Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

hope to see the support of personalized deep noise suppression #3

Open
wendongj opened this issue May 31, 2023 · 1 comment
Open

hope to see the support of personalized deep noise suppression #3

wendongj opened this issue May 31, 2023 · 1 comment

Comments

@wendongj
Copy link

thanks for such a great work, in DNS challenge, personalized speech enhancement is gradually replaced the non-personalized speech enhancement, this is a challenge and interesting task, since it need to use speaker recognition model to extract speaker embedding, which is already existed in We-Speaker

@StuartIanNaylor
Copy link

StuartIanNaylor commented May 31, 2023

https://github.com/BUTSpeechFIT/speakerbeam

There was this paper that looks really interesting but can not find any code examples https://arxiv.org/abs/2106.02934

Also with Esspresif's laterial thought on BSS you could replace the KWS on each signal output with VAD that maybe uses ondevice training of a smaller model to bias to a voice profile?
VAD can often be lighter than KWS and maybe post process KWS or upstream.

I agree totally that extraction of a known than cancellation of the unknown seems to be the way forward and the above paper seems the only alternative to Googles VoiceFilterLite (Please opensource :) ) so that it can run downstream on lite hardware or concurrent multiple instances.

Also https://arxiv.org/abs/2106.04129 but again no code.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants