search

sandrohanea / whisper.net

Whisper.net. Speech to text made simple using Whisper Models
MIT License
512 stars 78 forks source link

Identifying two speakers #87

Closed yakovw closed 1 year ago

yakovw commented 1 year ago

Is there any built-in way to identify two speakers? Thank you

sandrohanea commented 1 year ago

Currently, the only option to identify different speakers is if each speaker will be recorded in a different channel, and in this case, diarization can be used as in example: https://github.com/sandrohanea/whisper.net/blob/main/examples/Diarization/Program.cs

However, if the speakers are not isolated in different channels, there is this on-going effort in the nartive library: https://github.com/ggerganov/whisper.cpp/pull/1058

If that will work and will be merged, this dotnet binding will probably add support as well.

yakovw commented 1 year ago

Thank you