git.gag.com Git - debian/freetts/blob - com/sun/speech/freetts/en/us/CMUVoice.java

   1 /**
   2  * Portions Copyright 2003 Sun Microsystems, Inc.
   3  * Portions Copyright 1999-2001 Language Technologies Institute,
   4  * Carnegie Mellon University.
   5  * All Rights Reserved.  Use is subject to license terms.
   6  *
   7  * See the file "license.terms" for information on usage and
   8  * redistribution of this file, and for a DISCLAIMER OF ALL
   9  * WARRANTIES.
  10  */
  11 package com.sun.speech.freetts.en.us;
  12
  13 import java.io.IOException;
  14 import java.util.List;
  15 import java.util.Locale;
  16
  17 import com.sun.speech.freetts.Age;
  18 import com.sun.speech.freetts.FeatureSet;
  19 import com.sun.speech.freetts.Gender;
  20 import com.sun.speech.freetts.PartOfSpeech;
  21 import com.sun.speech.freetts.PartOfSpeechImpl;
  22 import com.sun.speech.freetts.PhoneDurations;
  23 import com.sun.speech.freetts.PhoneDurationsImpl;
  24 import com.sun.speech.freetts.PhoneSet;
  25 import com.sun.speech.freetts.PhoneSetImpl;
  26 import com.sun.speech.freetts.Segmenter;
  27 import com.sun.speech.freetts.Tokenizer;
  28 import com.sun.speech.freetts.UtteranceProcessor;
  29 import com.sun.speech.freetts.Voice;
  30 import com.sun.speech.freetts.cart.CARTImpl;
  31 import com.sun.speech.freetts.cart.Durator;
  32 import com.sun.speech.freetts.cart.Intonator;
  33 import com.sun.speech.freetts.cart.Phraser;
  34 import com.sun.speech.freetts.en.ContourGenerator;
  35 import com.sun.speech.freetts.en.PartOfSpeechTagger;
  36 import com.sun.speech.freetts.en.PauseGenerator;
  37 import com.sun.speech.freetts.relp.AudioOutput;
  38 import com.sun.speech.freetts.util.BulkTimer;
  39
  40 /**
  41  * Provides generic support for a CMU Voice
  42  */
  43 public abstract class CMUVoice extends Voice {
  44     private PhoneSet phoneSet;
  45
  46     /**
  47      * Creates a simple voice
  48      *
  49      * @param name the name of the voice
  50      * @param gender the gender of the voice
  51      * @param age the age of the voice
  52      * @param description a human-readable string providing a
  53      * description that can be displayed for the users.
  54      * @param locale the locale of the voice
  55      * @param domain the domain of this voice.  For example,
  56      * @param organization the organization which created the voice
  57      * &quot;general&quot;, &quot;time&quot;, or
  58      * &quot;weather&quot;.
  59      * @param lexicon the lexicon to load
  60      */
  61     public CMUVoice(String name, Gender gender,
  62             Age age, String description, Locale locale, String domain,
  63             String organization, CMULexicon lexicon) {
  64         super(name, gender, age, description, locale, domain,
  65                 organization);
  66         setLexicon(lexicon);
  67     }
  68
  69     // overrides Voice.loader
  70
  71     /**
  72      * Called by <code> Voice </code>  during loading, derived voices
  73      * should override this to provide customized loading.
  74      */
  75     protected void loader() throws IOException {
  76         setupFeatureSet();
  77         setupUtteranceProcessors();
  78         setupFeatureProcessors();
  79     }
  80
  81     /**
  82      * Sets up the FeatureSet for this Voice.
  83      *
  84      * @throws IOException if an I/O error occurs
  85      */
  86     protected void setupFeatureSet() throws IOException {
  87         BulkTimer.LOAD.start("FeatureSet");
  88         FeatureSet features = getFeatures();
  89         features.setString(FEATURE_SILENCE, "pau");
  90         features.setString("join_type", "simple_join");
  91         BulkTimer.LOAD.stop("FeatureSet");
  92     }
  93
  94
  95     /**
  96      * Sets up the utterance processors for this voice. Subclasses
  97      * should define this method to setup the utterance processors for
  98      * the voice.
