recognize_audio_samples

Transcribe pre-segmented speech

This event can be used to transcribe pre-segmented speech and should provide the following payload:

• audio samples serialized as bytes
• audio sample type string (one of ‘int16’, ‘int32’, ‘float32’, ‘float64’)

JavaScript

const response = await fetch('test_audio.wav')
const blob = await response.blob()
const arrayBuffer = await blob.arrayBuffer()
const audioData = new Int16Array(arrayBuffer)

socketLocal.emit('recognize_audio_samples', audioData, 'int16')

The server will respond by immediately emitting a request_received event with the following payload:

JavaScript

{  
    request_id:  str  
}

When request processing is completed, the server will emit a recognition event with the following payload:

JavaScript

{  
    text:  str  
    request_id:  str  
}

The text value may be null if the processed segment didn’t contain any speech, or if request processing failed.

The event may be captured as follows:

JavaScript

socketLocal.on('recognition', (data) => {
  console.log('recognition', data.text, new Date())
}