Verfügbarkeit: Text-to-speech synthesis

Text-to-speech synthesis:

Gespeichert in:

Bibliographische Detailangaben
1. Verfasser:	Taylor, Paul (VerfasserIn)
Format:	Buch
Sprache:	German
Veröffentlicht:	Cambridge [u.a.] Cambridge Univ. Pr. 2009
Ausgabe:	1. publ.
Schlagworte:	Sprachcodierung Synthèse automatique de la parole Speech synthesis
Online-Zugang:	Inhaltsverzeichnis
Beschreibung:	Literaturverz. S. [556] - 582
Beschreibung:	XXVIII, 597 S. graph. Darst.
ISBN:	9780521899277

Internformat

MARC


LEADER	00000nam a2200000 c 4500
001	BV035039831
003	DE-604
005	20090212
007	t
008	080905s2009 d\|\|\| \|\|\|\| 00\|\|\| ger d
020			\|a 9780521899277 \|9 978-0-521-89927-7
035			\|a (OCoLC)221147648
035			\|a (DE-599)BVBBV035039831
040			\|a DE-604 \|b ger \|e rakwb
041	0		\|a ger
049			\|a DE-19 \|a DE-11
050		0	\|a TK7882.S65
082	0		\|a 006.54 \|2 22
084			\|a ES 950 \|0 (DE-625)27936: \|2 rvk
100	1		\|a Taylor, Paul \|e Verfasser \|4 aut
245	1	0	\|a Text-to-speech synthesis \|c Paul Taylor
250			\|a 1. publ.
264		1	\|a Cambridge [u.a.] \|b Cambridge Univ. Pr. \|c 2009
300			\|a XXVIII, 597 S. \|b graph. Darst.
336			\|b txt \|2 rdacontent
337			\|b n \|2 rdamedia
338			\|b nc \|2 rdacarrier
500			\|a Literaturverz. S. [556] - 582
650		7	\|a Sprachcodierung \|2 swd
650		7	\|a Synthèse automatique de la parole \|2 ram
650		4	\|a Speech synthesis
856	4	2	\|m HEBIS Datenaustausch \|q application/pdf \|u http://bvbr.bib-bvb.de:8991/F?func=service&doc_library=BVB01&local_base=BVB01&doc_number=016708674&sequence=000001&line_number=0001&func_code=DB_RECORDS&service_type=MEDIA \|3 Inhaltsverzeichnis
999			\|a oai:aleph.bib-bvb.de:BVB01-016708674

Datensatz im Suchindex

_version_	1804137974089121792
adam_text	TEXT-TO-SPEECH SYNTHESIS PAUL TAYLOR UNIVERSITY OF CAMBRIDGE CAMBRIDGE UNIVERSITY PRESS CONTENTS FOREWORD PAGE XXIII PREFACE XXVII INTRODUCTION 1 1.1 WHAT ARE TEXT-TO-SPEECH SYSTEMS FOR? 2 1.2 WHAT SHOULD THE GOALS OF TEXT-TO-SPEECH SYSTEM DEVELOPMENT BE? 3 1.3 THE ENGINEERING APPROACH 4 1.4 OVERVIEW OF THE BOOK 5 1.4.1 VIEWPOINTS WITHIN THE BOOK 5 1.4.2 READERS BACKGROUNDS 6 1.4.3 BACKGROUND AND SPECIALIST SECTIONS 7 COMMUNICATION AND LANGUAGE 8 9 10 11 12 13 14 16 17 18 18 19 20 21 22 22 23 24 2.1 2.2 2.3 2.4 2.5 TYPES 2.1.1 2.1.2 2.1.3 2.1.4 2.1.5 OF COMMUNICATION AFFECTIVE COMMUNICATION ICONIC COMMUNICATION SYMBOLIC COMMUNICATION COMBINATIONS OF SYMBOLS MEANING, FORM AND SIGNAL HUMAN COMMUNICATION 2.2.1 2.2.2 2.2.3 2.2.4 VERBAL COMMUNICATION LINGUISTIC LEVELS AFFECTIVE PROSODY