search

wiseman / py-webrtcvad

Python interface to the WebRTC Voice Activity Detector
Other
2.02k stars 406 forks source link

Functional Description of Algorithm #46

Closed schmiph2 closed 4 years ago

schmiph2 commented 4 years ago

Hi Py-WebrtcVAD-Team

Thank you for wrapping the webrtcvad. Everything worked straightforward and the algo had a pretty good performance. I want to integrate the algorithm in an embedded solution. As i'm already calculating features related to the log-power of the six bands used by the VAD on my device, I am interested in the detailed design of the algorithm to adapt it to my features. Have you any description? Like a paper or a flowchart? What I found out from the code is, that some sort of adaptive GMM (sequential EM?) is involved. Do you have any further information?

Thank you in advance?

wiseman commented 4 years ago

Hello! The VAD code is by Google, and the last time I checked they gave no description of what the code does beyond comments in the code. Sorry!