Deepspeech silence detection

DeepSpeech
#1

Hi all,

I have audio files which are more than 7 to 10 min long. As deepspeech cant handle this, I am chunking it with 10 sec each. But my audio files don’t have words spoken throughout the audio. So is there a way to identify if there is no speech spoken? right now I am doing it manually. creating segments but eliminating silence parts in my audio.

thanks

#2

Voice Activity Detection is likely what you are looking for, e.g., https://pypi.org/project/webrtcvad/

1 Like
#3

Hi. I can confirm !! Webrtcvad is very powerful, and works perfectly !
You can, too, select detection level…