Add an Interface for Pitchtrackers and set VolTresholds

PitchTracker/PitchWrapper.cpp

@@ -39,17 +39,17 @@ void Analyzer_Process(Analyzer* analyzer){
     analyzer->process();
 }
 
-double Analyzer_GetPeak(Analyzer* analyzer){
+float Analyzer_GetPeak(Analyzer* analyzer){
 	if(!analyzer)
 		return -999;
-	return analyzer->getPeak();   
+	return static_cast<float>(analyzer->getPeakRaw());
 }
 
 double Analyzer_FindNote(Analyzer* analyzer, double minFreq, double maxFreq){
 	if(!analyzer)
 		return -1;
 	const Tone* tone = analyzer->findTone(minFreq, maxFreq);
-	return (tone == NULL) ? -1 : FreqToNote(tone->freq);    
+	return (tone == NULL) ? -1 : FreqToNote(tone->freq);
 }
 
 bool Analyzer_OutputFloat(Analyzer* analyzer, float* data, int sampleCt, float rate){
@@ -71,6 +71,18 @@ int PtAKF_GetNumHalfTones(){
 	return PtAKF::GetNumHalfTones();
 }
 
+void PtAKF_SetVolumeTreshold(PtAKF* analyzer, float threshold){
+	if(!analyzer)
+		return;
+	analyzer->SetVolumeThreshold(threshold);
+}
+
+float PtAKF_GetVolumeThreshold(PtAKF* analyzer){
+	if(!analyzer)
+		return 0.f;
+	return analyzer->GetVolumeThreshold();
+}
+
 void PtAKF_InputByte(PtAKF* analyzer, char* data, int sampleCt){
 	if(sampleCt <= 0 || !analyzer)
 		return;
@@ -82,7 +94,7 @@ void PtAKF_InputByte(PtAKF* analyzer, char* data, int sampleCt){
 	//analyzer->input(dataShort, dataShort + sampleCt);
 }
 
-int PtAKF_GetNote(PtAKF* analyzer, double* maxVolume, float* weights){
+int PtAKF_GetNote(PtAKF* analyzer, float* maxVolume, float* weights){
 	if(!analyzer)
 		return -1;
 	return analyzer->GetNote(maxVolume, weights);
@@ -97,6 +109,18 @@ void PtDyWa_Free(PtDyWa* analyzer){
 		delete analyzer;
 }
 
+void PtDyWa_SetVolumeTreshold(PtDyWa* analyzer, float threshold){
+	if(!analyzer)
+		return;
+	analyzer->SetVolumeThreshold(threshold);
+}
+
+float PtDyWa_GetVolumeThreshold(PtDyWa* analyzer){
+	if(!analyzer)
+		return 0.f;
+	return analyzer->GetVolumeThreshold();
+}
+
 void PtDyWa_InputByte(PtDyWa* analyzer, char* data, int sampleCt){
 	if(sampleCt <= 0 || !analyzer)
 		return;
@@ -105,16 +129,10 @@ void PtDyWa_InputByte(PtDyWa* analyzer, char* data, int sampleCt){
 	freeFloatArray(dataFloat);
 }
 
-void PtDyWa_Process(PtDyWa* analyzer){
-	if(!analyzer)
-		return;
-	analyzer->Process();
-}
-
-double PtDyWa_GetNote(PtDyWa* analyzer){
+double PtDyWa_FindNote(PtDyWa* analyzer, float* maxVolume){
 	if(!analyzer)
 		return -1;
-	return analyzer->GetNote();
+	return analyzer->FindNote(maxVolume);
 }
 
 /*

PitchTracker/PitchWrapper.h

@@ -12,21 +12,24 @@ DllExport void Analyzer_InputFloat(Analyzer* analyzer, float* data, int sampleCt
 DllExport void Analyzer_InputShort(Analyzer* analyzer, short* data, int sampleCt);
 DllExport void Analyzer_InputByte(Analyzer* analyzer, char* data, int sampleCt);
 DllExport void Analyzer_Process(Analyzer* analyzer);
-DllExport double Analyzer_GetPeak(Analyzer* analyzer);
+DllExport float Analyzer_GetPeak(Analyzer* analyzer);
 DllExport double Analyzer_FindNote(Analyzer* analyzer, double minFreq, double maxFreq);
 DllExport bool Analyzer_OutputFloat(Analyzer* analyzer, float* data, int sampleCt, float rate);
 
 DllExport PtAKF* PtAKF_Create(unsigned step);
 DllExport void PtAKF_Free(PtAKF* analyzer);
 DllExport int PtAKF_GetNumHalfTones();
+DllExport void PtAKF_SetVolumeThreshold(PtAKF* analyzer, float threshold);
+DllExport float PtAKF_GetVolumeThreshold(PtAKF* analyzer);
 DllExport void PtAKF_InputByte(PtAKF* analyzer, char* data, int sampleCt);
-DllExport int PtAKF_GetNote(PtAKF* analyzer, double* maxVolume, float* weights);
+DllExport int PtAKF_GetNote(PtAKF* analyzer, float* maxVolume, float* weights);
 
