Diarization

{
  "diarization": [
    {
      "speaker": "SPEAKER_00",
      "start": 15,
      "end": 30.5,
      "confidence": {
        "SPEAKER_00": 16,
        "SPEAKER_01": 93
      }
    }
  ],
  "exclusiveDiarization": [
    {
      "speaker": "SPEAKER_00",
      "start": 15,
      "end": 30.5,
      "confidence": {
        "SPEAKER_00": 16,
        "SPEAKER_01": 93
      }
    }
  ],
  "wordLevelTranscription": [
    {
      "start": 0.5,
      "end": 0.8,
      "text": "Hello",
      "speaker": "SPEAKER_00"
    }
  ],
  "turnLevelTranscription": [
    {
      "start": 0.5,
      "end": 2.3,
      "text": "Hello, how are you?",
      "speaker": "SPEAKER_00"
    }
  ],
  "error": "<string>",
  "warning": "<string>"
}

A diarization object represents the results of a diarization job.

diarization

object[]

required

List of diarization segments

Show child attributes

confidence

object

Confidence scores details. Only returned if confidence is set to true when job is created

Show child attributes

exclusiveDiarization

object[]

Exclusive diarization segments where only one speaker is active at a time. Only returned if exclusive is set to true when job is created.

Show child attributes

speechProbability

object

Frame-level speech probability curve. Only returned if speechProbability is set to true when job is created. Only available for the precision-3 model.

Show child attributes

crosstalkProbability

object

Frame-level overlapping speech (cross-talk) probability curve. Only returned if crosstalkProbability is set to true when job is created. Only available for the precision-3 model.

Show child attributes

speakerProbability

object

Per-speaker, frame-level probability scores. Only returned if speakerProbability is set to true when job is created. Only available for the precision-3 model.

Show child attributes

wordLevelTranscription

object[]

Word-level transcription segments with text. Only returned if transcription is set to true when job is created.

Show child attributes

Example:

[
  {
    "start": 0.5,
    "end": 0.8,
    "text": "Hello",
    "speaker": "SPEAKER_00"
  }
]

turnLevelTranscription

object[]

Turn-level (speaker turn) transcription segments with text. Only returned if transcription is set to true when job is created.

Show child attributes

Example:

[
  {
    "start": 0.5,
    "end": 2.3,
    "text": "Hello, how are you?",
    "speaker": "SPEAKER_00"
  }
]

error

string

Error message if any

warning

string

Warning message if any

Introduction

Test

Jobs

Streaming

Media

Schemas