Automatic silence detection
Up one level
Automatic silence detection
I have WAV files which recorded dialogues with each speaker in a separate channel. I would like to make two tiers each of which includes silence and non-silence spans of a speaker as a preprocessing for transcription. Reading the manual, I could not see the concrete procedure to use silence recogniser for this purpose. Which parts should I give the system as examples of silence span? Spans where both right and left channels are silence? Or spans where only either channel is silence? I could not see how to use "Add channel 1" and "Add channel 2" buttons. Thanks for your advice.
Re: Automatic silence detection
You can first select one or more segments with silence in channel 1 (ignoring channel 2 altogether) and click the Add Channel 1 button. Then select one or more silence segments in channel 2 (ignoring channel 1) and click the Add Channel 2 button. After running the recognizer you can create separate tiers for the channels (i.e. speakers) and you can specify whether yoy want both the silence and the speech segments converted to annotations.
-Han
Re: Re: Automatic silence detection
Han, thank you for the information. I read the passage you quoted in the manual, but could not see the way. After a several try and error, I succeeded.
Thanks again.
Third party update, regular expressions
