Changes between Version 13 and Version 14 of HMMVoiceCreation


Ignore:
Timestamp:
08/31/10 16:52:13 (14 years ago)
Author:
marcela_charfuelan
Comment:

--

Legend:

Unmodified
Added
Removed
Modified

  • HMMVoiceCreation

    v13 v14  
    11 
    2  
    3  
    4 = '''Voice Import Tools Tutorial : How to build a HMM-based voice for the MARY platform''' = 
    5  
    6 For creating HMM-based voices we use a version of the speaker dependent training scripts provided by [http://hts.sp.nitech.ac.jp/ HTS] that was adapted to the MARY 
    7 platform. The steps for building a HMM voice for the MARY platform can be summarised in:[[BR]] 
    8  
    9 I)   Checking the necessary programs and files[[BR]] 
    10 II)  Data preparation[[BR]] 
    11 III) Training of HMM models[[BR]] 
    12 IV)  Adding a new HMM voice in the Mary system.[[BR]] 
    13 V)   Creating other voice in German or English (__if you want to train HMMs with another speech database__).[[BR]] 
    14  
    15 The previous steps will be explained below creating a HMM voice using the HTS '''speaker dependent training demo'''.[[BR]] 
    16  
    17 For an explanation on how to create an adapted HMM voice using the '''speaker adaptation/adaptive training demo''' please see [wiki:HMMVoiceCreationAdapt]. [[BR]] 
    18  
    19 The training scripts used here are the latest versions, that is, it is required HTS_2.0.1 and SPTK-3.1. Some scripts have been added-modified to:[[BR]] 
     2= '''Voice Import Tools Tutorial : How to build a HMM-based voice for the MARY 4.1.0 platform''' = 
     3 
     4For creating HMM-based voices we use a version of the speaker dependent training scripts provided by [http://hts.sp.nitech.ac.jp/ HTS] that was adapted to the MARY 4.1.0 platform. The steps for building a HMM voice for the MARY platform can be summarised in:[[BR]] 
     5 
     6I) Download MARY TTS including Voice import tools[[BR]] 
     7II) Check necessary programs and files[[BR]] 
     8III) Check data: audio and text files[[BR]] 
     9IV) Run the Voice import tools [[BR]] 
     10V) Creating other voice in a language different from German or English (US). 
     11 
     12The training scripts used here are the latest versions, that is, it is required HTS_2.1 and SPTK-3.2. Some scripts have been added-modified to:[[BR]] 
    2013- Use MARY instead of festival as text analyzer.[[BR]] 
    21 - Train bandpass voicing strengths for mixed excitation.[[BR]] 
    22 - Process language specific settings as parameters.[[BR]] 
    23  
    24  
    25 ''' 
    26 === I) Checking the necessary programs and files: === 
    27 ''' 
    28  
    29 MARY requirements:[[BR]] 
    30 - Operating System - Linux[[BR]] 
    31 - MARY TTS Recent Version - Download Link: http://mary.dfki.de/Download [[BR]] 
    32 - Openmary - SVN from http://mary.opendfki.de [[BR]] 
    33 - MARY patch for HTS demo: HTS-2.0.1-demo_CMU-ARCTIC-SLT_for_Mary-3.6.0.patch [[BR]] 
    34  
    35  
    36 HTS requirements, please download and follow the instructions for installing:[[BR]] 
    37 - HTS-2.0.1_for_HTK-3.4.patch [[BR]] 
    38 - HTK-3.4 ( [wiki:CompilingHTK Important note on compiling HTK] )   [[BR]]   
    39 - SPTK-3.1 [[BR]] 
    40 - HTS-demo_CMU-ARCTIC-SLT (for HTS-2.0.1) [[BR]] 
    41  
    42 Other requirements, please download and follow the instructions for installing: [[BR]] 
    43 - EHMM for automatic labeling, available with festvox-2.1 (Recent Version) http://festvox.org/download.html [[BR]] 
    44 - sox, normally available in linux.  [[BR]] 
    45 - tcl-tk supporting snack, for example  ActiveTcl - Download Link: http://www.activestate.com/Products/ActiveTcl/ [[BR]] 
    46 - perl, normally available in linux.  [[BR]] 
    47  
    48 0.1) download and un-zip, un-tar the latest speaker dependent training demo for English. 