AUGMENTATIVE PROSODY COMMUNICATION PROCESSES 2.3.1 2.3.2 2.3.3 2.3.4 2.3.5 COMMUNICATION FACTORS GENERATION ENCODING DECODING UNDERSTANDING DISCUSSION SUMMARY CONTENTS THE TEXT-TO-SPEECH PROBLEM 26 3.1 SPEECH AND WRITING 26 3.1.1 PHYSICAL NATURE 27 3.1.2 SPOKEN FORM AND WRITTEN FORM 28 3.1.3 USE 29 3.1.4 PROSODIC AND VERBAL CONTENT 30 3.1.5 COMPONENT BALANCE 31 3.1.6 NON-LINGUISTIC CONTENT 32 3.1.7 SEMIOTIC SYSTEMS 33 3.1.8 WRITING SYSTEMS 34 3.2 READING ALOUD 35 3.2.1 READING SILENTLY AND READING ALOUD 35 3.2.2 PROSODY IN READING ALOUD 36 3.2.3 VERBAL CONTENT AND STYLE IN READING ALOUD 37 3.3 TEXT-TO-SPEECH SYSTEM ORGANISATION 37 3.3.1 THE COMMON-FORM MODEL 38 3.3.2 OTHER MODELS 39 3.3.3 COMPARISON 40 3.4 SYSTEMS 41 3.4.1 A SIMPLE TEXT-TO-SPEECH SYSTEM 41 3.4.2 CONCEPT TO SPEECH 42 3.4.3 CANNED SPEECH AND LIMITED-DOMAIN SYNTHESIS 43 3.5 KEY PROBLEMS IN TEXT-TO-SPEECH 44 3.5.1 TEXT CLASSIFICATION WITH RESPECT TO SEMIOTIC SYSTEMS 44 3.5.2 DECODING NATURAL-LANGUAGE TEXT 46 3.5.3 NATURALNESS 47 3.5.4 INTELLIGIBILITY: ENCODING THE MESSAGE IN SIGNAL 48 3.5.5 AUXILIARY GENERATION FOR PROSODY 49 3.5.6 ADAPTING THE SYSTEM TO THE SITUATION 50 3.6 SUMMARY 50 TEXT SEGMENTATION AND ORGANISATION 52 52 53 53 55 59 60 61 61 62 63 63 4.1 OVERVIEW OF THE PROBLEM 4.2 WORDS 4.2 4.2 4.2 4.2 4.2 4.2 4.2 4.2 .1 .2 .3 .4 .5 .6 .7 .8 4.3 TEXT SE AND SENTENCES WHAT IS A WORD? DEFINING WORDS IN TEXT-TO-SPEECH SCOPE AND MORPHOLOGY CONTRACTIONS AND CLITICS SLANG FORMS HYPHENATED FORMS WHAT IS A SENTENCE? THE LEXICON GMENTATION CONTENTS 4.3.1 TOKENISATION 64 4.3.2 TOKENISATION AND PUNCTUATION 65 4.3.3 TOKENISATION ALGORITHMS 66 4.3.4 SENTENCE SPLITTING 67 4.4 PROCESSING DOCUMENTS 68 4.4.1 MARKUP LANGUAGES 68 4.4.2 INTERPRETING CHARACTERS 70 4.5 TEXT-TO-SPEECH ARCHITECTURES 71 4.6 DISCUSSION 75 4.6.1 FURTHER READING 75 4.6.2 SUMMARY 76 TEXT DECODING: FINDING THE WORDS FROM THE TEXT 78 5.1 OVERVIEW OF TEXT DECODING 78 5.2 TEXT-CLASSIFICATION ALGORITHMS 79 5.2.1 FEATURES AND ALGORITHMS 79 5.2.2 TAGGING AND WORD-SENSE DISAMBIGUATION 82 5.2.3 AD-HOC APPROACHES 83 5.2.4 DETERMINISTIC RULE APPROACHES 83 5.2.5 DECISION LISTS , 85 5.2.6 NAIVE BAYES CLASSIFIER 86 5.2.7 DECISION TREES 87 5.2.8 PART-OF-SPEECH TAGGING 88 5.3 NON-NATURAL-LANGUAGE TEXT 92 5.3.1 SEMIOTIC CLASSIFICATION 92 5.3.2 SEMIOTIC DECODING 95 5.3.3 VERBALISATION 95 5.4 NATURAL-LANGUAGE TEXT 97 5.4.1 ACRONYMS AND LETTER SEQUENCES 99 5.4.2 HOMOGRAPH DISAMBIGUATION 99 5.4.3 NON-HOMOGRAPHS 101 5.5 NATURAL-LANGUAGE PARSING 102 5.5.1 CONTEXT-FREE GRAMMARS 