  99      *
 100      * @throws IOException throws an IOException if an error occurs
 101      */
 102     protected void setupUtteranceProcessors() throws IOException {
 103         List processors = getUtteranceProcessors();
 104
 105         BulkTimer.LOAD.start("CartLoading");
 106         CARTImpl numbersCart = new CARTImpl(getResource("nums_cart.txt"));
 107         CARTImpl phrasingCart = new CARTImpl(getResource("phrasing_cart.txt"));
 108         CARTImpl accentCart = new CARTImpl(getResource("int_accent_cart.txt"));
 109         CARTImpl toneCart = new CARTImpl(getResource("int_tone_cart.txt"));
 110         CARTImpl durzCart = new CARTImpl(getResource("durz_cart.txt"));
 111         BulkTimer.LOAD.stop("CartLoading");
 112
 113         BulkTimer.LOAD.start("UtteranceProcessors");
 114         PhoneDurations phoneDurations = new PhoneDurationsImpl(
 115             getResource("dur_stat.txt"));
 116         PronounceableFSM prefixFSM = new PrefixFSM
 117             (getResource("prefix_fsm.txt"));
 118         PronounceableFSM suffixFSM = new SuffixFSM
 119             (getResource("suffix_fsm.txt"));
 120
 121         processors.add(new TokenToWords(numbersCart, prefixFSM, suffixFSM));
 122         processors.add(new PartOfSpeechTagger());
 123         processors.add(new Phraser(phrasingCart));
 124         processors.add(new Segmenter());
 125         processors.add(new PauseGenerator());
 126         processors.add(new Intonator(accentCart, toneCart));
 127         processors.add(getPostLexicalAnalyzer());
 128         processors.add(new Durator(durzCart, 150.0f, phoneDurations));
 129         processors.add(new ContourGenerator
 130            (getResource("f0_lr_terms.txt"), 170.0f, 34.0f));
 131
 132
 133         processors.add(getUnitSelector());
 134         processors.add(getPitchmarkGenerator());
 135         processors.add(getUnitConcatenator());
 136         BulkTimer.LOAD.stop("UtteranceProcessors");
 137     }
 138
 139     // [[[TODO: currently a CMUVoice only allows customization of
 140     // the postlex, unit selector and wave synthesizer. This may
 141     // grow as time goes on ]]]
 142     /**
 143      * Returns the post lexical processor to be used by this voice.
 144      * Derived voices typically override this to customize behaviors.
 145      *
 146      * @return the post lexical analyzer in use by this voice
 147      *
 148      * @throws IOException if an IO error occurs while getting
 149      *     processor
 150      */
 151     protected UtteranceProcessor getPostLexicalAnalyzer() throws IOException {
 152         return new com.sun.speech.freetts.en.PostLexicalAnalyzer();
 153     }
 154
 155     /**
 156      * Returns the unit selector to be used by this voice
 157      * Derived voices typically override this to customize behaviors.
 158      *
 159      * @return the unit selector in use by this voice
 160      *
 161      * @throws IOException if an IO error occurs while getting
 162      *     processor
 163      */
 164     protected UtteranceProcessor getUnitSelector() throws IOException {
 165         return null;
 166     }
 167
 168     /**
 169      * Returns the pitch mark generator to be used by this voice
 170      * Derived voices typically override this to customize behaviors.
 171      *
 172      * @return the pitch mark generator to be used by this voice
 173      *
 174      * @throws IOException if an IO error occurs while getting
 175      *     processor
 176      */
 177     protected UtteranceProcessor getPitchmarkGenerator() throws IOException {
 178         return null;
 179     }
 180
 181     /**
 182      * Returns the unit concatenator to be used by this voice
 183      * Derived voices typically override this to customize behaviors.
 184      *
 185      * @return the Unit concatenator
 186      *
 187      * @throws IOException if an IO error occurs while getting
 188      *     processor
 189      */
 190     protected UtteranceProcessor getUnitConcatenator() throws IOException {
 191         return null;
 192     }
 193
 194
 195     /**
 196      * Sets up the FeatureProcessors for this Voice.