    49  
    50 http://hts.sp.nitech.ac.jp/archives/2.0.1/HTS-demo_CMU-ARCTIC-SLT.tar.bz2 for HTS-2.0.1 
    51  
    52 0.2) download and unzip the patch file for using MARY instead of Festival as text analyser. 
    53  
    54 https://mary.opendfki.de/repos/trunk/lib/hts/HTS-2.0.1-demo_CMU-ARCTIC-SLT_for_Mary-3.6.0.patch.zip  [[BR]] 
    55  
    56 apply the patch to the HTS-demo_CMU-ARCTIC-SLT directory:  [[BR]]  
    57 {{{ 
    58    patch -p1 -d . < HTS-2.0.1-demo_CMU-ARCTIC-SLT_for_Mary-3.6.0.patch 
    59 }}} 
    60  
    61 0.3) create a wav directory. 
    62  
    63 0.4) Run the VoiceImport program 
    64  
    65 First of all you need to set your MARY_BASE directory and then run the program:  [[BR]] 
    66 {{{ 
    67    export MARY_BASE="/dir/to/openmary" 
    68    java -jar -Xmx1024m  $MARY_BASE/java/voiceimport.jar 
    69 }}} 
    70  
    71 If you are not familiar or have problems with the VoiceImport program, please read and follow the instructions in the Voice Import Tools 
    72 Tutorial: http://mary.opendfki.de/wiki/VoiceImportToolsTutorial 
    73  
    74 If you want to create another voice in German or English please see the section V below. 
    75  
    76 Please remember that whenever you are in doubt about the settings of a particular component you can check its corresponding help for a description of the meaning 
    77 (and possible values) of each variable. 
    78  
    79 ''' 
    80 === II) Data preparation: === 
    81 ''' 
    82  
    83 1- Run the HMMVoiceDataPreparation of the HMM Voice Trainer group to check if text, wav and data/raw files are available and in the correct paths. 
    84 If just data/raw is provided, the program will do the conversion.  If no text files are available but data/utts in festival format, the program will do the  
    85 conversion as well. 
    86  
    87 2- Run the PhoneUnitFeatureComputer component of the Feature Extraction group to extract context feature vectors from the text data. This procedure will create a "phonefeatures" directory.  For running this component the MARY server should be running as well. 
    88  
    89 3- Run the EHMMlabeler component of the Automatic Labeling group to label automatically the wav files using the corresponding transcriptions. This procedure might 
     14- Train bandpass voicing strengths and Fourier magnitudes for mixed excitation.[[BR]] 
     15 
     16'''MARY requirements:'''[[BR]] 
     17- Operating System - Linux (tested on Ubuntu 9.04) [[BR]] 
     18- MARY TTS 4.1.0 including Voice import tools during installation - link: [http://mary.dfki.de/download/4.0/openmary-standalone-install-4.0.0.jar MARY TTS 4.1.0] [[BR]] 
     19 
     20 
     21''' 
     22== I) Download MARY TTS including Voice import tools == 
     23''' 
     24 
     25Click on the latest MARY release [http://mary.dfki.de/download/4.0/openmary-standalone-install-4.0.0.jar MARY download] or download the file and run it with: 
     26{{{ 
     27java -jar openmary-standalone-install-4.1.0.jar 
     28}}} 
     29 
     30 
     31''' 
     32== II) Check the necessary programs and files: == 
     33''' 
     34 
     35We provide an script to facilitate the checking and installation of the necessary external programs, once installed MARY TTS open a command line shell in your voice building directory and run the shell script: 
     36{{{ 
     37$MARY_BASE/lib/external/check_install_external_programs.sh 
     38}}} 
     39 
     40With the option '''-check''', this script will check if the necessary programs and versions are installed (that is, the programs can be found in the PATH or in the paths provided by the user).[[BR]] 
     41With the option '''-install''' this script will try to download and install the necessary programs in: $MARY_BASE/lib/external/bin (if problems, it will suggest how to install manually the programs). 