102 5.5.2 STATISTICAL PARSING 105 5.6 DISCUSSION 105 5.6.1 FURTHER READING 108 5.6.2 SUMMARY 109 PROSODY PREDICTION FROM TEXT 111 6.1 PROSODIC FORM 111 6.2 PHRASING 112 6.2.1 PHRASING PHENOMENA 112 6.2.2 MODELS OF PHRASING 113 XII CONTENTS 6.3 PROMINENCE 115 6.3.1 SYNTACTIC PROMINENCE PATTERNS 116 6.3.2 DISCOURSE PROMINENCE PATTERNS 118 6.3.3 PROMINENCE SYSTEMS, DATA AND LABELLING 119 6.4 INTONATION AND TUNE 121 6.5 PROSODIC MEANING AND FUNCTION 122 6.5.1 AFFECTIVE PROSODY 123 6.5.2 SUPRASEGMENTALITY 124 6.5.3 AUGMENTATIVE PROSODY 125 6.5.4 SYMBOLIC COMMUNICATION AND PROSODIC STYLE 126 6.6 DETERMINING PROSODY FROM THE TEXT 127 6.6.1 PROSODY AND HUMAN READING 127 6.6.2 CONTROLLING THE-DEGREE OF AUGMENTATIVE PROSODY 128 6.6.3 PROSODY AND SYNTHESIS TECHNIQUES 128 6.7 PHRASING PREDICTION 129 6.7.1 EXPERIMENTAL FORMULATION 129 6.7.2 DETERMINISTIC APPROACHES 130 6.7.3 CLASSIFIER APPROACHES 132 6.7.4 HMM APPROACHES 133 6.7.5 HYBRID APPROACHES 135 6.8 PROMINENCE PREDICTION . 136 6.8.1 COMPOUND-NOUN PHRASES 136 6.8.2 FUNCTION-WORD PROMINENCE 138 6.8.3 DATA-DRIVEN APPROACHES 138 6.9 INTONATIONAL-TUNE PREDICTION 139 6.10 DISCUSSION 139 6.10.1 LABELLING SCHEMES AND LABELLING ACCURACY 139 6.10.2 LINGUISTIC THEORIES AND PROSODY 141 6.10.3 SYNTHESISING SUPRASEGMENTAL AND TRUE PROSODY 142 6.10.4 PROSODY IN REAL DIALOGUES 143 6.10.5 CONCLUSION 144 6.10.6 SUMMARY 144 PHONETICS AND PHONOLOGY 146 7.1 ARTICULATORY PHONETICS AND SPEECH PRODUCTION 146 7.1.1 THE VOCAL ORGANS 147 7.1.2 SOUND SOURCES 147 7.1.3 SOUND OUTPUT 150 7.1.4 THE VOCAL-TRACT FILTER 150 7.1.5 VOWELS 151 7.1.6 CONSONANTS 153 7.1.7 EXAMINING SPEECH PRODUCTION 155 7.2 ACOUSTICS, PHONETICS AND SPEECH PERCEPTION 156 CONTENTS XIII 7.3 7.4 7.5 7.2.1 ACOUSTIC REPRESENTATIONS 7.2.2 ACOUSTIC CHARACTERISTICS THE COMMUNICATIVE USE OF SPEECH 7.3.1 COMMUNICATING DISCRETE INFORMATION WITH A CONTINUOUS CHANNEL 7.3.2 PHONEMES, PHONES AND ALLOPHONES 7.3.3 ALLOPHONIC VARIATION AND PHONETIC CONTEXT 7.3.4 COARTICULATION, TARGETS AND TRANSIENTS 7.3.5 THE CONTINUOUS NATURE OF SPEECH 7.3.6 TRANSCRIPTION 7.3.7 THE DISTINCTIVENESS OF SPEECH IN COMMUNICATION PHONOLOGY: THE LINGUISTIC ORGANISATION OF SPEECH 7.4.1 PHONOTACTICS 7.4.2 WORD FORMATION 7.4.3 DISTINCTIVE FEATURES AND PHONOLOGICAL THEORIES 7.4.4 SYLLABLES 7.4.5 LEXICAL STRESS DISCUSSION 7.5.1 FURTHER READING 7.5.2 SUMMARY 8 PRONUNCIATION 8.1 8.2 8.3 8.4 PRONUNCIATION REPRESENTATIONS 8.1.1 WHY BOTHER? 