marytts.signalproc.adaptation.gmm.jointgmm

Class JointGMMTransformer

java.lang.Object

marytts.signalproc.adaptation.BaselineTransformer

marytts.signalproc.adaptation.gmm.jointgmm.JointGMMTransformer

public class JointGMMTransformer
extends BaselineTransformer

Voice conversion transformation using Joint-GMM approach. Reference: A. Kain and M. Macon, “Spectral voice conversion for text-to-speech synthesis,” in Proc. of the IEEE ICASSP 1998, vol. 1, pp. 285-288.

Author:

Oytun Türk

Field Summary

Fields

Modifier and Type	Field and Description
JointGMMSet	jointGmmSet
JointGMMMapper	mapper
JointGMMTransformerParams	params
PitchMapping	pitchMapping

Fields inherited from class marytts.signalproc.adaptation.BaselineTransformer

featureExtractor, postprocessor, preprocessor

Constructor Summary

Constructors

Constructor and Description
JointGMMTransformer(BaselinePreprocessor pp, BaselineFeatureExtractor fe, BaselinePostprocessor po, JointGMMTransformerParams pa)

Method Summary

Methods

Modifier and Type	Method and Description
boolean	checkParams()
BaselineAdaptationSet	getInputSet(String inputFolder)
BaselineAdaptationSet	getOutputSet(BaselineAdaptationSet inputSet, String outputFolder)
static void	main(String[] args)
static void	mainHmmVoiceConversion(String[] args)
static void	mainIeeeTaslp2009_mary(String[] args)
static void	mainIeeeTaslp2009_rap(String[] args)
static void	mainInterspeech2008(String[] args)
static void	mainParametric(String inputFolder, String outputBaseFolder, String baseFile, String outputFolderInfoString, boolean isSourceVocalTractSpectrumFromModel, boolean isTemporalSmoothing, int smoothingNumNeighbours, boolean isPitchFromTargetFile, int pitchFromTargetMethod, boolean isDurationFromTargetFile, int durationFromTargetMethod, boolean isEnergyFromTargetFile, boolean isLsfsFromTargetFile, int targetAlignmentFileType, boolean isContextualGMMs, int contextClassificationType, int[] numComponents)
static void	mainQuickTest(String[] args)
static void	mainQuickTest2(String[] args) This example uses the ouput of the example: marytts.signalproc.adaptation.gmm.jointgmm.JointGMMParallelTrainer.mainQuickTest2() Input: /Neutral-Spike-Conversion/ouput/source2target/sourceF_X_targetF_99_10.jgs /Neutral-Spike-Conversion/source/wav/*.wav (files to convert, normally these files are different from the training set) Output: /Neutral-Spike-Conversion/output/source2target/gmmF_99/isSrc0_smooth1_10_mixes10_prosody1x0x0/*.wav transformed files
void	run()
void	transform(BaselineAdaptationSet inputSet, BaselineAdaptationSet outputSet)
static void	transformOneItem(BaselineAdaptationItem inputItem, BaselineAdaptationItem outputItem, JointGMMTransformerParams wctParams, JointGMMMapper jgMapper, JointGMMSet jgSet, PitchTransformationData pMap)

Methods inherited from class marytts.signalproc.adaptation.BaselineTransformer

isScalingsRequired, transformOneItem

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

params

public JointGMMTransformerParams params

mapper

public JointGMMMapper mapper

jointGmmSet

public JointGMMSet jointGmmSet

pitchMapping

public PitchMapping pitchMapping

Constructor Detail

JointGMMTransformer

public JointGMMTransformer(BaselinePreprocessor pp,
                   BaselineFeatureExtractor fe,
                   BaselinePostprocessor po,
                   JointGMMTransformerParams pa)

Method Detail

checkParams

public boolean checkParams()
                    throws IOException

Overrides:

checkParams in class BaselineTransformer

Throws:

IOException

run

public void run()
         throws IOException,
                UnsupportedAudioFileException

Throws:

IOException

UnsupportedAudioFileException

getInputSet

public BaselineAdaptationSet getInputSet(String inputFolder)

getOutputSet

public BaselineAdaptationSet getOutputSet(BaselineAdaptationSet inputSet,
                                 String outputFolder)

transform

public void transform(BaselineAdaptationSet inputSet,
             BaselineAdaptationSet outputSet)
               throws UnsupportedAudioFileException

Overrides:

transform in class BaselineTransformer

Throws:

UnsupportedAudioFileException

transformOneItem

public static void transformOneItem(BaselineAdaptationItem inputItem,
                    BaselineAdaptationItem outputItem,
                    JointGMMTransformerParams wctParams,
                    JointGMMMapper jgMapper,
                    JointGMMSet jgSet,
                    PitchTransformationData pMap)
                             throws UnsupportedAudioFileException,
                                    IOException

Throws:

UnsupportedAudioFileException

IOException

main

public static void main(String[] args)
                 throws IOException,
                        UnsupportedAudioFileException

Throws:

IOException

UnsupportedAudioFileException

mainHmmVoiceConversion

public static void mainHmmVoiceConversion(String[] args)
                                   throws UnsupportedAudioFileException,
                                          IOException

Throws:

UnsupportedAudioFileException

IOException

mainQuickTest

public static void mainQuickTest(String[] args)
                          throws UnsupportedAudioFileException,
                                 IOException

Throws:

UnsupportedAudioFileException

IOException

mainQuickTest2

public static void mainQuickTest2(String[] args)
                           throws UnsupportedAudioFileException,
                                  IOException

This example uses the ouput of the example: marytts.signalproc.adaptation.gmm.jointgmm.JointGMMParallelTrainer.mainQuickTest2() Input: /Neutral-Spike-Conversion/ouput/source2target/sourceF_X_targetF_99_10.jgs /Neutral-Spike-Conversion/source/wav/*.wav (files to convert, normally these files are different from the training set) Output: /Neutral-Spike-Conversion/output/source2target/gmmF_99/isSrc0_smooth1_10_mixes10_prosody1x0x0/*.wav transformed files

Parameters:

args - args

Throws:

UnsupportedAudioFileException - UnsupportedAudioFileException

IOException - IOException

mainInterspeech2008

public static void mainInterspeech2008(String[] args)
                                throws IOException,
                                       UnsupportedAudioFileException

Throws:

IOException

UnsupportedAudioFileException

mainIeeeTaslp2009_mary

public static void mainIeeeTaslp2009_mary(String[] args)
                                   throws IOException,
                                          UnsupportedAudioFileException

Throws:

IOException

UnsupportedAudioFileException

mainIeeeTaslp2009_rap

public static void mainIeeeTaslp2009_rap(String[] args)
                                  throws IOException,
                                         UnsupportedAudioFileException

Throws:

IOException

UnsupportedAudioFileException

mainParametric

public static void mainParametric(String inputFolder,
                  String outputBaseFolder,
                  String baseFile,
                  String outputFolderInfoString,
                  boolean isSourceVocalTractSpectrumFromModel,
                  boolean isTemporalSmoothing,
                  int smoothingNumNeighbours,
                  boolean isPitchFromTargetFile,
                  int pitchFromTargetMethod,
                  boolean isDurationFromTargetFile,
                  int durationFromTargetMethod,
                  boolean isEnergyFromTargetFile,
                  boolean isLsfsFromTargetFile,
                  int targetAlignmentFileType,
                  boolean isContextualGMMs,
                  int contextClassificationType,
                  int[] numComponents)
                           throws IOException,
                                  UnsupportedAudioFileException

Throws:

IOException

UnsupportedAudioFileException