-1
       String jsonRequest = "{\"config\": {\"languageCode\":\"en-US\"},\"audio\": {\"content\": \"" + base64Content+ "\"}}";
        String str = "";
        var speech = SpeechClient.Create();
        var response = 
        speech.Recognize(RecognizeRequest.Parser.ParseJson(jsonRequest));
        foreach (var result in response.Results)
        {
            foreach (var alternative in result.Alternatives)
            {
                Console.WriteLine(alternative.Transcript);
                str += alternative.Transcript;
            }
        }

This code is working fine with mono .wav files but it throws exception for stereo files. The exception says

Status(StatusCode=InvalidArgument, Detail="Must use single channel (mono) audio, but WAV header indicates 2 channels.")

So, my question is how can I add support for stereo files? How to convert multi channel audio to a single channel in c#? I have already tried this answer so plz don't refer to it. It is not working.

TheGeneral
  • 79,002
  • 9
  • 103
  • 141
Mean Coder
  • 304
  • 1
  • 12
  • 1
    _"I have already tried this answer so plz don't refer to it. It is not working"_ - but it contains the answer. If you can't get it to work, create a [mcve] and ask for help getting it to work. Or choose an existing audio library and call it to obtain the single channel or downmixed audio. – CodeCaster Jun 11 '18 at 08:30
  • Can you give more information on what you try to achieve? Where do the audiofiles come from? The best way is to fit the audiofile to fit the API and not converting it. Do you have the possibility of using mono files? – M Stoerzel Jun 11 '18 at 09:28
  • Audio can come from any where. It could be a song, a conversation or a lecture. I can freely browse a file and the code will convert it to text for me. If any selected file is mono then the above code works otherwise Iit doesn't work. I hope it makes sense. – Mean Coder Jun 11 '18 at 10:59

1 Answers1

0

You should take a look into sox which can convert nearly any format into another including sample rate conversion, and in your case interesting, channel conversion. In the documentation you find many examples on how to use it.

In your case I would advise against just extracting one channel from the audio because that could mean that the desired audio is only on another channel which is not selected.

If you want to have full control over the audio you could dive into bass.dll in liaison with bass.net.dll, which allows you to mix channels together or select a specific channel.

M Stoerzel
  • 300
  • 1
  • 13