annotate src/AudioEventMatcher.cpp @ 8:572564b7cb85

added calculation posterior into both onset and pitch processes

/*
 *  AudioEventMatcher.cpp
 *  MultipleAudioMathcher
 *
 *  Created by Andrew on 31/01/2012.
 *  Copyright 2012 QMUL. All rights reserved.
 *
 */

#include "AudioEventMatcher.h"


const int matchWindowWidth = 6000;

AudioEventMatcher::AudioEventMatcher(){

	setArraySizes();
	
	usingRealTime = false;
	bayesianStruct.realTimeMode = &usingRealTime;
	recentPitch = 0;
	currentAlignmentPosition = 0;
}

void AudioEventMatcher::setWindowDimensions(){
	double startHeight = recordedTracks.numberOfAudioTracks * recordedTracks.trackScreenHeight;
	double heightAvailable = 1 - startHeight;
	heightAvailable /= 3.0;
	
	bayesPositionWindow.setToRelativeSize(0, startHeight, 1, heightAvailable);
	bayesLikelihoodWindow.setToRelativeSize(0, startHeight + 1*heightAvailable, 1, heightAvailable);
	bayesTempoWindow.setToRelativeSize(0, startHeight + 2*heightAvailable, 1, heightAvailable);
	
	
}

void AudioEventMatcher::setArraySizes(){
	bayesianStruct.resetSpeedSize(200);
	bayesianStruct.setRelativeSpeedScalar(0.01);
	bayesianStruct.setSpeedPrior(1.0);
	bayesianStruct.relativeSpeedPrior.getMaximum();
	
	bayesianStruct.resetSize(matchWindowWidth);
	bayesianStruct.setPositionDistributionScalar(1);
	
}

void  AudioEventMatcher::startPlaying(){
	bayesianStruct.setStartPlaying();
	currentAlignmentPosition = 0;
	startTime = ofGetElapsedTimeMillis();
	//bayesianStruct.posterior.printArray();
}

void AudioEventMatcher::updateBestAlignmentPosition(){
	currentAlignmentPosition = bayesianStruct.posterior.offset + bayesianStruct.posterior.getIndexInRealTerms(bayesianStruct.posterior.MAPestimate);
	currentAlignmentPosition += (ofGetElapsedTimeMillis() - lastAlignmentTime) * bayesianStruct.relativeSpeedPosterior.getIndexInRealTerms(bayesianStruct.relativeSpeedPosterior.MAPestimate);
}

void AudioEventMatcher::draw(){
	//draw some outlines in blue
	ofSetColor(20,200,200);
	bayesPositionWindow.drawOutline();
	bayesTempoWindow.drawOutline();
	
	//draw the scrolling audio tracks
	recordedTracks.drawTracks();


	
	ofSetColor(255);
//	bayesianStruct.relativeSpeedPrior.drawVector(0, 200, bayesTempoWindow);
	
	drawBayesianDistributions();
	
	//	bayesianStruct.posterior.drawVector(0, bayesianStruct.posterior.getRealTermsAsIndex(screenWidthMillis), bayesPositionWindow);
	
	//bayesianStruct.posterior.drawVector(bayesianStruct.posterior.getRealTermsAsIndex(0), bayesianStruct.posterior.getRealTermsAsIndex(screenWidthMillis), bayesPositionWindow);
	
	//	bayesianStruct.relativeSpeedPosterior.drawVector(0, bayesianStruct.relativeSpeedPosterior.getRealTermsAsIndex(2), bayesTempoWindow);
	
	ofDrawBitmapString("pitch "+ofToString(recentPitch, 2)+", Time "+ofToString(recentTime, 0), 20, 20);
}
	
void AudioEventMatcher::drawBayesianDistributions(){
	
	
	double screenWidthMillis = recordedTracks.loadedAudioFiles[0].fileLoader.onsetDetect.framesToMillis(recordedTracks.loadedAudioFiles[0].fileLoader.onsetDetect.amplitudeNumber);
	

	double screenStartTimeMillis = recordedTracks.loadedAudioFiles[0].fileLoader.onsetDetect.framesToMillis(recordedTracks.loadedAudioFiles[0].fileLoader.onsetDetect.drawParams.windowStartFrame);
	double screenEndTimeMillis = screenStartTimeMillis + screenWidthMillis;
	int startIndex = bayesianStruct.posterior.getRealTermsAsIndex(screenStartTimeMillis);
	int endIndex = bayesianStruct.posterior.getRealTermsAsIndex(screenEndTimeMillis);
	
	bayesianStruct.posterior.drawConstrainedVector(startIndex, endIndex, 0, ofGetWidth(), bayesPositionWindow);
	
	string tmpString = "start "+ofToString(screenStartTimeMillis)+" (index "+ofToString(startIndex)+"), end "+ofToString(screenEndTimeMillis);
	ofDrawBitmapString(tmpString, bayesPositionWindow.x+20, bayesPositionWindow.y+20);
	
