public int Recognize(ISoundSignalReader signal, HiddenMarkovClassifier hmm, out string name,
SignalVisitor voiceVisitor = null)
{
var featureUtility = new FeatureUtility(_engineParameters);
signal.Reset();
var features = featureUtility.ExtractFeatures(signal, voiceVisitor).First();
var observations = _codeBook.Quantize(features.Select(item => new Point(item)).ToArray());
double[] responsabilities;
var ret = hmm.Compute(observations, out responsabilities);
var models = hmm.Models;
var likelyHoodValue = Double.MinValue;
name = string.Empty;
foreach (var model in models)
{
var val = model.Evaluate(observations);
if (val > likelyHoodValue)
{
likelyHoodValue = val;
name = model.Tag.ToString();
}
}
return ret;
}
public int Recognize(ISoundSignalReader signal, HiddenMarkovModel[] models, out string name,
SignalVisitor voiceVisitor = null)
{
var featureUtility = new FeatureUtility(_engineParameters);
signal.Reset();
var features = featureUtility.ExtractFeatures(signal, voiceVisitor).First();
var observations = _codeBook.Quantize(features.Select(item => new Point(item)).ToArray());
var likelyHoodValue = Double.MinValue;
name = string.Empty;
var index = 0;
var ret = 0;
foreach (var model in models)
{
var val = model.Evaluate(observations);
if (val > likelyHoodValue)
{
likelyHoodValue = val;
name = model.Tag.ToString();
ret = index;
}
index++;
}
return ret;
}
public VoiceActivityDetection(ISoundSignalReader signal, int frameSize, int emptyFrames = 3,
VoiceActivationDetectionEnhancement enhancements = VoiceActivationDetectionEnhancement.All)
{
_signal = signal;
_frameSize = frameSize;
_emptyFrames = emptyFrames;
_enhancements = enhancements;
Init();
}
public VoiceActivitySignalReader(ISoundSignalReader signal, int frameSize, int emptyFrames = 3,
VoiceActivationDetectionEnhancement enhancements = VoiceActivationDetectionEnhancement.All)
{
_voiceActivityDetection = new VoiceActivityDetection(signal, frameSize, emptyFrames, enhancements);
signal.Reset();
_soundSignal = signal;
_properties = new Dictionary<string, object> {{VoiceProperty, false}};
SupportedPropertiesSet.Add(VoiceProperty);
}
public void ExtractFeaturesAsync(ISoundSignalReader signal, Action<List<double[]>> action,
SignalVisitor voiceVisitor = null)
{
Action<List<double[]>> addfeatures = features =>
{
action.BeginInvoke(features, null, null);
};
ExtractFeaturesInternalUsingVad(signal, addfeatures, voiceVisitor);
}
public IEnumerable<List<double[]>> ExtractFeatures(ISoundSignalReader signal, SignalVisitor voiceVisitor = null)
{
List<List<Double[]>> allObservables = new List<List<Double[]>>();
Action<List<double[]>> addfeatures = features =>
{
allObservables.Add(features);
};
ExtractFeaturesInternalUsingVad(signal, addfeatures, voiceVisitor);
return allObservables;
}
public static IFeatureProvider GetProvider(FeatureProviderParameters parameters, ISoundSignalReader signal)
{
var lpcParam = parameters as LpcCepstrumParameters;
if (lpcParam!= null)
{
return new LpcCepstrumProvider(lpcParam.LpcOrder, lpcParam.NumberOfCoeff);
}
var melParam = parameters as MellFreqProviderParameters;
if (melParam != null)
{
//var mfcc = new MFCC(NumberOfFilterBanks, samplingRate: signal.SampleRate, lowerfrequency: 50);
return new MellFrequencyCepstrumProvider(melParam.NumberOfFilterBanks, samplingRate: signal.SampleRate, lowerfrequency: 50, numberOfCoef: melParam.NumberOfExtractedMelItems + 1);
}
return null;
}
public void RecognizeAsync(ISoundSignalReader signal, HiddenMarkovClassifier hmm, Action<string> handleMessage,
SignalVisitor voiceVisitor = null)
{
Action<List<double[]>> action = features =>
{
var observations = _codeBook.Quantize(features.Select(item => new Point(item)).ToArray());
double[] responsabilities;
var ret = hmm.Compute(observations, out responsabilities);
var models = hmm.Models;
