Speaker detection in live streams
Support for speaker detection in live streams
Timbra uses speaker recognition to split up captions and improve the transcript.
When editing the project, you can choose an event type with the expected number of speakers. This helps our speech recognition to distinguish different voices. When in doubt, pick the highest number of speakers.
If a speaker differentiator symbol is configured in your language profile, it will also appear in your live stream when that language profile is used.
Speaker differentiator is currently in beta and may change or be discontinued as our development and product plans evolve. Availability is not guaranteed. Please contact your Account Manager or Customer Success Manager for more information on testing with this feature.