8.1.2 PHONEMIC AND PHONETIC INPUT 8.1.3 DIFFICULTIES IN DERIVING PHONETIC INPUT 8.1.4 A STRUCTURED APPROACH TO PRONUNCIATION 8.1.5 ABSTRACT PHONOLOGICAL REPRESENTATIONS FORMULATING A PHONOLOGICAL REPRESENTATION SYSTEM 8.2.1 SIMPLE CONSONANTS AND VOWELS 8.2.2 DIFFICULT CONSONANTS 8.2.3 DIPHTHONGS AND AFFRICATES 8.2.4 APPROXIMANT-VOWEL COMBINATIONS 8.2.5 DEFINING THE FULL INVENTORY 8.2.6 PHONEME NAMES 8.2.7 SYLLABIC ISSUES THE LEXICON 8.3.1 LEXICON AND RULES 8.3.2 LEXICON FORMATS 8.3.3 THE OFFLINE LEXICON 8.3.4 THE SYSTEM LEXICON 8.3.5 LEXICON QUALITY 8.3.6 DETERMINING THE PRONUNCIATIONS OF UNKNOWN WORDS GRANHEME-TO-RJHONEME CONVERSION 156 159 160 161 162 166 168 169 170 171 172 172 179 181 184 186 189 189 190 192 192 192 193 194 195 196 197 197 199 201 201 203 204 206 207 208 210 213 214 215 216 218 XIV CONTENTS 8.4.1 8.4.2 8.4.3 8.4.4 8.4.5 8.4.6 RULE-BASED TECHNIQUES GRAPHEME-TO-PHONEME ALIGNMENT NEURAL NETWORKS PRONUNCIATION BY ANALOGY OTHER DATA-DRIVEN TECHNIQUES STATISTICAL TECHNIQUES 8.5 FURTHER ISSUES 8.5.1 8.5.2 8.5.3 MORPHOLOGY LANGUAGE ORIGIN AND NAMES POST-LEXICAL PROCESSING 8.6 SUMMARY 218 219 219 220 221 221 222 222 223 223 224 SYNTHESIS OF PROSODY 225 9.1 INTONATION OVERVIEW 225 9.1.1 F0 AND PITCH 226 9.1.2 INTONATIONAL FORM 226 9.1.3 MODELS OF F0 CONTOURS 227 9.1.4 MICRO-PROSODY 229 9.2 INTONATIONAL BEHAVIOUR 229 9.2.1 INTONATIONAL TUNE 229 9.2.2 DOWNDRIFT 230 9.2.3 PITCH RANGE 233 9.2.4 PITCH ACCENTS AND BOUNDARY TONES 234 9.3 INTONATION THEORIES AND MODELS 236 9.3.1 TRADITIONAL MODELS AND THE BRITISH SCHOOL 236 9.3.2 THE DUTCH SCHOOL 237 9.3.3 AUTOSEGMENTAL-METRICAL AND TOBI MODELS 237 9.3.4 THE INTSINT MODEL 239 9.3.5 THE FUJISAKI MODEL AND SUPERIMPOSITIONAL MODELS 239 9.3.6 THE TILT MODEL 242 9.3.7 COMPARISON 244 9.4 INTONATION SYNTHESIS WITH AM MODELS 245 9.4.1 PREDICTION OF AM LABELS FROM TEXT 246 9.4.2 DETERMINISTIC SYNTHESIS METHODS 246 9.4.3 DATA-DRIVEN SYNTHESIS METHODS 247 9.4.4 ANALYSIS WITH AUTOSEGMENTAL MODELS 248 9.5 INTONATION SYNTHESIS WITH DETERMINISTIC ACOUSTIC MODELS 248 9.5.1 SYNTHESIS WITH SUPERIMPOSITIONAL MODELS 249 9.5.2 SYNTHESIS WITH THE TILT MODEL 249 9.5.3 ANALYSIS WITH FUJISAKI AND TILT MODELS 250 9.6 DATA-DRIVEN INTONATION MODELS 250 9.6.1 UNIT-SELECTION-STYLE APPROACHES 251 9.6.2 DYNAMIC-SYSTEM MODELS 252 CONTENTS XV 9.7 9.8 9.6.3 9.6.4 HIDDEN MARKOV MODELS FUNCTIONAL MODELS TIMING 9.7.1 9.7.2 9.7.3 9.7.4 9.7.5 9.7.6 FORMULATION OF THE TIMING PROBLEM THE NATURE OF TIMING KLATT RULES THE SUMS-OF-PRODUCTS MODEL THE CAMPBELL MODEL OTHER REGRESSION TECHNIQUES DISCUSSION 9.8.1 9.8.2 SIGNALS AND FURTHER READING SUMMARY FILTERS 253 