public SoundRecorder(int sampleRate, VoiceActivityDetector vad)
{
this.vad = vad;
if (!AVAudioSession.SharedInstance().Category == AVAudioSession.CategoryRecord &&
!AVAudioSession.SharedInstance().Category == AVAudioSession.CategoryPlayAndRecord)
{
throw new AIServiceException("AVAudioCategory not set to RECORD or PLAY_AND_RECORD. Please set it using AVAudioSession class.");
}
audioStreamDescription = new AudioStreamBasicDescription
{
Format = AudioFormatType.LinearPCM,
FormatFlags =
AudioFormatFlags.LinearPCMIsSignedInteger |
AudioFormatFlags.LinearPCMIsPacked,
SampleRate = sampleRate,
BitsPerChannel = BitsPerChannel,
ChannelsPerFrame = Channels,
BytesPerFrame = (BitsPerChannel / 8) * Channels,
FramesPerPacket = 1,
Reserved = 0,
};
audioStreamDescription.BytesPerPacket = audioStreamDescription.BytesPerFrame * audioStreamDescription.FramesPerPacket;
inputQueue = CreateInputQueue(audioStreamDescription);
}
static void ProcessStream(VoiceActivityDetector vad, Stream inputStream)
{
var bufferSize = 3000 + new Random().Next(200);
var buffer = new byte[bufferSize];
int bytesRead = 0;
bytesRead = inputStream.Read(buffer, 0, bufferSize);
while (bytesRead > 0)
{
vad.ProcessBufferEx(buffer, bytesRead);
bufferSize = 3000 + new Random().Next(200);
buffer = new byte[bufferSize];
bytesRead = inputStream.Read(buffer, 0, bufferSize);
}
}
private void buttonStartTest_Click(object sender, RoutedEventArgs e)
{
try
{
// speexPreprocessor = new SpeexPreprocessFilter(AudioConstants.SamplesPerFrame, sampleRate, new MediaConfig(), null, "VAD test page Speex Preprocessor");
voiceActivityDetector = new VoiceActivityDetector(AudioFormat.Default, VoiceActivityDetector.Aggressiveness.Normal);
// DataContext = voiceActivityDetector. speexPreprocessor.GetVadViewModel(this.Dispatcher);
InitializeCaptureSource();
//buttonStartTest.IsEnabled = false;
}
catch (Exception ex)
{
ClientLogger.Debug(ex.ToString());
MessageBox.Show(ex.ToString());
}
}
public void TestEnabled()
{
var vad = new VoiceActivityDetector(SAMPLE_RATE);
vad.Enabled = false;
var inputStream = ReadFileFromResource("speech.raw");
speechEnd = false;
speechBegin = false;
vad.SpeechBegin += () => { speechBegin = true; };
vad.SpeechEnd += () => { speechEnd = true; };
ProcessStream(vad, inputStream);
Assert.IsTrue(speechBegin);
Assert.IsFalse(speechEnd);
}
public void TestSpeechDetect()
{
var vad = new VoiceActivityDetector(SAMPLE_RATE);
var inputStream = ReadFileFromResource("speech.raw");
speechEnd = false;
speechBegin = false;
speechNotDetected = false;
vad.SpeechBegin += () => { speechBegin = true; };
vad.SpeechEnd += () => { speechEnd = true; };
vad.SpeechNotDetected += () => { speechNotDetected = true; };
ProcessStream(vad, inputStream);
Assert.IsTrue(speechBegin);
Assert.IsTrue(speechEnd);
Assert.IsFalse(speechNotDetected);
Assert.IsTrue(vad.SpeechBeginTime > 800 && vad.SpeechBeginTime < 950, "vad.SpeechBeginTime " + vad.SpeechBeginTime);
Assert.IsTrue(vad.SpeechEndTime > 2400 && vad.SpeechEndTime < 2900, "vad.SpeechEndTime " + vad.SpeechEndTime);
}
public DtxFilter(AudioFormat audioFormat)
{
_audioFormat = audioFormat;
_vad = new VoiceActivityDetector(audioFormat, VoiceActivityDetector.Aggressiveness.Normal);
}
internal AudioStream(AudioRecord record, VoiceActivityDetector vad)
{
this.vad = vad;
audioRecord = record;
}
