Using SpeechBrain Speaker Diarisation to separate conversations into separate persons


I was wondering if it is possible for me to use SpeechBrain to perform Speaker Diarisation to cut a audio file containing the conversations of 4 persons, into 4 separate audio files for each of the persons?


I think so yes, maybe @nauman can help with that.

Hi Steven,
You can look at the final output in the RTTM file and use start time and duration as information for cutting. For actually cutting the audio into chunks, you can use some audio tools. I am not sure if cutting is possible within SpeechBrain as of now. But it is not there in diarization recipe. I think sox should do that easily.