[C#] Riconoscimento vocale con classi System.Speech

[KaNnA]

Ciao a tutti,
sto provando a fare una piccola applicazione per il riconoscimento vocale utilizzando le librerie System.Speech.Recognition.
Premesso che il linguaggio del computer è Italiano (e quindi non posso attivare le funzionalità integrate di riconoscimento vocale da pannello di controllo), ho trovato, scartabellando per internet, questo codice:

using System;
using System.Speech;
using System.Speech.Recognition;
using System.Threading;
using System.Diagnostics;
using System.Globalization;

namespace SpeechTest
{
    class Program
    {

        static void Main(string[] args)
        {
            EnumerateEngines();
            args = CheckArguments(args);

            CommandListener listener = new CommandListener();
            listener.SetCommands(args);
            listener.Run();
        }

        private static string[] CheckArguments(string[] args)
        {
            if (args.Length == 0)
            {
                Console.WriteLine("Setting up default words. Enter your own words via command line arguments");
                args = new string[] { "music", "word" };
            }

            Console.Write("Now listening for...");
            foreach (string word in args)
            {
                Console.Write(String.Format("'{0}', ", word));
            }
            Console.WriteLine("\r\n(Press a key to stop listening)");
            return args;
        }

        private static void EnumerateEngines()
        {
            Console.WriteLine("The following engines are available");
            foreach (RecognizerInfo config in SpeechRecognitionEngine.InstalledRecognizers())
            {
                Console.WriteLine(config.Name + " " + config.Description);
            }
        }
    }

    class CommandListener
    {
        SpeechRecognitionEngine _speechRecogniser;
        public CommandListener()
        {
            _speechRecogniser = new SpeechRecognitionEngine();
            _speechRecogniser.SetInputToDefaultAudioDevice();

            _speechRecogniser.SpeechDetected += new EventHandler<SpeechDetectedEventArgs>(_speechRecogniser_SpeechDetected);
            _speechRecogniser.SpeechHypothesized += new EventHandler<SpeechHypothesizedEventArgs>(_speechRecogniser_SpeechHypothesized);
            _speechRecogniser.SpeechRecognitionRejected += new EventHandler<SpeechRecognitionRejectedEventArgs>(_speechRecogniser_SpeechRecognitionRejected);
            _speechRecogniser.SpeechRecognized += new EventHandler<SpeechRecognizedEventArgs>(_speechRecogniser_SpeechRecognized);
            _speechRecogniser.AudioStateChanged += new EventHandler<AudioStateChangedEventArgs>(_speechRecogniser_AudioStateChanged);
            _speechRecogniser.AudioSignalProblemOccurred += new EventHandler<AudioSignalProblemOccurredEventArgs>(_speechRecogniser_AudioSignalProblemOccurred);

            NotifyEvent("Loaded default engine - {0}", _speechRecogniser.RecognizerInfo.Name);
        }

        #region Event Handlers
        void _speechRecogniser_AudioSignalProblemOccurred(object sender, AudioSignalProblemOccurredEventArgs e)
        {
            NotifyEvent("Audio problem {0}", e.AudioSignalProblem.ToString("G"));
        }

        void _speechRecogniser_AudioStateChanged(object sender, AudioStateChangedEventArgs e)
        {
            NotifyEvent("Audio state now {0}", e.AudioState.ToString("G"));
        }

        void _speechRecogniser_SpeechRecognized(object sender, SpeechRecognizedEventArgs e)
        {
            NotifyEvent("Recognized '{0}' at {1} seconds", e.Result.Text, e.Result.Audio.AudioPosition);
        }

        void _speechRecogniser_SpeechRecognitionRejected(object sender, SpeechRecognitionRejectedEventArgs e)
        {
            NotifyEvent("RecognitionRejected '{0}' at {1} seconds", e.Result.Text, e.Result.Audio.AudioPosition);
        }

        void _speechRecogniser_SpeechHypothesized(object sender, SpeechHypothesizedEventArgs e)
        {

            NotifyEvent("Hypothesized '{0}' at seconds", e.Result.Text);
        }

        void _speechRecogniser_SpeechDetected(object sender, SpeechDetectedEventArgs e)
        {
            NotifyEvent("Detected speech at {0} seconds", e.AudioPosition.TotalSeconds);
        }
        #endregion

        public void SetCommands(string[] commands)
        {
            GrammarBuilder grammarBuilder = new GrammarBuilder();

            grammarBuilder.Culture = _speechRecogniser.RecognizerInfo.Culture;
            grammarBuilder.Append(new Choices(commands));
            Grammar grammar = new Grammar(grammarBuilder);

            _speechRecogniser.UnloadAllGrammars();
            _speechRecogniser.LoadGrammar(grammar);

            _speechRecogniser.RecognizeAsync(RecognizeMode.Multiple);
        }

        public void Run()
        {
            while (!Console.KeyAvailable)
            {
                Thread.Sleep(100);
            }

            _speechRecogniser.Dispose();
        }

        private static void NotifyEvent(string format, params object[] args)
        {
            Console.WriteLine(string.Format(format, args));
        }
    }
} 

Ok, il problema è, dopo aver aggiunto il riferimento del progetto a System.speech, compilando mi da questo errore su questa linea:

_speechRecogniser.SetInputToDefaultAudioDevice(); 

Errore HRESULT E_FAIL restituito da una chiamata a un componente COM.

Non ne so moltissimo di C#, preferisco altri linguaggi, ma sembra che il problema sia il "defaultAudioDevice". Controllando tra le periferiche,il mio microfono è abilitato e funzionante.
Provando a cambiare la suddetta riga così:

_speechRecogniser.SetInputToNull(); 

Il risutalto è uguale. Il problema quindi non è il microfono.

Cosa diavolo può essere?
Grazie a tutti!