     42 
     43If you have already installed some of the required programs, '''please include their paths in the PATH variable or provide the paths''', for example: 
     44{{{ 
     45$MARY_BASE/lib/external/check_install_external_programs.sh -check /your/path/to/htk/bin /your/path/to/Festival/festvox/src/ehmm/bin 
     46}}} 
     47 
     48This script generates a '''$MARY_BASE/lib/external/externalBinaries.config''' file that will be used by the Voice import tools to locate the necessary external programs. 
     49 
     50The necessary programs that this script checks are:[[BR]] 
     51 
     52'''HTS requirements:'''[[BR]] 
     53- [http://hts.sp.nitech.ac.jp/archives/2.1/HTS-2.1_for_HTK-3.4.tar.bz2 HTS-2.1_for_HTK-3.4.patch] [[BR]] 
     54- HTK-3.4 and HDecode patched with HTS-2.1_for_HTK-3.4.patch links: 
     55    * [http://htk.eng.cam.ac.uk/ftp/software/HTK-3.4.tar.gz HTK-3.4] (you will need to register first) [[BR]]  
     56    * [http://htk.eng.cam.ac.uk/prot-docs/hdecode.shtml HDecode] (you will need to register first) [[BR]] 
     57- [http://downloads.sourceforge.net/sp-tk/SPTK-3.2.tar.gz SPTK-3.2] [[BR]] 
     58- [http://downloads.sourceforge.net/hts-engine/hts_engine_API-1.01.tar.gz hts_engine_API-1.01] [[BR]] 
     59 
     60'''Other requirements:'''[[BR]] 
     61- awk normally available in linux [[BR]] 
     62- perl normally available in linux [[BR]] 
     63- bc normally available in linux [[BR]] 
     64- sox, v13.0 or greater [http://sox.sourceforge.net/ SoX], normally available in linux.  [[BR]] 
     65- tcl supporting snack, for example  [http://www.activestate.com/Products/ActiveTcl/ ActiveTcl.] Note that only ActiveTcl 8.4 includes snack; 8.5+ requires manual installation. [[BR]] 
     66- [http://www.speech.kth.se/snack/download.html snack] library for tcl.  [[BR]] 
     67- EHMM for automatic labeling, available with [http://festvox.org/download.html festvox-2.1] [[BR]] 
     68 
     69 
     70 
     71 
     72''' 
     73== III) Check data: audio and text files[[BR]] == 
     74''' 
     75 
     76In your voice building directory execute the step-by-step procedure in [http://mary.opendfki.de/wiki/VoiceImportToolsTutorial VoiceImportToolsTutorial] to make 
     77sure that the data, sound (wav) and text files are in the correct place and format.[[BR]] 
     78 
     79As a result of this step your voice building directory should contain a wav and text directories. 
     80 
     81 
     82''' 
     83== IV) Run the Voice Import tools == 
     84''' 
     85 
     86In your voice building directory run the voice import tools: 
     87{{{ 
     88export MARY_BASE="/your/path/to/MARY TTS/" 
     89java -Xmx1024m -jar $MARY_BASE/java/voiceimport.jar 
     90}}} 
     91 
     92After starting the Voice Import Tools check the global settings of the voice, make sure that the allophones file is provided and exists: 
     93{{{ 
     94db.alophonesSet = $MARY_BASE/lib/modules/xx/lexicon/allophones.xx.xml  (where xx is the corresponding language) 
     95}}} 
     96 
     97 
     98And run the following components: 
     99 
     100 
     101'''1-''' Run the HMMVoiceDataPreparation of the HMM Voice Trainer group to set up the environment to create a HMM voice and check if required external programs and text and wav files are available and in the correct paths.  