 DllExport PtDyWa* PtDyWa_Create(unsigned step);
 DllExport void PtDyWa_Free(PtDyWa* analyzer);
+DllExport void PtDyWa_SetVolumeTreshold(PtDyWa* analyzer,float threshold);
+DllExport float PtDyWa_GetVolumeThreshold(PtDyWa* analyzer);
 DllExport void PtDyWa_InputByte(PtDyWa* analyzer, char* data, int sampleCt);
-DllExport void PtDyWa_Process(PtDyWa* analyzer);
-DllExport double PtDyWa_GetNote(PtDyWa* analyzer);
+DllExport double PtDyWa_FindNote(PtDyWa* analyzer);
 
 /*
 public ref class CTone {

PitchTracker/dywapitchtrack/dywapitchtrack.cpp

@@ -102,7 +102,7 @@ typedef struct _minmax {
 	struct _minmax *next;
 } minmax;
 
-double _dywapitch_computeWaveletPitch(double * samples, int startsample, int samplecount) {
+double _dywapitch_computeWaveletPitch(double * samples, int startsample, int samplecount, float* maxVolume, double volThreshold) {
 	double pitchF = 0.0;
 	
 	int i, j;
@@ -143,6 +143,9 @@ double _dywapitch_computeWaveletPitch(double * samples, int startsample, int sam
 		maxValue = maxValue - theDC;
 		minValue = minValue - theDC;
 		double amplitudeMax = (maxValue > -minValue ? maxValue : -minValue);
+		*maxVolume = static_cast<float>(amplitudeMax + theDC);
+		if(*maxVolume < volThreshold)
+			return 0.0;
 		
 		ampltitudeThreshold = amplitudeMax*maximaThresholdRatio;
 		//asLog("dywapitch theDC=%f ampltitudeThreshold=%f\n", theDC, ampltitudeThreshold);
@@ -439,8 +442,8 @@ void dywapitch_inittracking(dywapitchtracker *pitchtracker) {
 	pitchtracker->_pitchConfidence = -1;
 }
 
-double dywapitch_computepitch(dywapitchtracker *pitchtracker, double * samples, int startsample, int samplecount) {
-	double raw_pitch = _dywapitch_computeWaveletPitch(samples, startsample, samplecount);
+double dywapitch_computepitch(dywapitchtracker *pitchtracker, double * samples, int startsample, int samplecount, float* maxVolume, double volThreshold) {
+	double raw_pitch = _dywapitch_computeWaveletPitch(samples, startsample, samplecount, maxVolume, volThreshold);
 	return _dywapitch_dynamicprocess(pitchtracker, raw_pitch);
 }
 

PitchTracker/dywapitchtrack/dywapitchtrack.h

@@ -100,7 +100,7 @@ void dywapitch_inittracking(dywapitchtracker *pitchtracker);
 // startsample : the index of the first sample to use in the sample buffer
 // samplecount : the number of samples to use to compte the pitch
 // return 0.0 if no pitch was found (sound too low, noise, etc..)
-double dywapitch_computepitch(dywapitchtracker *pitchtracker, double * samples, int startsample, int samplecount);
+double dywapitch_computepitch(dywapitchtracker *pitchtracker, double * samples, int startsample, int samplecount, float* maxVolume, double volThreshold);
 
 #ifdef __cplusplus
 } // extern "C"

PitchTracker/dywapitchtrack/ptDyWa.cpp

@@ -5,23 +5,27 @@
 PtDyWa::PtDyWa(unsigned step){
 	_Step = step;
 	dywapitch_inittracking(&_State);
+	_VolTreshold = 0.01f;
+	_LastMaxVol = 0.f;
 }
 
-void PtDyWa::Process(){
+void PtDyWa::SetVolumeThreshold(float threshold){
+	_VolTreshold = threshold;
+}
+
+double PtDyWa::FindNote(float* maxVolume){
 	double AnaylsisBuf[_SampleCt];
 	size_t size = _AnalysisBuf.size();
 	if(size > _SampleCt)
 		_AnalysisBuf.pop(size - _SampleCt);
-	if(!_AnalysisBuf.read(AnaylsisBuf, AnaylsisBuf + _SampleCt))
-		return;
+	if(!_AnalysisBuf.read(AnaylsisBuf, AnaylsisBuf + _SampleCt)){
+		*maxVolume = _LastMaxVol * 0.85f;
+		return -1;
+	}
 	_AnalysisBuf.pop(_Step);
-	double pitch = dywapitch_computepitch(&_State, AnaylsisBuf,0,_SampleCt);
+	double pitch = dywapitch_computepitch(&_State, AnaylsisBuf, 0, _SampleCt, maxVolume, _VolTreshold);
 	if(pitch == 0.0)
-		_LastNote = -1;
+		return -1;
 	else
-		_LastNote = FreqToNote(pitch);
-}
-
-double PtDyWa::GetNote(){
-	return _LastNote;
+		return FreqToNote(pitch);
 }