var likelyHoodValue = Double.MinValue;
foreach (var model in models)
{
var val = model.Evaluate(observations);
if (val > likelyHoodValue)
{
likelyHoodValue = val;
}
}
handleMessage(hmm[ret].Tag.ToString());
};
var featureUtility = new FeatureUtility(_engineParameters);
featureUtility.ExtractFeaturesAsync(signal, action, voiceVisitor);
}
public PreemphasisFilter(ISoundSignalReader signalReader, float filterRatio)
{
_signalReader = signalReader;
_filterRatio = filterRatio;
Length = signalReader.Length;
}
public void RecognizeAsync(ISoundSignalReader signal, Action<string> handleMessage,
SignalVisitor visitor = null)
{
Action<List<double[]>> action = features =>
{
var observations = _codeBook.Quantize(features.Select(item => new Point(item)).ToArray());
var likelyHoodValue = Double.MinValue;
HiddenMarkovModel bestFit = null;
var modelsToSearchFor = _models.Values.Where(item =>
{
var idProp = (IdentificationProperties) item.Tag;
var rateLength = Math.Abs(idProp.MeanFeaturesLength - observations.Length)/idProp.MeanFeaturesLength;
return rateLength < 0.1;
}).ToArray();
foreach (var model in modelsToSearchFor)
{
var val = model.Evaluate(observations);
if (val > likelyHoodValue)
{
likelyHoodValue = val;
bestFit = model;
}
}
if (bestFit != null)
{
var idProp = (IdentificationProperties)bestFit.Tag;
handleMessage(idProp.Label);
}
};
var featureUtility = new FeatureUtility(_engineParameters);
featureUtility.ExtractFeaturesAsync(signal, action, visitor);
}
public int Recognize(ISoundSignalReader signal, out string name,
SignalVisitor visitor = null)
{
var featureUtility = new FeatureUtility(_engineParameters);
signal.Reset();
var features = featureUtility.ExtractFeatures(signal, visitor).First();
var observations = _codeBook.Quantize(features.Select(item => new Point(item)).ToArray());
var likelyHoodValue = Double.MinValue;
HiddenMarkovModel bestFit = null;
var index = 0;
var ret = 0;
foreach (var model in _models.Values)
{
var val = model.Evaluate(observations);
if (val > likelyHoodValue)
{
likelyHoodValue = val;
bestFit = model;
ret = index;
}
index++;
}
if (bestFit != null)
{
var idProp = (IdentificationProperties) bestFit.Tag;
name = idProp.Label;
}
else
name = string.Empty;
return ret;
}
private void ExtractFeaturesInternalUsingVad(ISoundSignalReader signal, Action<List<double[]>> featureExtracted,
SignalVisitor voiceVisitor)
{
var featureProvider = FeaturesProviderFactory.GetProvider(ProviderParameters, signal);
var frameSize = (int) Math.Floor(signal.SampleRate*FrameSizeMiliseconds/1000.0);
var stepSize = (int) Math.Floor(signal.SampleRate*StepSizeMiliseconds/1000.0);
var filteredSignal = new PreemphasisFilter(signal, 0.95f);
float[] frame;
var voiceStream = new VoiceActivitySignalReader(filteredSignal, frameSize, 8);
if (voiceVisitor != null)
{
voiceStream.Accept(voiceVisitor);
}
int index = 0, silentSamples = 0, noOfItems = ProviderParameters.NumberOfCoeff - 1;
var observables = new List<double[]>();
bool isVoice;
while (voiceStream.Read(frameSize, stepSize, out frame, out isVoice))
{
if (isVoice)
{
bool isEmpty;
var features = featureProvider.Extract(frame, out isEmpty);
silentSamples = 0;
observables.Add(features);
if (featureProvider.ComputeDelta)
{
ComputeDelta(observables, index - 1, 1, noOfItems);
ComputeDelta(observables, index - 2, noOfItems + 1, noOfItems, true);
}
index++;
}
else if (observables.Count > 0 && silentSamples > SilenceThreshHold)
{
if (index >= MinWordLength)
{
if (featureProvider.ComputeDelta)
{
ComputeDelta(observables, index - 1, 1, noOfItems);
ComputeDelta(observables, index - 2, noOfItems + 1, noOfItems, true);
ComputeDelta(observables, index - 1, noOfItems + 1, noOfItems, true);
}
featureExtracted(observables);
}
observables = new List<double[]>();
index = 0;
}
else
{
silentSamples++;
}
}
}
