Detect different speakers in an audio recording

This page describes how to get labels for different speakers in audio data transcribed by Speech-to-Text.

Sometimes, audio data contains samples of more than one person talking. For example, audio from a telephone call usually features voices from two or more people. A transcription of the call ideally includes who speaks at which times.

Speaker diarization

Speech-to-Text can recognize multiple speakers in the same audio clip. When you send an audio transcription request to Speech-to-Text, you can include a parameter telling Speech-to-Text to identify the different speakers in the audio sample. This feature, called speaker diarization, detects when speakers change and labels by number the individual voices detected in the audio.

When you enable speaker diarization in your transcription request, Speech-to-Text attempts to distinguish the different voices included in the audio sample. The transcription result tags each word with a number assigned to individual speakers. Words spoken by the same speaker bear the same number. A transcription result can include numbers up to as many speakers as Speech-to-Text can uniquely identify in the audio sample.

When you use speaker diarization, Speech-to-Text produces a running aggregate of all the results provided in the transcription. Each result includes the words from the previous result. Thus, the words array in the final result provides the complete, diarized results of the transcription.

Review the language support page to see if this feature is available for your language.

Enable speaker diarization in a request

To enable speaker diarization, you need to set the diarization_config field in RecognitionFeatures. You must set the min_speaker_count and max_speaker_count values according to how many speakers you expect in the transcript.

Speech-to-Text supports speaker diarization for all speech recognition methods: speech:recognize and Streaming.

Use a local file

The following code snippet demonstrates how to enable speaker diarization in a transcription request to Speech-to-Text using a local file

Protocol

Refer to the speech:recognize API endpoint for complete details.

To perform synchronous speech recognition, make a POST request and provide the appropriate request body. The following shows an example of a POST request using curl. The example uses the Google Cloud CLI to generate an access token. For instructions on installing the gcloud CLI, see the quickstart.

curl -s -H "Content-Type: application/json" \
    -H "Authorization: Bearer $(gcloud auth application-default print-access-token)" \
    https://speech.googleapis.com/v2/projects/{project}/locations/{location}/recognizers/{recognizer}:recognize \
    --data '{
    "config": {
        "features": {
            "diarizationConfig": {
              "minSpeakerCount": 2,
              "maxSpeakerCount": 2
            },
        }
    },
    "uri": "gs://cloud-samples-tests/speech/commercial_mono.wav"
}' > speaker-diarization.txt

If the request is successful, the server returns a 200 OK HTTP status code and the response in JSON format, saved to a file named speaker-diarization.txt.

{
  "results": [
    {
      "alternatives": [
        {
          "transcript": "hi I'd like to buy a Chromecast and I was wondering whether you could help me with that certainly which color would you like we have blue black and red uh let's go with the black one would you like the new Chromecast Ultra model or the regular Chrome Cast regular Chromecast is fine thank you okay sure we like to ship it regular or Express Express please terrific it's on the way thank you thank you very much bye",
          "confidence": 0.92142606,
          "words": [
            {
              "startOffset": "0s",
              "endOffset": "1.100s",
              "word": "hi",
              "speakerLabel": "2"
            },
            {
              "startOffset": "1.100s",
              "endOffset": "2s",
              "word": "I'd",
              "speakerLabel": "2"
            },
            {
              "startOffset": "2s",
              "endOffset": "2s",
              "word": "like",
              "speakerLabel": "2"
            },
            {
              "startOffset": "2s",
              "endOffset": "2.100s",
              "word": "to",
              "speakerLabel": "2"
            },
            ...
            {
              "startOffset": "6.500s",
              "endOffset": "6.900s",
              "word": "certainly",
              "speakerLabel": "1"
            },
            {
              "startOffset": "6.900s",
              "endOffset": "7.300s",
              "word": "which",
              "speakerLabel": "1"
            },
            {
              "startOffset": "7.300s",
              "endOffset": "7.500s",
              "word": "color",
              "speakerLabel": "1"
            },
            ...
          ]
        }
      ],
      "languageCode": "en-us"
    }
  ]
}

Go

To learn how to install and use the client library for Speech-to-Text, see Speech-to-Text client libraries. For more information, see the Speech-to-Text Go API reference documentation.

