问题
I finally have built a program to listen to the internal audio loopback using NAudio, and output recognized text. The problem is it listens, and always says, eg:
Recognized text: had
Recognized text: had
Recognized text: had
Recognized text: had
Recognized text: had had phone Le K add phone Laton
Recognized text: had phone looked had phone looked had phone looked had phone lo
oked zone
Recognized text: had phone lines to had, had phone looked had phone looked had p
hone line had phone
Recognized text: had phone line had phone looked had phone
Recognized text: had phone looked had phone looked had phone line had phone
Recognized text: had phone looked had phone look to had pot they had phone lit o
nly had phone
Recognized text: had phone line had phone looked had phone line to had to had ph
one
Recognized text: had phone line had phone looked had phone looked had phone
Recognized text: had phone line had phone looked had phone looked had phone line
10 only T had phone
Recognized text: had phone line had
Recognized text: had phone line had phone looked had phone line had
Recognized text: had phone Le tone looked had
Recognized text: had phone looked had phone looked had phone
Recognized text: had phone line had phone line had phone licked had phone
Recognized text: had phone lines to had popped the own
and similar nonsense, but even when I pause audio it just shows "Recognized text: had" or "an" again and again and again. When I unpause audio it keeps unsuccessfully recognizing the internal audio. Is there a way I can fix this, or at least get a wav of what it's trying to send to the Microsoft speech recognition recognizer?
using System;
using System.Speech.Recognition;
using NAudio.Wave;
using NAudio.CoreAudioApi.Interfaces;
using NAudio.CoreAudioApi;
using System.IO;
using System.Speech.AudioFormat;
using NAudio.Wave.SampleProviders;
using NAudio.Utils;
using System.Threading;
using System.Collections.Generic;
namespace SpeechRecognitionApp
{
class SpeechStreamer : Stream
{
private AutoResetEvent _writeEvent;
private List<byte> _buffer;
private int _buffersize;
private int _readposition;
private int _writeposition;
private bool _reset;
public SpeechStreamer(int bufferSize)
{
_writeEvent = new AutoResetEvent(false);
_buffersize = bufferSize;
_buffer = new List<byte>(_buffersize);
for (int i = 0; i < _buffersize; i++)
_buffer.Add(new byte());
_readposition = 0;
_writeposition = 0;
}
public override bool CanRead
{
get { return true; }
}
public override bool CanSeek
{
get { return false; }
}
public override bool CanWrite
{
get { return true; }
}
public override long Length
{
get { return -1L; }
}
public override long Position
{
get { return 0L; }
set { }
}
public override long Seek(long offset, SeekOrigin origin)
{
return 0L;
}
public override void SetLength(long value)
{
}
public override int Read(byte[] buffer, int offset, int count)
{
int i = 0;
while (i < count && _writeEvent != null)
{
if (!_reset && _readposition >= _writeposition)
{
_writeEvent.WaitOne(100, true);
continue;
}
buffer[i] = _buffer[_readposition + offset];
_readposition++;
if (_readposition == _buffersize)
{
_readposition = 0;
_reset = false;
}
i++;
}
return count;
}
public override void Write(byte[] buffer, int offset, int count)
{
for (int i = offset; i < offset + count; i++)
{
_buffer[_writeposition] = buffer[i];
_writeposition++;
if (_writeposition == _buffersize)
{
_writeposition = 0;
_reset = true;
}
}
_writeEvent.Set();
}
public override void Close()
{
_writeEvent.Close();
_writeEvent = null;
base.Close();
}
public override void Flush()
{
}
}
class FakeStreamer : Stream
{
public bool bExit = false;
Stream stream;
Stream client;
public FakeStreamer(Stream client)
{
this.client = client;
this.stream = client;
}
public override bool CanRead
{
get { return stream.CanRead; }
}
public override bool CanSeek
{
get { return false; }
}
public override bool CanWrite
{
get { return stream.CanWrite; }
}
public override long Length
{
get { return -1L; }
}
public override long Position
{
get { return 0L; }
set { }
}
public override long Seek(long offset, SeekOrigin origin)
{
return 0L;
}
public override void SetLength(long value)
{
stream.SetLength(value);
}
public override int Read(byte[] buffer, int offset, int count)
{
int len = 0, c = count;
while (c > 0 && !bExit)
{
//try {
len = stream.Read(buffer, offset, c);
/*}
catch (Exception e)
{
Console.WriteLine("ouch");
}
if (!client.Connected || len == 0)
{
//Exit read loop
return 0;
}*/
offset += len;
c -= len;
}
return count;
}
public override void Write(byte[] buffer, int offset, int count)
{
stream.Write(buffer, offset, count);
}
public override void Close()
{
stream.Close();
base.Close();
}
public override void Flush()
{
stream.Flush();
}
}
class Program
{
static void Main(string[] args)
{
// Create an in-process speech recognizer for the en-US locale.