 197      *
 198      * @throws IOException if an I/O error occurs
 199      */
 200     protected void setupFeatureProcessors() throws IOException {
 201         BulkTimer.LOAD.start("FeatureProcessing");
 202         PartOfSpeech pos = new PartOfSpeechImpl(
 203             getResource("part_of_speech.txt"),
 204             "content");
 205
 206         phoneSet  = new PhoneSetImpl(getResource("phoneset.txt"));
 207
 208         addFeatureProcessor("word_break", new FeatureProcessors.WordBreak());
 209         addFeatureProcessor("word_punc", new FeatureProcessors.WordPunc());
 210         addFeatureProcessor("gpos", new FeatureProcessors.Gpos(pos));
 211         addFeatureProcessor("word_numsyls",new FeatureProcessors.WordNumSyls());
 212         addFeatureProcessor("ssyl_in", new FeatureProcessors.StressedSylIn());
 213         addFeatureProcessor("syl_in", new FeatureProcessors.SylIn());
 214         addFeatureProcessor("syl_out", new FeatureProcessors.SylOut());
 215         addFeatureProcessor("ssyl_out", new
 216                 FeatureProcessors.StressedSylOut());
 217         addFeatureProcessor("syl_break", new FeatureProcessors.SylBreak());
 218         addFeatureProcessor("old_syl_break", new FeatureProcessors.SylBreak());
 219         addFeatureProcessor("num_digits", new FeatureProcessors.NumDigits());
 220         addFeatureProcessor("month_range", new FeatureProcessors.MonthRange());
 221         addFeatureProcessor("token_pos_guess",
 222                 new FeatureProcessors.TokenPosGuess());
 223         addFeatureProcessor("segment_duration",
 224                 new FeatureProcessors.SegmentDuration());
 225         addFeatureProcessor("sub_phrases", new FeatureProcessors.SubPhrases());
 226         addFeatureProcessor("asyl_in", new FeatureProcessors.AccentedSylIn());
 227         addFeatureProcessor("last_accent", new FeatureProcessors.LastAccent());
 228         addFeatureProcessor("pos_in_syl", new FeatureProcessors.PosInSyl());
 229         addFeatureProcessor("position_type", new
 230                 FeatureProcessors.PositionType());
 231
 232         addFeatureProcessor("ph_cplace", new FeatureProcessors.PH_CPlace());
 233         addFeatureProcessor("ph_ctype", new FeatureProcessors.PH_CType());
 234         addFeatureProcessor("ph_cvox", new FeatureProcessors.PH_CVox());
 235         addFeatureProcessor("ph_vc", new FeatureProcessors.PH_VC());
 236         addFeatureProcessor("ph_vfront", new FeatureProcessors.PH_VFront());
 237         addFeatureProcessor("ph_vheight", new FeatureProcessors.PH_VHeight());
 238         addFeatureProcessor("ph_vlng", new FeatureProcessors.PH_VLength());
 239         addFeatureProcessor("ph_vrnd", new FeatureProcessors.PH_VRnd());
 240
 241         addFeatureProcessor("seg_coda_fric", new
 242                 FeatureProcessors.SegCodaFric());
 243         addFeatureProcessor("seg_onset_fric", new
 244                 FeatureProcessors.SegOnsetFric());
 245
 246         addFeatureProcessor("seg_coda_stop", new
 247                 FeatureProcessors.SegCodaStop());
 248         addFeatureProcessor("seg_onset_stop", new
 249                 FeatureProcessors.SegOnsetStop());
 250
 251         addFeatureProcessor("seg_coda_nasal", new
 252                 FeatureProcessors.SegCodaNasal());
 253         addFeatureProcessor("seg_onset_nasal", new
 254                 FeatureProcessors.SegOnsetNasal());
 255
 256         addFeatureProcessor("seg_coda_glide", new
 257                 FeatureProcessors.SegCodaGlide());
 258         addFeatureProcessor("seg_onset_glide", new
 259                 FeatureProcessors.SegOnsetGlide());
 260
 261         addFeatureProcessor("seg_onsetcoda", new
 262                 FeatureProcessors.SegOnsetCoda());
 263         addFeatureProcessor("syl_codasize", new
 264                 FeatureProcessors.SylCodaSize());
 265         addFeatureProcessor("syl_onsetsize", new
 266                 FeatureProcessors.SylOnsetSize());
 267         addFeatureProcessor("accented", new FeatureProcessors.Accented());
 268         BulkTimer.LOAD.stop("FeatureProcessing");
 269     }
 270
 271     /**
 272      * Given a phoneme and a feature name, return the feature
 273      *
 274      * @param phone the phoneme of interest
 275      * @param featureName the name of the feature of interest
 276      *
 277      * @return the feature with the given name
 278      */
 279     public String getPhoneFeature(String phone, String featureName) {
 280         return phoneSet.getPhoneFeature(phone, featureName);
 281     }
 282
 283     /**
 284      * Returns the AudioOutput processor to be used by this voice
 285      * Derived voices typically override this to customize behaviors.
 286      *
 287      * @return the audio output processor
 288      *
 289      * @throws IOException if an IO error occurs while getting
 290      *     processor
 291      */
 292     protected UtteranceProcessor getAudioOutput() throws IOException {
 293         return new AudioOutput();
 294     }
 295
 296     /**
 297      * Gets a tokenizer for this voice
 298      *
 299      * @return the tokenizer
 300      */
 301     public Tokenizer getTokenizer() {
 302         Tokenizer tokenizer = new com.sun.speech.freetts.en.TokenizerImpl();
 303         tokenizer.setWhitespaceSymbols(USEnglish.WHITESPACE_SYMBOLS);
 304         tokenizer.setSingleCharSymbols(USEnglish.SINGLE_CHAR_SYMBOLS);
 305         tokenizer.setPrepunctuationSymbols(USEnglish.PREPUNCTUATION_SYMBOLS);
 306         tokenizer.setPostpunctuationSymbols(USEnglish.PUNCTUATION_SYMBOLS);
 307         return tokenizer;
 308     }
 309
 310     /**
 311      * Converts this object to its String representation
 312      *
 313      * @return the string representation of this object
 314      */
 315     public String toString() {
 316         return "CMUVoice";
 317     }
 318 }
 319