Google speech recognition api - bad request

Question

I have an issue with google speech recognition API. I am making successful calls with the examples provided in the documentation.

However, my input is in mp3 format (8 Khz). I suspect that the transformation to FLAC which I did with an online tool may be the issue.

Here is the body of my call:

{
  "config": {
      "encoding":"FLAC",
      "sampleRateHertz": 8000,
      "languageCode": "en-US",
      "enableWordTimeOffsets": false
  },
  "audio": {
      "uri":"gs://speech-demo/phone3.flac"
  }
}

I get a bad request. If I use FLAC, 16 KHz for the example provided I get a full transcript.

Any idea what I was doing wrong, and if it's the conversion part, how I should be converting?

Answer 1

"languageCode": "en-IN"

8000 is defined for en-IN.