PitchTracker/dywapitchtrack/ptDyWa.h

@@ -11,12 +11,14 @@ public:
 		_AnalysisBuf.insert(begin, end);
 	}
 
-	void Process();
-	double GetNote();
+	double FindNote(float* maxVolume);
+	void SetVolumeThreshold(float threshold);
+	float GetVolumeThreshold(){return _VolTreshold;}
 private:
 	dywapitchtracker _State;
 	constexpr static size_t _SampleCt = 2048;
 	RingBuffer<_SampleCt * 2> _AnalysisBuf;
 	unsigned _Step;
-	double _LastNote;
+	float _VolTreshold;
+	float _LastMaxVol;
 };

PitchTracker/performous/pitch.hh

@@ -86,6 +86,7 @@ public:
 	fft_t const& getFFT() const { return m_fft; }
 	/** Get the peak level in dB (negative value, 0.0 = clipping). **/
 	double getPeak() const { return 10.0 * log10(m_peak); }
+	double getPeakRaw() const { return m_peak; }
 	/** Get a list of all tones detected. **/
 	tones_t const& getTones() const { return m_tones; }
 	/** Find a tone within the singing range; prefers strong tones around 200-400 Hz. **/

PitchTracker/ptAKF.cpp

@@ -47,6 +47,8 @@ PtAKF::PtAKF(unsigned step){
 		}
 	}
 	_Step = step;
+	_VolTreshold = 0.01f;
+	_LastMaxVol = 0.f;
 }
 
 PtAKF::~PtAKF(){
@@ -59,13 +61,19 @@ PtAKF::~PtAKF(){
 	}
 }
 
-int PtAKF::GetNote(double* restrict maxVolume, float* restrict weights){
+void PtAKF::SetVolumeThreshold(float threshold){
+	_VolTreshold = threshold;
+}
+
+int PtAKF::GetNote(float* restrict maxVolume, float* restrict weights){
 	float AnaylsisBuf[_SampleCt];
 	size_t size = _AnalysisBuf.size();
 	if(size > _SampleCt)
 		_AnalysisBuf.pop(size - _SampleCt);
-	if(!_AnalysisBuf.read(AnaylsisBuf, AnaylsisBuf + _SampleCt))
+	if(!_AnalysisBuf.read(AnaylsisBuf, AnaylsisBuf + _SampleCt)){
+		*maxVolume = _LastMaxVol * 0.85f;
 		return -1;
+	}
 	_AnalysisBuf.pop(_Step);
 	return _GetNote(AnaylsisBuf, maxVolume, weights);
 }
@@ -96,7 +104,7 @@ void PtAKF::AddPeak(SPeak peaks[], float curWeight, int toneIndex){
 	}
 }
 
-int PtAKF::_GetNote(float* restrict samples, double* restrict maxVolume, float* restrict weights){
+int PtAKF::_GetNote(float* restrict samples, float* restrict maxVolume, float* restrict weights){
 	// Calculate maximum volume
 
 	float maxVolumeL = 0;
@@ -105,9 +113,12 @@ int PtAKF::_GetNote(float* restrict samples, double* restrict maxVolume, float*
 		if(vol > maxVolumeL)
 			maxVolumeL = vol;
 	}
-	//maxVolumeL /= 32767.;
+	//maxVolumeL /= 32767.f;
 	*maxVolume = maxVolumeL;
 
+	if(maxVolumeL < _VolTreshold)
+		return -1;
+
 	float samplesWindowed[_SampleCt];
 
 	for(int i = 0; i < _SampleCt; i++){

PitchTracker/ptAKF.h

@@ -29,7 +29,9 @@ public:
 		_AnalysisBuf.insert(begin, end);
 	}
 
-	int GetNote(double* restrict maxVolume, float* restrict weights);
+	int GetNote(float* restrict maxVolume, float* restrict weights);
+	void SetVolumeThreshold(float threshold);
+	float GetVolumeThreshold(){return _VolTreshold;}
 	static int GetNumHalfTones(){ return _MaxHalfTone - _MinHalfTone + 1;}
 
 private:
@@ -48,8 +50,11 @@ private:
 	constexpr static size_t _SampleCt = 2048;
 	RingBuffer<_SampleCt * 2> _AnalysisBuf;
 	unsigned _Step;
+	float _VolTreshold;
+	float _LastMaxVol;
+
+	int _GetNote(float* restrict samples, float* restrict maxVolume, float* restrict weights);
 
-	static int _GetNote(float* restrict samples, double* restrict maxVolume, float* restrict weights);
 	static float _AnalyzeBySampleCt(float* restrict samples, float* restrict samplesWindowed, float samplesPerPeriodD);
 	static inline float _AnalyzeByTone(float* restrict samples, float* restrict samplesWindowed, int toneIndex);
 	static float _AKFBySampleCt(float* restrict samples, float samplesPerPeriodD);