     102 
     103'''2-''' Run the AllophonesExtractor of the Automatic Labeling group to create the '''prompt_allophones''' directory required in the next step. This component requires the MARY server. [[BR]] 
     104 
     105'''3-''' Run the EHMMlabeler component of the Automatic Labeling group to label automatically the wav files using the corresponding transcriptions. This procedure might 
    90106take several hours. For running EHMMLabeler, please use the settings editor of this component to set, according to your festvox installation, the variable: 
    91107{{{ 
    92108   EHMMLabeler.ehmm  = ../festvox/src/ehmm/bin/ 
    93109}}} 
    94  
    95 4- Run the LabelPauseDeleter component of the Automatic Labeling group. Please use the settings editor of this component to set the variable: 
     110The result of this step is a '''ehmm/lab''' directory. 
     111 
     112'''4-''' Run the LabelPauseDeleter component of the Automatic Labeling group. Please use the settings editor of this component to set the variable: 
    96113{{{ 
    97114   LabelPauseDeleter.threshold  =  10 
    98115}}} 
    99  
    100 5- Run the PhoneUnitLabelComputer component of the Labels and Pause Correction group. This procedure will create a "phonelab" directory. 
    101  
    102 6- Run the PhonelabelFeatureAligner component of the Labels and Pause Correction group. This procedure will verify alignment between "phonefeatures" and "phonelabels". 
    103  
    104  
    105 ''' 
    106 === III) HMM models training: === 
    107 ''' 
    108  
    109 7- Run the HMMVoiceConfigure component of the HMM Voice trainer group, the default setting values of this component are already fixed for the 
    110 HTS-demo_CMU-ARCTIC-SLT voice. 
     116The result of this step is a '''lab''' directory. 
     117 
     118'''5-''' Run the TranscriptionAligner component of the Label-Transcript Alignment group.  This program will create the '''allophones''' directory. 
     119 
     120'''6-''' Run the PhoneUnitLabelComputer component of the Label-Transcript Alignment group. This procedure has as input the '''lab''' directory and will create as an output the  '''phonelab''' directory.  
     121 
     122'''7-''' Run the FeatureSelelection component of the Feature Extraction group. This program will create a '''mary/features.txt''' file, it requires the MARY server running. Select here all the features and save the file. 
     123 
     124'''8-''' Run the PhoneUnitFeatureComputer component of the Feature Extraction group to extract context feature vectors from the text data. This procedure will create a '''phonefeatures''' directory. For running this component the MARY server should be running as well.  
     125 
     126'''9-''' Run the PhonelabelFeatureAligner component of the Verify Alignment group. This procedure will verify alignment between "phonefeatures" and "phonelabels".[[BR]] 
     127 
     128As a result of previous steps we should have:[[BR]] 
     129- phonefeatures directory [[BR]] 
     130- phonelab directory [[BR]] 
     131- mary/features.txt file [[BR]] 
     132- $MARY_BASE/lib/external/externalBinaries.config 
     133 
     134 
     135''' 
     136=== HMM models training: === 
     137''' 
     138 
     139'''10-''' Run the HMMVoiceConfigure component of the HMM Voice trainer group. The default setting values are already fixed for the arctic slt voice, some path settings depend on your installation, and will be taken from $MARY_BASE/lib/external/externalBinaries.config 
    111140 
    112141If running configure for other voice, for example a male German voice, please use the settings editor of this component to set the variables: 
    113142{{{ 
    114143  HMMVoiceConfigure.dataSet      =  german_set_name 
    115   HMMVoiceConfigure.featureList  =  feature_list_de.pl  (the set of context features used for this voice can be change in this file). 
    116144  HMMVoiceConfigure.speaker      =  speaker_name  
    117145  HMMVoiceConfigure.lowerF0      =  40  (male=40,  female=80)   
    118146  HMMVoiceConfigure.upperF0      =  280 (male=280, female=350) 
    119   HMMVoiceConfigure.voiceLang    =  de 
    120 }}} 
    121  
    122 Using the settings editor of this component you can also change other variables like using LSP instead og MGC, sampling frequency, etc., 
    123 the same as you would do when running "make configure + parameters" with the original HTS scripts. 
