How to increase the time for which the Microsoft Speech Service SDK listens in a single go?

Question

I am using MS speech service sdk for speech to text conversion. When I speak, my speech is converted to text after 60 seconds even if I haven't stopped speaking. It basically considers it one chunk and starts processing it. What can I do to increase this timeout?

I want the service to keep listening even if I speak for 80 seconds and convert the speech to text only after I have finished speaking.

Answer 1

Thanks for reaching out to us and posting this query.

Could you please use the SegmentationSilenceTimeoutMs property and check if that helps ?

See this documentation which talks about this setting in detail:

The default behaviors for non-speech silence in input audio may not produce the expected results if the user is speaking faster or slower than usual. Please refer this documentation, which explains how to handle this timeout.

Also this SO thread and this article has some suggestions on the timeout settings like InitialSilenceTimeout and EndSilenceTimeout available within the speech sdk. Hope this helps.