Gpt-4o-transcribe unsupported format with .wav

API
, ,
1

I’ve been using whisper-1 for my audio process and it’s been fine so far, today I tried to switch and test the gpt-4o-transcribe, I know it has stricter output format, but input-wise, the api doc said it accept .wav, which is what i’ve been using smoothly with whisper. When I switch to the 4o-transcribe model, I keep getting :ERROR:main:Audio processing error: Error code: 400 - {‘error’: {‘message’: ‘This model does not support the format you provided.’, ‘type’: ‘invalid_request_error’, ‘param’: ‘messages’, ‘code’: ‘unsupported_format’}}

Here’s my code:

def process_audio(contents):
    try:
        with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as temp:
            temp.write(contents)
            temp_path = temp.name  # Store the file path for later use
        
        # Reopen the file in binary mode and pass it to the API
        with open(temp_path, "rb") as temp_file:
            transcript = client.audio.transcriptions.create(
                model="gpt-4o-transcribe", #whereas before it was whisper-1
                file=temp_file,
                response_format="json"
            )

    # Clean up temporary file
    os.unlink(temp_path)
    
    return transcript
except Exception as e:
    logger.error(f"Audio processing error: {e}")
    raise
2

i have been having the exact same issue

3

This would happen if it was some other format (e.g. mp3) named as .wav. Could you double check that the file inside is indeed wav and not some other format?

1 Like