//	bayesianStruct.likelihood.drawConstrainedVector(startIndex, endIndex, 0, ofGetWidth(), bayesLikelihoodWindow);
	
	bayesianStruct.relativeSpeedPosterior.drawConstrainedVector(0, bayesianStruct.relativeSpeedPosterior.arraySize, 0, ofGetWidth(), bayesTempoWindow);
	
	string tmpStr = "zero is "+ofToString(bayesianStruct.posterior.getRealTermsAsIndex(0));
	tmpStr += " offsetis "+ofToString(bayesianStruct.posterior.offset);
	tmpStr += " screenWidth = "+ofToString(bayesianStruct.posterior.getRealTermsAsIndex(screenWidthMillis));
	ofDrawBitmapString(tmpStr, 20,140);
	tmpStr = "best est "+ofToString(bayesianStruct.bestEstimate);
	ofDrawBitmapString(tmpStr, 20, 180);
	
	ofDrawBitmapString("screenwidth "+ofToString(screenWidthMillis), 20, 800);
	
	ofSetColor(0,255,0);
	double currentEstimateIndex = (currentAlignmentPosition - screenStartTimeMillis)*ofGetWidth()/screenWidthMillis;
	ofLine(currentEstimateIndex, bayesPositionWindow.y, currentEstimateIndex, bayesPositionWindow.y + bayesPositionWindow.height);
	
	//draw track by track likelihoods
	for (int i = 0; i <recordedTracks.numberOfAudioTracks;i++){
		ofSetColor(200,255,50);
		likelihoodVisualisation[i].drawConstrainedVector(likelihoodVisualisation[i].getRealTermsAsIndex(screenStartTimeMillis), likelihoodVisualisation[i].getRealTermsAsIndex(screenEndTimeMillis), 0, ofGetWidth(), recordedTracks.loadedAudioFiles[i].fileLoader.onsetDetect.window);		
		ofSetColor(255);
		ofDrawBitmapString("recent event "+ofToString(recentEventTime[i]), recordedTracks.loadedAudioFiles[i].fileLoader.onsetDetect.window.x + 20, recordedTracks.loadedAudioFiles[i].fileLoader.onsetDetect.window.y + recordedTracks.loadedAudioFiles[i].fileLoader.onsetDetect.window.height - 10);
	}

	int priorStartIndex = recentPrior.getRealTermsAsIndex(screenStartTimeMillis);
	int priorEndIndex = recentPrior.getRealTermsAsIndex(screenEndTimeMillis);
	ofSetColor(0,200,200);
	recentPrior.drawConstrainedVector(priorStartIndex, priorEndIndex, 0, ofGetWidth(), bayesPositionWindow);

//	bayesianStruct.prior.addTriangularShape(100, 20, 0.4);
	ofSetColor(255,0,100);
	bayesianStruct.prior.drawConstrainedVector(bayesianStruct.prior.getRealTermsAsIndex(screenStartTimeMillis), bayesianStruct.prior.getRealTermsAsIndex(screenEndTimeMillis), 0, ofGetWidth(), bayesLikelihoodWindow);
	
}

void AudioEventMatcher::newPitchEvent(const int& channel, const double& pitchIn, const double& timeIn){
	if (pitchIn > 0){
	liveInput.addPitchEvent(pitchIn, timeIn);
	
	//tmp print stuff
	printf("New pitch MAP post estimate now %i, ", bayesianStruct.posterior.MAPestimate);
	double tmp  = bayesianStruct.posterior.getMAPestimate();
	printf(" getting it %f and offset %f == %f ms\n", tmp,  bayesianStruct.posterior.offset, bayesianStruct.posterior.getIndexInRealTerms(tmp));
	
	matchNewPitchEvent(channel, pitchIn, timeIn);//main pitch matching fn

	likelihoodVisualisation[1] = bayesianStruct.likelihood;
		
	recentPitch = pitchIn;//for drawing
	recentTime = timeIn;
	}
	
	
}

void AudioEventMatcher::newKickEvent(const double& timeIn){	
//	liveInput.addKickEvent(timeIn);
	matchNewOnsetEvent(0, timeIn);
	likelihoodVisualisation[0] = bayesianStruct.likelihood;
}

void AudioEventMatcher::newKickEvent(const int& channel, const double& timeIn){
//	liveInput.addKickEvent(timeIn);
	matchNewOnsetEvent(channel, timeIn);
	likelihoodVisualisation[0] = bayesianStruct.likelihood;
}


void AudioEventMatcher::newSnareEvent(const double& timeIn){
	matchNewOnsetEvent(2, timeIn);
	likelihoodVisualisation[2] = bayesianStruct.likelihood;
}


void AudioEventMatcher::newSnareEvent(const int& channel, const double& timeIn){
	matchNewOnsetEvent(channel, timeIn);
	likelihoodVisualisation[2] = bayesianStruct.likelihood;
}

//Needs just to set bounds for the matching process, not have TimeIn
void AudioEventMatcher::matchNewOnsetEvent(const int& channel, const double& timeIn){

	
	bayesianStruct.updateBayesianDistributions(timeIn);//moves the posterior up into prior given the time interval and calculates new offsets
	