254 254 255 255 256 257 258 259 259 260 260 10 SIGNALS AND FILTERS 262 10.1 ANALOGUE SIGNALS 262 10.1.1 SIMPLE PERIODIC SIGNALS: SINUSOIDS 263 10.1.2 GENERAL PERIODIC SIGNALS 265 10.1.3 SINUSOIDS AS COMPLEX EXPONENTIALS 266 10.1.4 FOURIER ANALYSIS 269 10.1.5 THE FREQUENCY DOMAIN 270 10.1.6 THE FOURIER TRANSFORM 275 10.2 DIGITAL SIGNALS 278 10.2.1 DIGITAL WAVEFORMS 279 10.2.2 DIGITAL REPRESENTATIONS 280 10.2.3 THE DISCRETE-TIME FOURIER TRANSFORM 280 10.2.4 THE DISCRETE FOURIER TRANSFORM 281 10.2.5 THE Z-TRANSFORM 282 10.2.6 THE FREQUENCY DOMAIN FOR DIGITAL SIGNALS 283 10.3 PROPERTIES OF TRANSFORMS 284 10.3.1 LINEARITY 284 10.3.2 TIME AND FREQUENCY DUALITY 284 10.3.3 SCALING 285 10.3.4 IMPULSE PROPERTIES 285 10.3.5 TIME DELAY 286 10.3.6 FREQUENCY SHIFT 286 10.3.7 CONVOLUTION 287 10.3.8 ANALYTICAL AND NUMERICAL ANALYSIS 287 10.3.9 STOCHASTIC SIGNALS 288 10.4 DIGITAL FILTERS 288 10.4.1 DIFFERENCE EQUATIONS 289 10.4.2 THE IMPULSE RESPONSE 289 10.4.3 THE FILTER CONVOLUTION SUM 292 10.4.4 THE FILTER TRANSFER FUNCTION 293 XVI CONTENTS 10.4.5 THE TRANSFER FUNCTION AND THE IMPULSE RESPONSE 293 10.5 DIGITAL FILTER ANALYSIS AND DESIGN 294 10.5.1 POLYNOMIAL ANALYSIS: POLES AND ZEROS 294 10.5.2 FREQUENCY INTERPRETATION OF THE Z-DOMAIN TRANSFER FUNCTION 297 10.5.3 FILTER CHARACTERISTICS 298 10.5.4 PUTTING IT ALL TOGETHER 304 10.6 SUMMARY 305 11 ACOUSTIC MODELS OF SPEECH PRODUCTION 309 11.1 THE ACOUSTIC THEORY OF SPEECH PRODUCTION 309 11.1.1 COMPONENTS IN THE MODEL 309 11.2 THE PHYSICS OF SOUND 311 11.2.1 RESONANT SYSTEMS 311 11.2.2 TRAVELLING WAVES 313 11.2.3 ACOUSTIC WAVES 315 11.2.4 ACOUSTIC REFLECTION 317 11.3 THE VOWEL-TUBE MODEL 318 11.3.1 DISCRETE TIME AND DISTANCE 319 11.3.2 JUNCTION OF TWO TUBES 320 11.3.3 SPECIAL CASES OF JUNCTIONS 322 11.3.4 THE TWO-TUBE VOCAL-TRACT MODEL 323 11.3.5 THE SINGLE-TUBE MODEL 325 11.3.6 THE MULTI-TUBE VOCAL-TRACT MODEL 327 11.3.7 THE ALL-POLE RESONATOR MODEL 329 11.4 SOURCE AND RADIATION MODELS 330 11.4.1 RADIATION 330 11.4.2 THE GLOTTAL SOURCE 330 11.5 MODEL REFINEMENTS 333 11.5.1 MODELLING THE NASAL CAVITY 333 11.5.2 SOURCE POSITIONS IN THE ORAL CAVITY 334 11.5.3 MODELS WITH VOCAL-TRACT LOSSES 335 11.5.4 SOURCE AND RADIATION EFFECTS 336 11.6 DISCUSSION 336 11.6.1 FURTHER READING 339 11.6.2 SUMMARY 339 12 ANALYSIS OF SPEECH SIGNALS 341 12.1 SHORT-TERM SPEECH ANALYSIS 341 12.1.1 WINDOWING 342 12.1.2 SHORT-TERM SPECTRAL REPRESENTATIONS 343 12.1.3 FRAME LENGTHS AND SHIFTS 345 12.1.4 THE SPECTROGRAM 348 12.1.5 AUDITORY SCALES 351 CONTENTS XVII 12.2 