Search code examples
c#httpwebrequestspeech-to-text

Trying to use Google Speech2Text in C#


The following simple code tries to post a wave file to Google Speech2Text service, but always fails with either a "Gateway Timeout (504)" or general exception "The operation timed out". Can anyone help please?

  public void ProcessWaveFile(string path)
    {
        HttpWebRequest request = 
             (HttpWebRequest)HttpWebRequest.Create(
                 "https://www.google.com/"+ 
                 "speech-api/v1/recognize?"+ 
                 "xjerr=1&client=speech2text&lang=en-US&maxresults=10");

        ServicePointManager.ServerCertificateValidationCallback += 
                     delegate { return true; };

        request.Timeout = 60000;
        request.Method = "POST";
        request.KeepAlive = true;           
        request.ContentType = "audio/wav";
        request.UserAgent = "speech2text";

        FileInfo fInfo = new FileInfo(path);
        long numBytes = fInfo.Length;
        byte[] data;

        using (FileStream fStream = new FileStream(
                                          path, 
                                          FileMode.Open, 
                                          FileAccess.Read))
        {
            data = new byte[fStream.Length];
            fStream.Read(data, 0, (int)fStream.Length);
            fStream.Close();
        }

        using (Stream wrStream = request.GetRequestStream())
            wrStream.Write(data, 0, data.Length);

        try
        {
            HttpWebResponse response = (HttpWebResponse)request.GetResponse();
            var resp = response.GetResponseStream();

            if (resp != null)
            {
                StreamReader sr = new StreamReader(resp);
                MessageBox.Show(sr.ReadToEnd());

                resp.Close();
                resp.Dispose();
            }
        }
        catch (System.Exception ee)
        {
            MessageBox.Show(ee.Message);
        }           
    }

Many thanks.

Shujaat


Solution

  • Your code works for me with the following change:

    Request.ContentType = "audio/x-flac; rate=8000";
    

    and the file that you provide needs to be in FLAC format.

    I recorded a small sample with Windows Sound Recorder which produces an WMA file. Then I used the VLC Player to convert the WMA file to FLAC (using the Convert options, make sure to output in RAW format, one channel and 8000 kbps)

    A good non c# reference with as well some extended api doc in the comments