    124  
    125 8- Run the HMMVoiceMakeData component of the HMM Voice trainer group to run the HTS procedure "make data". This procedure is the same as in the original scripts with additional sections for calculating strengths (for mixed excitation), global variance, and handling of MARY context features. 
    126  
    127 Particular procedures can be repeated isolated fixing the particular settings for this component. For example, if the procedure that creates strengths (in the str directory) has to be repeated with a different set of filters (data/filters/), please remove the old str directory and set: 
     147}}} 
     148 
     149Using the settings editor of this component you can also change other variables like using LSP instead og MGC, sampling frequency, etc., the same as you would do when running "make configure + parameters" with the original HTS scripts. 
     150 
     151'''11-''' Run the HMMVoiceFeatureSelection component of the HMM Voice trainer group. This program reads the '''mary/features.txt''' file (created in step 11), and generates the file '''mary/hmmFeatures.txt'''. This file contains extra features, apart from phone and phonological features, that will be used to train HMMs. When running this program a small set of features will be presented on top, separated by an empty line:[[BR]] 
     152{{{ 
     153   pos_in_syl 
     154   syl_break 
     155   prev_syl_break 
     156   position_type 
     157    
     158   accented 
     159   accented_syls_from_phrase_end 
     160   accented_syls_from_phrase_start 
     161   breakindex 
     162   edge 
     163   ... 
     164}}} 
     165If you are not sure about using other features, use the first four, delete the others and save the file. 
     166 
     167'''12-''' Run the HMMVoiceMakeData component of the HMM Voice trainer group to run the HTS procedure "make data". This procedure require the following files: 
     168{{{ 
     169   HMMVoiceMakeData.allophonesFile   = allophones.en_US.xml  # allophones set (language dependent) 
     170   HMMVoiceMakeData.featureListFile  = mary/hmmFeatures.txt  # extra context features used for training HMMs. 
     171}}} 
     172 
     173The allophones set file is language dependent, it can be found in $MARY_BASE/lib/modules/en/us/lexicon/allophones.en_US.xml[[BR]] 
     174The hmmFeatures.txt is the file created in step 15 and contains additional context features, apart from phone and phonological features, used for training HMMs.[[BR]] 
     175 
     176The HMMVoiceMakeData procedure is similar to the original HTS scripts with additional sections for calculating strengths, Fourier magnitudes (for mixed excitation), global variance and composing training data files from mgc, lf0, str and mag files. This component will execute in the hts/data/ directory:  
     177{{{ 
     178  make all-mary  or 
     179  make mgc lf0 str-mary mag-mary cmp-mary gv-mary gv list scp  
     180}}} 
     181The '''label''' directory and the '''mlf''' files in MARY are done with the Voice Import Tools: HMMVoiceMakeData.makeLabels()[[BR]] 
     182The '''questions''' file in MARY is done with the Voice Import Tools: HMMVoiceMakeData.makeQuestions()  
     183 
     184 
     185Particular procedures can be repeated isolated fixing the particular settings for this component. For example, if the procedure that creates strengths (in the str directory) has to be repeated with a different set of filters (data/filters/), set: 
    128186{{{ 
    129187  HMMVoiceMakeData.makeSTR       =  1 
    130188  HMMVocieMakeData.makeCMPMARY   =  1 
    131189}}} 
    132 all the other variables in 0, and run again the component. (In this case you need to run as well makeCMPMARY because you need to compose again the vectors mgc+lf0+str). 
    133  
    134 The procedures can be repeated manually as well, going to the data directory and running "make data" or "make str", as is normally done with the original HTS scripts. 
    135  
    136 NOTE: the Makefile in data/ includes a gv: section copied from HTS-2.1alpha version to calculate global variance files. In MARY, these files are generated little endian and contain a header of size one short to indicate the size of the vectors it contains. 
    137  
    138 9- Run the HMMVoiceMakeVoice component of the HMM Voice trainer group, here again particular training steps can be repeated selecting them (setting in 1, all the others in 0) from the settings of this component. This is equivalent to run again: 
    139 {{{ 
    140    perl scripts/Training.pl scripts/Config.pm 
     190all the other variables in 0, and run again the component. (In this case you need to run makeCMPMARY again because you need to compose again the vectors mgc+lf0+str+mag). 