FILTER-BANK ANALYSIS 352 12.3 THE CEPSTRUM 353 12.3.1 CEPSTRUM DEFINITION 353 12.3.2 TREATING THE MAGNITUDE SPECTRUM AS A SIGNAL 353 12.3.3 CEPSTRAL ANALYSIS AS DECONVOLUTION 355 12.3.4 CEPSTRAL ANALYSIS DISCUSSION 356 12.4 LINEAR-PREDICTION ANALYSIS 357 12.4.1 FINDING THE COEFFICIENTS: THE COVARIANCE METHOD 358 12.4.2 THE AUTOCORRELATION METHOD 360 12.4.3 LEVINSON-DURBIN RECURSION 361 12.5 SPECTRAL-ENVELOPE AND VOCAL-TRACT REPRESENTATIONS 362 12.5.1 LINEAR-PREDICTION SPECTRA 362 12.5.2 TRANSFER-FUNCTION POLES 364 12.5.3 REFLECTION COEFFICIENTS 366 12.5.4 LOG AREA RATIOS 367 12.5.5 LINE-SPECTRUM FREQUENCIES 367 12.5.6 LINEAR-PREDICTION CEPSTRA 369 12.5.7 MEL-SCALED CEPSTRA 370 12.5.8 PERCEPTUAL LINEAR PREDICTION 370 12.5.9 FORMANT TRACKING 370 12.6 SOURCE REPRESENTATIONS , 372 12.6.1 RESIDUAL SIGNALS 372 12.6.2 CLOSED-PHASE ANALYSIS 374 12.6.3 OPEN-PHASE ANALYSIS 377 12.6.4 IMPULSE/NOISE MODELS 378 12.6.5 PARAMETERISATION OF GLOTTAL-FLOW SIGNALS 379 12.7 PITCH AND EPOCH DETECTION 379 12.7.1 PITCH DETECTION 379 12.7.2 EPOCH DETECTION: FINDING THE INSTANT OF GLOTTAL CLOSURE 381 12.8 DISCUSSION 384 12.8.1 FURTHER READING 385 12.8.2 SUMMARY 386 13 SYNTHESIS TECHNIQUES BASED ON VOCAL-TRACT MODELS 387 13.1 SYNTHESIS SPECIFICATION: THE INPUT TO THE SYNTHESISER 387 13.2 FORMANT SYNTHESIS 388 13.2.1 SOUND SOURCES 389 13.2.2 SYNTHESISING A SINGLE FORMANT 390 13.2.3 RESONATORS IN SERIES AND PARALLEL 391 13.2.4 SYNTHESISING CONSONANTS 392 13.2.5 A COMPLETE SYNTHESISER 394 13.2.6 THE PHONETIC INPUT TO THE SYNTHESISER 394 13.2.7 FORMANT-SYNTHESIS QUALITY 397 XVIII CONTENTS 13.3 CLASSICAL LINEAR-PREDICTION SYNTHESIS 13.3.1 COMPARISON WITH FORMANT SYNTHESIS 13.3.2 THE IMPULSE/NOISE SOURCE MODEL 13.3.3 LINEAR-PREDICTION DIPHONE-CONCATENATIVE SYNTHESIS 13.3.4 A COMPLETE SYNTHESISER 13.3.5 PROBLEMS WITH THE SOURCE 13.4 ARTICULATORY SYNTHESIS 13.5 DISCUSSION 13.5.1 FURTHER READING 13.5.2 SUMMARY 14 SYNTHESIS BY CONCATENATION AND SIGNAL-PROCESSING MODIFICATION 14.1 SPEECH UNITS IN SECOND-GENERATION SYSTEMS 14.1.1 CREATING A DIPHONE INVENTORY 14.1.2 OBTAINING DIPHONES FROM SPEECH 14.2 PITCH-SYNCHRONOUS OVERLAP AND ADD (PSOLA) 14.2.1 TIME-DOMAIN PSOLA 14.2.2 EPOCH MANIPULATION 14.2.3 HOW DOES PSOLA WORK? 14.3 RESIDUAL-EXCITED LINEAR PREDICTION (RELP) 14.3.1 RESIDUAL MANIPULATION 14.3.2 LINEAR-PREDICTION PSOLA 14.4 SINUSOIDAL MODELS 14.4.1 PURE SINUSOIDAL MODELS 14.4.2 HARMONIC/NOISE MODELS 14.5 MBROLA 14.6 SYNTHESIS FROM CEPSTRAL COEFFICIENTS 14.7 CONCATENATION ISSUES 14.8 DISCUSSION 14.8.1 FURTHER READING 14.8.2 SUMMARY 15 HIDDEN-MARKOV-MODEL SYNTHESIS 15.1 THE HMM FORMALISM 15.1.1 OBSERVATION PROBABILITIES 15.1.2 DELTA COEFFICIENTS 15.1.3 ACOUSTIC REPRESENTATIONS AND COVARIANCE 15. 