Google Cloud Speech-to-Text: “INVALID_ARGUMENT: Invalid recognition 'config': bad encoding..” codec audio encoding error

问题

I'm recording short audio files (a few seconds) in Chrome using mediaDevices.getUserMedia(), saving the file to Firebase Storage, and then trying to send the files to Google Cloud Speech-to-Text from a Firebase Cloud Function. I'm getting back this error message:

INVALID_ARGUMENT: Invalid recognition 'config': bad encoding.

Google's documentation says that this error message means

Your audio data might not be encoded correctly or is encoded with a codec different than what you've declared in the RecognitionConfig. Check the audio input and make sure that you've set the encoding field correctly.

In the browser I set up the microphone:

navigator.mediaDevices.getUserMedia({ audio: true, video: false })
.then(stream => {

var options = {
   audioBitsPerSecond : 128000,
   mimeType : 'audio/webm;codecs=opus'
};

const mediaRecorder = new MediaRecorder(stream, options);
mediaRecorder.start();
...

According to this answer Chrome only supports two codecs:

audio/webm
audio/webm;codecs=opus

Actually, that's one media format and one codec. This blog post also says that Chrome only supports the Opus codec.

I set up my Firebase Cloud Function:

// Imports the Google Cloud client library
const speech = require('@google-cloud/speech');

// Creates a client
const client = new speech.SpeechClient();

const gcsUri = 'gs://my-app.appspot.com/my-file';
const encoding = 'Opus';
const sampleRateHertz = 128000;
const languageCode = 'en-US';

const config = {
   encoding: encoding,
   sampleRateHertz: sampleRateHertz,
   languageCode: languageCode,
};
const audio = {
   uri: gcsUri,
};

const request = {
   config: config,
   audio: audio,
};

// Detects speech in the audio file
return response = client.recognize(request) // square brackets in ES6 construct an array
.then(function(response) {
console.log(response);
...

The audio encoding matches between the browser and the Google Speech-to-Text request. Why does Google Speech tell me that the audio encoding is bad?

I also tried using the default options in the browser, with the same error message:

navigator.mediaDevices.getUserMedia({ audio: true, video: false })
.then(stream => {

const mediaRecorder = new MediaRecorder(stream);
mediaRecorder.start();

In the Firebase Cloud Function I tried leaving out the line const encoding = 'Opus';, which resulted in an error encoding is not defined. I tried this line const encoding = ''; which resulted in the INVALID_ARGUMENT: Invalid recognition 'config': bad encoding.. error.

I'm getting a similar error message from IBM Watson Speech-to-Text. The file plays back without a problem.

来源：https://stackoverflow.com/questions/60747880/google-cloud-speech-to-text-invalid-argument-invalid-recognition-config-ba

标签

getusermedia

google-cloud-speech