	//start at beginning but OPTIMISE later
	double onsetLikelihoodToNoise = 0.5;
	
	double likelihoodWidth = 40;
	
	bayesianStruct.likelihood.offset = bayesianStruct.prior.offset;
	bayesianStruct.likelihood.zero();//set to zero
	
	double quantity = 1;//likelihoodToNoiseRatio / numberOfMatches;
	int numberOfMatchesFound = 0;

	
	double startTime = bayesianStruct.likelihood.offset;
	double endTime = bayesianStruct.likelihood.offset + matchWindowWidth;
	
	if (channel <= recordedTracks.numberOfAudioTracks){
		for (int i = 0;i < recordedTracks.loadedAudioFiles[channel].fileLoader.onsetDetect.chromaOnsets.size();i++){
			double millisTime = recordedTracks.loadedAudioFiles[channel].fileLoader.onsetDetect.chromaOnsets[i].millisTime;
			if (millisTime >= startTime && millisTime <= endTime){
				bayesianStruct.likelihood.addGaussianShapeFromRealTime(millisTime, likelihoodWidth,  quantity);
				numberOfMatchesFound++;
		//		printf("Adding Gaussian for onset at time %f offset %f\n", millisTime, bayesianStruct.likelihood.offset);
				
			}
		}
	}
	
//	bayesianStruct.likelihood.addConstant((1-likelihoodToNoiseRatio)/bayesianStruct.likelihood.length);
	bayesianStruct.likelihood.addConstant(numberOfMatchesFound*(1-onsetLikelihoodToNoise)/(onsetLikelihoodToNoise*bayesianStruct.likelihood.length));
	bayesianStruct.likelihood.renormalise();
	
	bayesianStruct.calculatePosterior();
	
	lastAlignmentTime = ofGetElapsedTimeMillis();
	recentEventTime[channel] = ofGetElapsedTimeMillis() - startTime;
	
}



void AudioEventMatcher::matchNewPitchEvent(const int& channel, const double& pitchIn, const double& timeIn){
	//start at beginning but OPTIMISE later
	
	
	bayesianStruct.updateBayesianDistributions(timeIn);//moves the posterior up into prior given the time interval and calculates new offsets

	//set the lielihoods by matching the pitched note
	
	double pitchLikelihoodToNoise = 0.5;//more noise
	int numberOfMatches = 0;
	bayesianStruct.likelihood.zero();//set to zero
	
	double quantity = 0;
	if (channel <= recordedTracks.numberOfAudioTracks){
		for (int i = 0;i < recordedTracks.loadedAudioFiles[channel].fileLoader.onsetDetect.chromaOnsets.size();i++){
			
			if (checkMatch(recordedTracks.loadedAudioFiles[channel].fileLoader.onsetDetect.chromaOnsets[i].aubioPitch, pitchIn)) {
				quantity = getPitchDistance(recordedTracks.loadedAudioFiles[channel].fileLoader.onsetDetect.chromaOnsets[i].aubioPitch, pitchIn, 10);
				bayesianStruct.likelihood.addGaussianShapeFromRealTime(recordedTracks.loadedAudioFiles[channel].fileLoader.onsetDetect.chromaOnsets[i].millisTime, 30, quantity);
				recordedTracks.loadedAudioFiles[channel].fileLoader.onsetDetect.chromaOnsets[i].matched = true;
				numberOfMatches++;
			} 
			else{
				recordedTracks.loadedAudioFiles[channel].fileLoader.onsetDetect.chromaOnsets[i].matched = false;
			}
			
		}
	}

	recentPrior = bayesianStruct.prior;
	
	
	if (numberOfMatches > 0){//no point updating unless there is a match
	
		bayesianStruct.likelihood.addConstant(numberOfMatches*(1-pitchLikelihoodToNoise)/(pitchLikelihoodToNoise*bayesianStruct.likelihood.length));

	//tmp set likelihood constant and calculate using that
	//bayesianStruct.likelihood.zero();
	//bayesianStruct.likelihood.addConstant(1);
		
		bayesianStruct.calculatePosterior();
	}
	
	lastAlignmentTime = ofGetElapsedTimeMillis();
	recentEventTime[channel] = ofGetElapsedTimeMillis() - startTime;
}

double AudioEventMatcher::getPitchDistance(const double& pitchOne, const double& pitchTwo, const double& scale){
	
	double distance = abs(pitchOne - pitchTwo);
	if (distance < scale)
		distance = 1 - (distance/scale);
	else
		distance = 0;
	
//	printf("[pitch distance %f vs %f = %f\n", pitchOne, pitchTwo, distance);
	return distance;

}


bool AudioEventMatcher::checkMatch(const double& recordedPitch, const double& livePitch){
	if (abs(recordedPitch - livePitch) < 40)
		return true;
	else
		return false;
}



void AudioEventMatcher::windowResized(const int& w, const int& h){
	recordedTracks.windowResized(w,h);
	bayesTempoWindow.resized(w,h);
	bayesPositionWindow.resized(w,h);
}