15. 15. 15. 15. .4 STATES AND TRANSITIONS .5 RECOGNISING WITH HMMS .6 LANGUAGE MODELS .7 THE VITERBI ALGORITHM .8 TRAINING HMMS 15.1.9 CONTEXT-SENSITIVE MODELLING 399 399 400 401 403 404 405 407 409 410 412 413 414 414 415 416 417 421 421 423 423 424 425 426 429 429 431 433 433 433 435 435 436 438 439 440 440 443 444 447 451 CONTENTS XIX 15.1.10 ARE HMMS A GOOD MODEL OF SPEECH? 454 15.2 SYNTHESIS FROM HIDDEN MARKOV MODELS 456 15.2.1 FINDING THE LIKELIEST OBSERVATIONS GIVEN THE STATE SEQUENCE 457 15.2.2 FINDING THE LIKELIEST OBSERVATIONS AND STATE SEQUENCE 459 15.2.3 ACOUSTIC REPRESENTATIONS 459 15.2.4 CONTEXT-SENSITIVE SYNTHESIS MODELS 463 15.2.5 DURATION MODELLING 464 15.2.6 SIGNAL PROCESSING IN HMM SYNTHESIS 464 15.2.7 HMM SYNTHESIS SYSTEMS 465 15.3 LABELLING DATABASES WITH HMMS 467 15.3.1 DETERMINING THE WORD SEQUENCE 467 15.3.2 DETERMINING THE PHONE SEQUENCE 468 15.3.3 DETERMINING THE PHONE BOUNDARIES 468 15.3.4 MEASURING THE QUALITY OF THE ALIGNMENTS 470 15.4 OTHER DATA-DRIVEN SYNTHESIS TECHNIQUES 471 15.5 DISCUSSION 471 15.5.1 FURTHER READING 471 15.5.2 SUMMARY 472 16 UNIT-SELECTION SYNTHESIS 474 16.1 FROM CONCATENATIVE SYNTHESIS TO UNIT SELECTION 474 16.1.1 EXTENDING CONCATENATIVE SYNTHESIS 475 16.1.2 THE HUNT AND BLACK ALGORITHM 477 16.2 FEATURES, 479 16.2.1 BASE TYPES 479 16.2.2 LINGUISTIC AND ACOUSTIC FEATURES 480 16.2.3 CHOICE OF FEATURES 481 16.2.4 TYPES OF FEATURES 482 16.3 THE INDEPENDENT-FEATURE TARGET-FUNCTION FORMULATION 484 16.3.1 THE PURPOSE OF THE TARGET FUNCTION 484 16.3.2 DEFINING A PERCEPTUAL SPACE 485 16.3.3 PERCEPTUAL SPACES DEFINED BY INDEPENDENT FEATURES 486 16.3.4 SETTING THE TARGET WEIGHTS USING ACOUSTIC DISTANCES 488 16.3.5 LIMITATIONS OF THE INDEPENDENT-FEATURE FORMULATION 491 16.4 THE ACOUSTIC-SPACE TARGET-FUNCTION FORMULATION 493 16.4.1 DECISION-TREE CLUSTERING 494 16.4.2 GENERAL PARTIAL-SYNTHESIS FUNCTIONS 496 16.5 JOIN FUNCTIONS 497 16.5.1 BASIC ISSUES IN JOINING UNITS 497 16.5.2 PHONE-CLASS JOIN COSTS 498 16.5.3 ACOUSTIC-DISTANCE JOIN COSTS 499 16.5.4 COMBINING CATEGORICAL AND AND ACOUSTIC JOIN COSTS 500 16.5.5 PROBABILISTIC AND SEQUENCE JOIN COSTS 501 XX CONTENTS 16.5.6 JOIN CLASSIFIERS 502 16.6 SEARCHING 504 16.6.1 BASE TYPES AND SEARCHING 505 16.6.2 PRUNING 508 16.6.3 PRE-SELECTION 508 16.6.4 BEAM PRUNING 509 16.6.5 MULTI-PASS SEARCHING 509 16.7 DISCUSSION 510 16.7.1 UNIT SELECTION AND SIGNAL PROCESSING 511 16.7.2 FEATURES, COSTS AND PERCEPTION 511 16.7.3 EXAMPLE UNIT-SELECTION SYSTEMS 512 16.7.4 FURTHER READING 514 16.7.5 SUMMARY 515 17 FURTHER ISSUES 517 17.1 DATABASES 517 17.1.1 UNIT-SELECTION DATABASES 517 17.1.2 TEXT MATERIALS 518 17.1.3 PROSODY DATABASES 518 17.1.4 LABELLING 519 17.1.5 WHAT EXACTLY IS HAND LABELLING? 