using (
SpeechRecognitionEngine recognizer =
new SpeechRecognitionEngine(
new System.Globalization.CultureInfo("en-US")))
{
// Create and load a dictation grammar.
recognizer.LoadGrammar(new DictationGrammar());
// Add a handler for the speech recognized event.
recognizer.SpeechRecognized +=
new EventHandler<SpeechRecognizedEventArgs>(recognizer_SpeechRecognized);
// Configure input to the speech recognizer.
//recognizer.SetInputToDefaultAudioDevice();
WasapiLoopbackCapture capture = new WasapiLoopbackCapture();
BufferedWaveProvider WaveBuffer = new BufferedWaveProvider(capture.WaveFormat);
WaveBuffer.DiscardOnBufferOverflow = true;
//WaveBuffer.ReadFully = false;
WaveToSampleProvider sampleStream = new WaveToSampleProvider(WaveBuffer);
StereoToMonoSampleProvider monoStream = new StereoToMonoSampleProvider(sampleStream)
{
LeftVolume = 1f,
RightVolume = 1f
};
//Downsample to 8000 https://stackoverflow.com/questions/48233099/capture-audio-from-wasapiloopbackcapture-and-convert-to-mulaw
WdlResamplingSampleProvider resamplingProvider = new WdlResamplingSampleProvider(monoStream, 16000);
SampleToWaveProvider16 ieeeToPcm = new SampleToWaveProvider16(resamplingProvider);
var arr = new byte[128];
Stream captureConvertStream = new System.IO.MemoryStream();
capture.StartRecording();
//outputStream = new MuLawConversionProvider(ieeeToPcm);
Stream captureStream = new System.IO.MemoryStream();
//Stream buffStream = new FakeStreamer(captureStream);
capture.DataAvailable += (s, a) =>
{
//It is getting here.
//captureStream.Write(a.Buffer, 0, a.BytesRecorded);
WaveBuffer.AddSamples(a.Buffer, 0, a.BytesRecorded);
};
Console.WriteLine(capture.WaveFormat.AverageBytesPerSecond);
Console.WriteLine(capture.WaveFormat.BitsPerSample);
//var newFormat = new WaveFormat(8000, 16, 1);
//using (var conversionStream = new WaveFormatConversionStream(newFormat, capture)
//capture.StartRecording();
//using (var resampler = new MediaFoundationResampler(new NAudio.Wave.RawSourceWaveStream(captureStream, capture.WaveFormat), newFormat))
//{
//resampler.ResamplerQuality = 60;
//WaveFileWriter.WriteWavFileToStream(captureConvertStream, resampler);
//recognizer.SetInputToDefaultAudioDevice();
//Stream buffStream = new FakeStreamer(captureConvertStream);
Stream buffStream = new SpeechStreamer(2048);
//recognizer.SetInputToWaveStream(buffStream);
recognizer.SetInputToAudioStream(buffStream, new SpeechAudioFormatInfo(
16000, AudioBitsPerSample.Eight, AudioChannel.Mono));
// Start asynchronous, continuous speech recognition.
