Submit Diarization Job
- Retrieve and store any output that you need before it is deleted.
- See Data retention for the full policy.
- How to Diarize Audio - Complete diarization workflow
- Diarization and Speech-to-Text Merge - Combining diarization with transcription
- How to Upload Files - Working with local files
- Speaker Configuration - Configuring speaker counts and overlapping speech
- VAD and Cross-talk Sensitivity - Tuning detection sensitivity (precision-3)
- Output Scores - Confidence scores, and frame-level speech, cross-talk and speaker probabilities
Authorizations
Body
Number of speakers. Only use if the number of speakers is known in advance. Number of speakers is detected automatically if not provided. Setting this value results in better overall diarization performance. In rare cases where we cannot honor this request (e.g. short files and large number of speakers), a warning will be added to the output. Equivalent to sending minSpeakers==maxSpeakers
x >= 12
Controls the sensitivity of voice activity detection (VAD). Float in [-5.0, 5.0], defaults to 0.0 (balanced precision/recall). Higher values increase recall (less speech missed). Lower values increase precision (less non-speech detected as speech). Only available for the precision-3 model.
-5 <= x <= 50
Controls the sensitivity of overlapping speech (cross-talk) detection. Float in [-5.0, 5.0], defaults to 0.0 (balanced precision/recall). Higher values increase recall (detect more overlapping speech). Lower values increase precision (detect only clear overlaps). Only available for the precision-3 model.
-5 <= x <= 50
Include confidence values in the output. Output is considerably larger when this option is enabled. Output includes a list of confidence scores with a resolution. Not supported by the precision-3 model, which returns an error if this is set; use speechProbability, crosstalkProbability or speakerProbability instead.
true