我理解用source或者driving的音频都有需求,下周我们会加一个选项,让用户自己选择。

Thank you for the feedback. It’s a good question. I understand there is a need for using either source or driving audio. Next week, we will add an option for users to choose themselves.

henjicc commented 1 month ago

非常感谢!

cleardusk commented 1 month ago

I have temporarily changed the priority: driving audio first, source audio second (if driving audio is NULL). @henjicc

cleardusk commented 1 month ago

Just pull the latest main branch.

henjicc commented 1 month ago

好的,非常感谢! Okay, thank you very much!