recognizer.RecognizeAsync(RecognizeMode.Multiple);
/*System.Threading.Thread.Sleep(5000);
works when playing anything
var floata = new float[128];
while(monoStream.Read(floata, 0, floata.Length) > 0 )
{
Console.WriteLine(arr.Length);
}*/
while (ieeeToPcm.Read(arr, 0, arr.Length) > 0)
{
//Console.Write("Writing PCM ");
//Console.WriteLine(arr.Length);
//captureConvertStream.Write(arr, 0, arr.Length);
buffStream.Write(arr, 0, arr.Length);
}
Console.WriteLine("end");
/*capture.StartRecording();
//Never getting to the resampler, the read is always zero!? even if waiting 5s for the audio to buffer.
System.Threading.Thread.Sleep(5000);
var arr = new byte[128];
while (resampler.Read(arr, 0, arr.Length) > 0)
{
captureConvertStream.Write(arr, 0, arr.Length);
Console.WriteLine("Never getting here");
}
// Keep the console window open.
while (true)
{
Console.ReadLine();
}*/
//}
}
}
// Handle the SpeechRecognized event.
static void recognizer_SpeechRecognized(object sender, SpeechRecognizedEventArgs e)
{
Console.WriteLine("Recognized text: " + e.Result.Text);
}
}
}
回答1:
That SpeechStreamer class has some problems, I cannot really see what its purpose is. I tried. Also looking at wavefile dumps from your implementation, the audio is really choppy, with long pauses between the samples. This might be what is throwing the speech recognizer off. This is an example: Windows Volume Adjutment Sound From Your Code
As you may hear, it is pretty choppy with a lot of silence between. The Voice Recognition part recognizes this as : "ta ta ta ta ta ta..."
I had to rewrite your code a bit to dump a wave file, since the Read method of your SpeechStream causes an eternal loop when used to read its contents.
To dump a wave file you could do the following:
var buffer = new byte[2048];
using (var writer = new WaveFileWriter("tmp.wav", ieeeToPcm.WaveFormat))
{
//buffStream is changed to a MemoryStream for this to work.
buffStream.Seek(0,SeekOrigin.Begin);
while (buffStream.Read(buffer, 0, buffer.Length)>0)
{
writer.Write(buffer, 0, buffer.Length);
}
}
Or you can do it when you read from your SampleToWaveProvider16:
var writer = new WaveFileWriter("dump.wav", ieeeToPcm.WaveFormat);
while (ieeeToPcm.Read(arr, 0, arr.Length) > 0)
{
if (Console.KeyAvailable && Console.ReadKey().Key == ConsoleKey.Escape)
break;
buffStream.Write(arr, 0, arr.Length);
writer.Write(arr, 0, arr.Length);
}
I just added the ability to hit Escape to exit the loop.
Now I do wonder why you are using NAudio? Why not use the methods native to the Sound.Speech API?
class Program
{
private static ManualResetEvent _done;
static void Main(string[] args)
{
_done = new ManualResetEvent(false);
using (SpeechRecognitionEngine recognizer = new SpeechRecognitionEngine(new CultureInfo("en-US")))
{
recognizer.LoadGrammar(new DictationGrammar());
recognizer.SpeechRecognized += RecognizedSpeech;
recognizer.SetInputToDefaultAudioDevice();
recognizer.RecognizeAsync(RecognizeMode.Multiple);
_done.WaitOne();
}
}
private static void RecognizedSpeech(object sender, SpeechRecognizedEventArgs e)
{
if (e.Result.Text.Contains("exit"))
{
_done.Set();
}
Console.WriteLine(e.Result.Text);
}
}
来源：https://stackoverflow.com/questions/58678228/wasapiloopbackcapture-internal-audio-recognition-gives-jibberish-and-text-when-n