SpeechRecognitionEngine in BackgroundWorker

Question

I am trying to write a C# application using Windows Forms and System.Speech to convert a WAV file to text. I've seen plenty of samples online of how to do this, but none that are very robust. I was hoping to write an application that could parse smaller pieces of a large WAV file using BackgroundWorker threads, but I keep getting the following exception in my threads' DoWork function when it calls engine.Recognize():

"No audio input is supplied to this recognizer. Use the method SetInputToDefaultAudioDevice if a microphone is connected to the system, otherwise use SetInputToWaveFile, SetInputToWaveStream or SetInputToAudioStream to perform speech recognition from pre-recorded audio"

Here is the code in my DoWork() function:

SpeechRecognitionEngine engine = new SpeechRecognitionEngine(new    System.Globalization.CultureInfo("en-US"));
engine.SetInputToWaveFile(fname);
engine.LoadGrammar(new DictationGrammar());
engine.BabbleTimeout = TimeSpan.FromSeconds(10.0);
engine.EndSilenceTimeout = TimeSpan.FromSeconds(10.0);
engine.EndSilenceTimeoutAmbiguous = TimeSpan.FromSeconds(10.0);
engine.InitialSilenceTimeout = TimeSpan.FromSeconds(10.0);

BackgroundWorker w = (BackgroundWorker)sender;
while (true)
{    
RecognitionResult data = engine.Recognize();
if (data == null)
    break;
if (w == null) //our thread died from beneath us
    break;
if (!w.IsBusy) //our thread died from beneath us
    break;
if (w.CancellationPending) //notice to cancel
    break;
w.ReportProgress(0, data.Text);
}

I am launching multiple BackgroundWorker threads that run this code. If i use a single thread, I don't see this problem.

Answer 1

You can try this approach. I tested it for Console and Windows Forms application types.

class Program {
    public static void Main() {
        var r1 = new Recognizer(@"c:\proj\test.wav");
        r1.Completed += (sender, e) => Console.WriteLine(r1.Result.Text);

        var r2 = new Recognizer(@"c:\proj\test.wav");
        r2.Completed += (sender, e) => Console.WriteLine(r2.Result.Text);

        Console.ReadLine();
    }
}

class Recognizer {
    private readonly string _fileName;
    private readonly AsyncOperation _operation;
    private volatile RecognitionResult _result;

    public Recognizer(string fileName) {
        _fileName = fileName;
        _operation = AsyncOperationManager.CreateOperation(null);            
        _result = null;

        var worker = new Action(Run);
        worker.BeginInvoke(delegate(IAsyncResult result) {
            worker.EndInvoke(result);
        }, null);            
    }

    private void Run() {
        try {
            SpeechRecognitionEngine engine = new SpeechRecognitionEngine(new System.Globalization.CultureInfo("en-US"));
            engine.SetInputToWaveFile(_fileName);
            engine.LoadGrammar(new DictationGrammar());
            engine.BabbleTimeout = TimeSpan.FromSeconds(10.0);
            engine.EndSilenceTimeout = TimeSpan.FromSeconds(10.0);
            engine.EndSilenceTimeoutAmbiguous = TimeSpan.FromSeconds(10.0);
            engine.InitialSilenceTimeout = TimeSpan.FromSeconds(10.0);
            _result = engine.Recognize();
        }
        finally {
            _operation.PostOperationCompleted(delegate {
                RaiseCompleted();
            }, null);
        }
    }

    public RecognitionResult Result {
        get { return _result; }
    }

    public event EventHandler Completed;

    protected virtual void OnCompleted(EventArgs e) {
        if (Completed != null)
            Completed(this, e);
    }

    private void RaiseCompleted() {
        OnCompleted(EventArgs.Empty);
    }
}