Recognize.js on Audio Stream is not very realtime

[rcoenen]

I have been testing extensively with node recognize.js listen (to microphone input)

It appears to me that the real-time transcription of text is not nearly as realtime as compared to using Google Search microphone input on Chrome or for example my Android device.

Is there a way to make it more realtime by for example making the chunk size smaller or force more frequent updates?

[jmdobry]

You can receive interim results, e.g.

const request = {
  config: {
    encoding: encoding,
    sampleRateHertz: sampleRateHertz,
    languageCode: languageCode
  },
  interimResults: true
};

See https://cloud.google.com/speech/reference/rpc/google.cloud.speech.v1#google.cloud.speech.v1.StreamingRecognitionConfig

[fhinkel]

If the issue still persists, please file it here. Thank you.