I am currently developing a project of a DialerService. One of the functions is to interpret the recorded .wav media files into plaint text. I used the SpeechRecognitionEngine trying to interpret the contents, and I got some results that are not accurate or, sometimes broken sentences that doesn't make any sense.
The .wav files are the recorded files from a telephone conversation between two or more clients, the file I tested is a very simple and short conversation I made with my colleague.
So my question is how can I improve the accuracy of the interpretation and what to do to improve my code for this purpose? I know adding the grammar will help recognize some keywords, but what I need is to generally interpret contents that I recorded from users.
Here blow is my working code:
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.IO;
using System.Threading.Tasks;
using System.Speech.Recognition;
using System.Speech.AudioFormat;
using System.Web;
namespace VoiceRecognition
{
class Program
{
static bool completed;
static void Main(string[] args)
{
using (
SpeechRecognitionEngine recognizer =
new SpeechRecognitionEngine(
new System.Globalization.CultureInfo("en-US")))
{
// Create and load a grammar.
Grammar dictation = new DictationGrammar();
dictation.Name = "Dictation Grammar";
recognizer.LoadGrammar(new DictationGrammar());
recognizer.SetInputToWaveFile(@"C:\Projects2\VoiceRecognition2\conf_with_vincent_1.wav");
// Attach event handlers for the results of recognition.
//recognizer.AudioLevelUpdated += new EventHandler<AudioLevelUpdatedEventArgs>(recognizer_AudioLevelUpdated);
//recognizer.AudioStateChanged += new EventHandler<AudioStateChangedEventArgs>(recognizer_AudioStateChanged);
recognizer.SpeechRecognized += new EventHandler<SpeechRecognizedEventArgs>(recognizer_SpeechRecognized);
recognizer.RecognizeCompleted += new EventHandler<RecognizeCompletedEventArgs>(recognizer_RecognizeCompleted);
// Perform recognition on the entire file.
Console.WriteLine("Starting asynchronous recognition...");
completed = false;
//recognizer.RecognizeAsync();
recognizer.RecognizeAsync(RecognizeMode.Multiple);
// Keep the console window open.
while (!completed)
{
Console.ReadLine();
}
Console.WriteLine("Done.");
}
Console.WriteLine();
Console.WriteLine("Press any key to exit...");
Console.ReadKey();
}
// Handle the Audio state event.
static void recognizer_AudioStateChanged(object sender, AudioStateChangedEventArgs e)
{
Console.WriteLine("The new audio state is: " + e.AudioState);
}
static void recognizer_AudioLevelUpdated(object sender, AudioLevelUpdatedEventArgs e)
{
Console.WriteLine("The audio level is now: {0}.", e.AudioLevel);
}
// Handle the SpeechRecognized event.
static void recognizer_SpeechRecognized(object sender, SpeechRecognizedEventArgs e)
{
if (e.Result != null && e.Result.Text != null)
{
Console.WriteLine(" Recognized text = {0}", e.Result.Text);
}
else
{
Console.WriteLine(" Recognized text not available.");
}
}
// Handle the RecognizeCompleted event.
static void recognizer_RecognizeCompleted(object sender, RecognizeCompletedEventArgs e)
{
if (e.Error != null)
{
Console.WriteLine(" Error encountered, {0}: {1}",
e.Error.GetType().Name, e.Error.Message);
}
if (e.Cancelled)
{
Console.WriteLine(" Operation cancelled.");
}
if (e.InputStreamEnded)
{
Console.WriteLine(" End of stream encountered.");
}
completed = true;
}
}
}
Another class is:
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Speech.Recognition;
public class SpeechReconizer
{
SpeechRecognitionEngine _speechRecognitionEngine;
public SpeechReconitionResult ReadResult { get; set; }
public SpeechReconizer()
{
Grammar dictation = new DictationGrammar();
dictation.Name = "Dictation Grammar";
_speechRecognitionEngine = new SpeechRecognitionEngine();
_speechRecognitionEngine.SetInputToDefaultAudioDevice();
_speechRecognitionEngine.LoadGrammar(dictation);
_speechRecognitionEngine.InitialSilenceTimeout = TimeSpan.FromSeconds(3);
_speechRecognitionEngine.BabbleTimeout = TimeSpan.FromSeconds(2);
_speechRecognitionEngine.EndSilenceTimeout = TimeSpan.FromSeconds(1);
_speechRecognitionEngine.EndSilenceTimeoutAmbiguous = TimeSpan.FromSeconds(1.5);
_speechRecognitionEngine.RecognizeAsync(RecognizeMode.Multiple);
_speechRecognitionEngine.SpeechRecognized += RecognizerSpeechRecognized;
_speechRecognitionEngine.RecognizeCompleted += RecognizerRecognizeCompleted;
}
public SpeechReconitionResult ReadSpeech(string sourceAudio)
{
ReadResult = new SpeechReconitionResult();
_speechRecognitionEngine.SetInputToWaveFile(sourceAudio);
_speechRecognitionEngine.Recognize();
return ReadResult;
}
private void RecognizerSpeechRecognized(object sender, SpeechRecognizedEventArgs e)
{
if (e.Result != null && e.Result.Text != null)
{
ReadResult.Success = true;
ReadResult.Text = e.Result.Text;
}
else
{
ReadResult.Text = "Recognized text not available.";
}
}
private void RecognizerRecognizeCompleted(object sender, RecognizeCompletedEventArgs e)
{
if (e.Error != null)
{
ReadResult.Success = false;
ReadResult.ErrorMessage = string.Format("{0}: {1}",
e.Error.GetType().Name, e.Error.Message);
}
if (e.Cancelled)
{
ReadResult.Success = false;
ReadResult.ErrorMessage = "Operation cancelled.";
}
}
}
public class SpeechReconitionResult
{
public string Text { get; set; }
public bool Success { get; set; }
public string ErrorMessage { get; set; }
public bool Complete { get; set; }
}
The test result is(in console):
Starting asynchronous recognition...
Recognized text = Helence and the globe or east
Recognized text = alarmed
Recognized text = and client thanks
Recognized text = what aren't going to do and that they
Recognized text = aren't goint to rule
Recognized text = working to dear E
Recognized text = N
Recognized text = at dinner
Recognized text = and
Recognized text = that going there
Recognized text = and you have a 98 no problem bars
End of stream encountered.
What the actual content is: -Hello Vincent. -Hello Boris. -How are you? -I am fine. -What are you going to do today? -I am going to watch TV, have dinner and go home. -Thank you, have a nice day. -No problem.