     191 
     192The procedures can be repeated manually as well, going to the hts/data directory and running "make str-mary" and "make cmp-mary". 
     193 
     194NOTE: the Makefile in data/ includes a gv: section to calculate global variance files. In MARY, these files are generated little endian and contain a header of size one short to indicate the size of the vectors it contains. 
     195 
     196 
     197'''13-''' Run the HMMVoiceMakeVoice component of the HMM Voice trainer group, here again particular training steps can be repeated selecting them (setting in 1, all the others in 0) from the settings of this component. This is equivalent to run again: 
     198{{{ 
     199   perl scripts/Training.pl scripts/Config.pm > logfile & 
    141200}}} 
    142201after modifying the Config.pm file, as is normally done with the original HTS scripts. 
    … …  
    148207 
    149208''' 
    150 === IV) Adding a new voice in the MARY platform: === 
    151 ''' 
    152  
    153 10- Run the HMMVoiceInstaller component of the Install Voice group. The default setting values of this component are already fixed for the HTS-demo_CMU-ARCTIC-SLT voice. If you are training other voice  please use the settings editor of this component to set: 
    154 {{{ 
    155   HMMVoiceInstaller.FeaList     =  data/feature_list_xx.pl  
    156                                    make sure that this file is the one used during training, xx=en for English or xx=de for German. 
    157   HMMVoiceInstaller.Flab        =  data/labels/gen/xx.lab  
    158                                    this is an example of a label file in HTSCONTEXT format for synthesise during start-up.  
    159   HMMVoiceInstaller.useMixExc   =  true 
     209=== Adding a new voice in the MARY platform: === 
     210''' 
     211 
     212'''14-''' Run the HMMVoicePackager component of the Install Voice group. The default setting values of this component are already fixed for the HTS-demo_CMU-ARCTIC-SLT voice. Some settings of the voice can be changed here, for example: 
     213{{{ 
     214  HMMVoicePackager.useMixExc   =  true 
    160215                                   set this variable to true if using mixed excitation 
    161   HMMVoiceInstaller.useGV       =  true  
     216  HMMVoicePackager.useGV       =  true  
    162217                                   set this variable to true if using global variance in parameter generation. 
    163 }}} 
    164  
    165 The VoiceInstaller will:  [[BR]] 
    166 - Create a new mary config file in: $MARY_BASE/conf/german-hmm-voice.config [[BR]] 
    167 - Add the files corresponding to this voice in: $MARY_BASE/lib/voices/hmm-voice/  [[BR]] 
    168 - copy features list: data/feature_list_en.pl to  $MARY_BASE/lib/voices/hmm-voice  [[BR]] 
    169 - copy one example of phonelab for testing the synthesiser: data/labels/gen/gen_cmu_us_arctic_slt_xxxx.lab to $MARY_BASE/lib/voices/hmm-voice  [[BR]] 
    170 - copy the HTS trees: voices/qst001/ver1/*.inf to $MARY_BASE/lib/voices/hmm-voice  [[BR]] 
    171 - copy the HTS PDF models: voices/qst001/ver1/*.pdf to $MARY_BASE/lib/voices/hmm-voice [[BR]] 
    172 - copy global variance models (if useGV is set to true): data/gv/gv-*-littend.pdf to $MARY_BASE/lib/voices/hmm-voice [[BR]] 
    173 - copy filter taps for mixed excitation: data/filters/mix_excitation_filters.txt to $MARY_BASE/lib/voices/hmm-voice [[BR]] 
    174  
    175 After successfully installing a new voice, it can be used with the mary_server and the mary_client. 
    176  
    177  
    178 ''' 
    179 === V) Creating other voice in German or English. === 
    180 ''' 
    181  
    182 If using German: 
    183  
    184 For creating a new German voice it is necessary: [[BR]] 
    185   * a wav or raw directory with the speech files you will use for training the German voice. [[BR]] 
    186   * transcriptions of the files, one text file per speech file, or transcriptions in festival format if available. [[BR]] 
    187  
    188 Then we use as a base the original HTS-demo_CMU-ARCTIC-SLT directory: 
    189  
    190 - Download and un-zip, un-tar the HTS-demo_CMU-ARCTIC-SLT for HTS-2.0.1 
    191  
    192 - Rename this directory as your new voice name, for example german_voice, and delete the directories data/raw and data/utt. 