To authenticate to Speech-to-Text, set up Application Default Credentials. For more information, see Set up authentication for a local development environment.

import (
	"context"
	"fmt"
	"io"
	"os"
	"strings"

	speech "cloud.google.com/go/speech/apiv1"
	"cloud.google.com/go/speech/apiv1/speechpb"
)

// transcribe_diarization_gcs_beta Transcribes a remote audio file using speaker diarization.
func transcribe_diarization(w io.Writer) error {

	ctx := context.Background()
	client, err := speech.NewClient(ctx)
	if err != nil {
		return fmt.Errorf("NewClient: %w", err)
	}
	defer client.Close()

	diarizationConfig := &speechpb.SpeakerDiarizationConfig{
		EnableSpeakerDiarization: true,
		MinSpeakerCount:          2,
		MaxSpeakerCount:          2,
	}

	recognitionConfig := &speechpb.RecognitionConfig{
		Encoding:          speechpb.RecognitionConfig_LINEAR16,
		SampleRateHertz:   8000,
		LanguageCode:      "en-US",
		DiarizationConfig: diarizationConfig,
	}

	// Get the contents of the local audio file
	content, err := os.ReadFile("../resources/commercial_mono.wav")
	if err != nil {
		return fmt.Errorf("error reading file %w", err)
	}
	audio := &speechpb.RecognitionAudio{
		AudioSource: &speechpb.RecognitionAudio_Content{Content: content},
	}

	longRunningRecognizeRequest := &speechpb.LongRunningRecognizeRequest{
		Config: recognitionConfig,
		Audio:  audio,
	}

	operation, err := client.LongRunningRecognize(ctx, longRunningRecognizeRequest)
	if err != nil {
		return fmt.Errorf("error running recognize %w", err)
	}

	response, err := operation.Wait(ctx)
	if err != nil {
		return err
	}

	// Speaker Tags are only included in the last result object, which has only one
	// alternative.
	alternative := response.Results[len(response.Results)-1].Alternatives[0]

	wordInfo := alternative.GetWords()[0]
	currentSpeakerTag := wordInfo.GetSpeakerTag()

	var speakerWords strings.Builder

	speakerWords.WriteString(fmt.Sprintf("Speaker %d: %s", wordInfo.GetSpeakerTag(), wordInfo.GetWord()))

	// For each word, get all the words associated with one speaker, once the speaker changes,
	// add a new line with the new speaker and their spoken words.
	for i := 1; i < len(alternative.Words); i++ {
		wordInfo := alternative.Words[i]
		if currentSpeakerTag == wordInfo.GetSpeakerTag() {
			speakerWords.WriteString(" ")
			speakerWords.WriteString(wordInfo.GetWord())
		} else {
			speakerWords.WriteString(fmt.Sprintf("\nSpeaker %d: %s",
				wordInfo.GetSpeakerTag(), wordInfo.GetWord()))
			currentSpeakerTag = wordInfo.GetSpeakerTag()
		}
	}
	fmt.Fprintf(w, speakerWords.String())
	return nil
}

Python

To learn how to install and use the client library for Speech-to-Text, see Speech-to-Text client libraries. For more information, see the Speech-to-Text Python API reference documentation.

To authenticate to Speech-to-Text, set up Application Default Credentials. For more information, see Set up authentication for a local development environment.

from google.cloud import speech_v1p1beta1 as speech

client = speech.SpeechClient()

speech_file = "resources/commercial_mono.wav"

with open(speech_file, "rb") as audio_file:
    content = audio_file.read()

audio = speech.RecognitionAudio(content=content)

diarization_config = speech.SpeakerDiarizationConfig(
    enable_speaker_diarization=True,
    min_speaker_count=2,
    max_speaker_count=10,
)

config = speech.RecognitionConfig(
    encoding=speech.RecognitionConfig.AudioEncoding.LINEAR16,
    sample_rate_hertz=8000,
    language_code="en-US",
    diarization_config=diarization_config,
)

print("Waiting for operation to complete...")
response = client.recognize(config=config, audio=audio)

# The transcript within each result is separate and sequential per result.
# However, the words list within an alternative includes all the words
# from all the results thus far. Thus, to get all the words with speaker
# tags, you only have to take the words list from the last result:
result = response.results[-1]

words_info = result.alternatives[0].words

# Printing out the output:
for word_info in words_info:
    print(f"word: '{word_info.word}', speaker_tag: {word_info.speaker_tag}")

return result