Get hands-on experience with 20+ free Google Cloud products and $300 in free credit for new customers.

For Speech-to-Text, how long can API's streaming session remain open after inactivity period?

In Google Cloud Speech-to-Text:

When no audio data (including silent audio) is being sent, the Google Cloud Speech API's streaming session remains open but inactive, waiting for new audio data to process.

How long can the API's streaming session remain open after some inactivity period?
10 seconds?

NB: this is different from VAD timeouts with streaming silence, where audio data representing silence is still being sent and processed to the API's streaming session.

1 2 636
2 REPLIES 2

You have the option to activate voice activity timeouts by adjusting parameters within the streaming_features, specifically by configuring values for the voice_activity_timeout message. These timeouts must be set within the range of greater than 500ms and less than 60 seconds. Additionally, it's possible to define speech begin and end timeouts separately.

As I explicitly wrote, I'm not talking about streaming silence.

I'm talking about when API's streaming session remains open but inactive - waiting for new audio data to process.

How long can this inactivity period last before API's streaming session is ended?