    193  
    194 - Apply the MARY patch to the german_voice directory. [[BR]] 
    195   patch -p1 -d . < HTS-2.0.1-demo_CMU-ARCTIC-SLT_for_Mary-3.5.0.patch 
    196  
    197 - Move your speech files to this directory, if you have a wav directory, this should be copied in the current directory (german_voice/wav). If you have a raw directory, this should be copied in the data directory (german_voice/data/raw). 
    198  
    199 - Move your transcription files to this directory, if you have a text directory containing the transcription of each file in separate files, this should be copied in the current directory (german_voice/text). If you have transcriptions in festival format please copy this directory in the data/utts directory (german_voice/data/utts/). 
    200  
    201 - Now run the VoiceImport program and follow the instructions as normal. Provide general settings for: 
     218  HMMVoicePackager.useAcousticModels = true 
     219                                       set this variable to true to allow prosody modification specified in MARYXML 
     220}}} 
     221 
     222The HMMVoicePackager will pack in a zip file located in MARY_BASE/download the following files:  [[BR]] 
     223- A mary config file: german-hsmm-voice.config [[BR]] 
     224- HMM files corresponding to this voice: [[BR]] 
     225  - one example of phonefeatures for testing the synthesiser: data/phonefeatures/cmu_us_arctic_slt_xxxx.pfeats  [[BR]] 
     226  - the HTS trees: voices/qst001/ver1/*.inf  [[BR]] 
     227  - the HTS PDF models: voices/qst001/ver1/*.pdf [[BR]] 
     228  - global variance models (if useGV is set to true): data/gv/gv-*-littend.pdf [[BR]] 
     229  - filter taps for mixed excitation: data/filters/mix_excitation_filters.txt [[BR]] 
     230  - trickyPhones.txt file, if one was created during training [[BR]] 
     231 
     232After successfully packing a new voice, you must run the MARY Component Installer to install the voice! 
     233 
     234 
     235''' 
     236=== V) Creating other voice in a language different from German or English (US). === 
     237''' 
     238 
     239If you are creating a voice in other language you will need to specify: 
     240 
     241- '''Minimal NLP components''': if you are creating a new voice from scratch, for example following the steps in [http://mary.opendfki.de/wiki/NewLanguageSupport NewLanguageSupport], you will need to create Minimal NLP components for the new language. These minimal components are necessary to run the MARY server in the new language and extract context features ('''phonefeatures''' directory). 
     242 
     243- '''Phoneme set''':  contained in $MARY_BASE/lib/modules/xx/lexicon/allophones.xx.xml , where xx corresponds to the new language. 
     244 
     245- After creating the minimal components, you will need wav files (in a wav directory) and the corresponding transcriptions (one file per wav file in a text directory). [[BR]] 
     246Afterwards follow the instructions as normal from step 1. Provide general settings for: 
    202247{{{ 
    203248   db.gender    =  male  (or female) 
    204    db.locale    =  de 
     249   db.locale    =  new_language locale (according to your minimal NLP components, ex. tr for Turkish, te for Telugu, etc.) 
    205250   db.marybase  =  /path/to/mary/base/ 
    206    db.voicename =  german_voice 
    207 }}} 
     251   db.voicename =  new_language_voice_name 
     252}}} 
     253 
     254 
     255   
    208256 
    209257[[BR]] 
    … …  
    211259 
    212260Marcela Charfuelan[[BR]] 
    213 DFKI - Fri May  9 11:54:00 CEST 2008 
    214  
    215  
    216  
     261Tue Aug 31 16:27:09 CEST 2010