marytts.htsengine

Class HTSEngineTest

java.lang.Object

marytts.htsengine.HTSEngineTest

public class HTSEngineTest
extends Object

Several functions for running the htsEngine or other components stand alone

Author:

Marcela Charfuelan

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
class	HTSEngineTest.PhonemeDuration

Constructor Summary

Constructors

Constructor and Description
HTSEngineTest()

Method Summary

Methods

Modifier and Type	Method and Description
void	generateParameters() Stand alone testing using a TARGETFEATURES file as input.
void	getSptkMfcc() Calculate mfcc using SPTK, uses sox to convert wav→raw
void	getSptkSnackLf0() Calculate mfcc using SPTK, uses sox to convert wav→raw
float	loadDurationsForAlignment(String fileName, Vector<HTSEngineTest.PhonemeDuration> alignDur) Load durations for phone alignment when the durations have been generated by EHMMs.
void	loadF0contour(String lf0File, int totalDurationFrames, HTSParameterGeneration pdf2par) Load logf0, in HTS format, create a voiced array and set this values in pdf2par This contour should be aligned with the durations, so the total duration in frames should be the same as in the lf0 file
static void	main(String[] args)
void	synthesisWithContinuousFeatureProcessors() Generation of speech using external specification of duration: using ContinuousFeatureProcessors of TARGETFEATURES Input: a TARGETFEATURES (.pfeats) file, this file should contain ContinuousFeatureProcessors: unit_duration float unit_logf0 float unit_logf0delta float The features unit_duration and unit_logf0 are used as external prosody, unit_logf0Delta is not used.
void	synthesisWithProsodySpecificationInExternalFiles() Generation of speech using external specification of duration: duration and logf0 in external files Input: a TARGETFEATURES (.pfeats) file

Methods inherited from class java.lang.Object

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

Constructor Detail

HTSEngineTest

public HTSEngineTest()

Method Detail

synthesisWithContinuousFeatureProcessors

public void synthesisWithContinuousFeatureProcessors()
                                              throws Exception

Generation of speech using external specification of duration: using ContinuousFeatureProcessors of TARGETFEATURES Input: a TARGETFEATURES (.pfeats) file, this file should contain ContinuousFeatureProcessors: unit_duration float unit_logf0 float unit_logf0delta float The features unit_duration and unit_logf0 are used as external prosody, unit_logf0Delta is not used. The TARGETFEATURES (.pfeats) file including ContinuousFeatureProcessors values can be generated with a unitselection voice or a mbrola voice, it can NOT be generated with HMM voices.

Throws:

Exception - Exception

synthesisWithProsodySpecificationInExternalFiles

public void synthesisWithProsodySpecificationInExternalFiles()
                                                      throws Exception

Generation of speech using external specification of duration: duration and logf0 in external files Input: a TARGETFEATURES (.pfeats) file

Throws:

Exception - Exception

loadDurationsForAlignment

public float loadDurationsForAlignment(String fileName,
                              Vector<HTSEngineTest.PhonemeDuration> alignDur)

Load durations for phone alignment when the durations have been generated by EHMMs.

Parameters:

fileName - the format is the same as for phonelab.

alignDur - alignDur

Returns:

totalDuration

loadF0contour

public void loadF0contour(String lf0File,
                 int totalDurationFrames,
                 HTSParameterGeneration pdf2par)
                   throws Exception

Load logf0, in HTS format, create a voiced array and set this values in pdf2par This contour should be aligned with the durations, so the total duration in frames should be the same as in the lf0 file

Parameters:

lf0File - : in HTS formant

totalDurationFrames - : the total duration in frames can be calculated as: totalDurationFrames = totalDurationInSeconds / (framePeriodInSamples / SamplingFrequencyInHz)

pdf2par - : HTSParameterGeneration object

Throws:

Exception - If the number of frames in the lf0 file is not the same as represented in the total duration (in frames).

generateParameters

public void generateParameters()
                        throws IOException,
                               InterruptedException,
                               Exception

Stand alone testing using a TARGETFEATURES file as input. Generates duration: file.lab, duration state level: file.slab, f0: file.f0, mfcc: file.mfcc and sound file: file.wav out of HMM models

Throws:

IOException - IOException

InterruptedException - InterruptedException

Exception - Exception

getSptkMfcc

public void getSptkMfcc()
                 throws IOException,
                        InterruptedException,
                        Exception

Calculate mfcc using SPTK, uses sox to convert wav→raw

Throws:

IOException - IOException

InterruptedException - InterruptedException

Exception - Exception

getSptkSnackLf0

public void getSptkSnackLf0()
                     throws IOException,
                            InterruptedException,
                            Exception

Calculate mfcc using SPTK, uses sox to convert wav→raw

Throws:

IOException - IOException

InterruptedException - InterruptedException

Exception - Exception

main

public static void main(String[] args)
                 throws Exception

Throws:

Exception