519 17.1.6 AUTOMATIC LABELLING 521 17.1.7 AVOIDING EXPLICIT LABELS 521 17.2 EVALUATION 522 17.2.1 SYSTEM TESTING: INTELLIGIBILITY AND NATURALNESS 523 17.2.2 WORD-RECOGNITION TESTS 523 17.2.3 NATURALNESS TESTS 524 17.2.4 TEST DATA 525 17.2.5 UNIT OR COMPONENT TESTING 525 17.2.6 COMPETITIVE EVALUATIONS 526 17.3 AUDIO-VISUAL SPEECH SYNTHESIS 527 17.3.1 SPEECH CONTROL 528 17.4 SYNTHESIS OF EMOTIONAL AND EXPRESSIVE SPEECH 529 17.4.1 DESCRIBING EMOTION 529 17.4.2 SYNTHESISING EMOTION WITH PROSODY CONTROL 529 17.4.3 SYNTHESISING EMOTION WITH VOICE TRANSFORMATION 530 17.4.4 UNIT SELECTION AND HMM TECHNIQUES 531 17.5 SUMMARY 531 18 CONCLUSION 533 18.1 SPEECH TECHNOLOGY AND LINGUISTICS 533 18.2 FUTURE DIRECTIONS 536 18.3 CONCLUSION 539 CONTENTS XXI APPENDIX A PROBABILITY 540 A. 1 DISCRETE PROBABILITIES 540 A. 1.1 DISCRETE RANDOM VARIABLES 540 A. 1.2 PROBABILITY MASS FUNCTIONS 541 A. 1.3 EXPECTED VALUES 541 A. 1.4 MOMENTS OF A PMF 542 A.2 PAIRS OF DISCRETE RANDOM VARIABLES 542 A.2.1 MARGINAL DISTRIBUTIONS 543 A.2.2 INDEPENDENCE 543 A.2.3 EXPECTED VALUES 543 A.2.4 MOMENTS OF A JOINT DISTRIBUTION 544 A.2.5 HIGHER-ORDER MOMENTS AND COVARIANCE 544 A.2.6 CORRELATION 544 A.2.7 CONDITIONAL PROBABILITY 545 A.2.8 BAYES RULE 545 A.2.9 SUM OF RANDOM VARIABLES 545 A.2.10 THE CHAIN RULE 546 A.2.11 ENTROPY 546 A.3 CONTINUOUS RANDOM VARIABLES 547 A.3.1 CONTINUOUS RANDOM VARIABLES 548 A.3.2 EXPECTED VALUES 548 A.3.3 THE GAUSSIAN DISTRIBUTION 549 A.3.4 THE UNIFORM DISTRIBUTION 549 - A.3.5 CUMULATIVE DENSITY FUNCTIONS 549 A.4 PAIRS OF CONTINUOUS RANDOM VARIABLES 550 A.4.1 INDEPENDENT VERSUS UNCORRELATED 551 A.4.2 THE SUM OF TWO RANDOM VARIABLES 551 A.4.3 ENTROPY 552 A.4.4 KULLBACK-LEIBLER DISTANCE 552 APPENDIX B PHONE DEFINITIONS 553 REFERENCES 556 INDEX 583
adam_txt	TEXT-TO-SPEECH SYNTHESIS PAUL TAYLOR UNIVERSITY OF CAMBRIDGE CAMBRIDGE UNIVERSITY PRESS CONTENTS FOREWORD PAGE XXIII PREFACE XXVII INTRODUCTION 1 1.1 WHAT ARE TEXT-TO-SPEECH SYSTEMS FOR? 2 1.2 WHAT SHOULD THE GOALS OF TEXT-TO-SPEECH SYSTEM DEVELOPMENT BE? 3 1.3 THE ENGINEERING APPROACH 4 1.4 OVERVIEW OF THE BOOK 5 1.4.1 VIEWPOINTS WITHIN THE BOOK 5 1.4.2 READERS'BACKGROUNDS 6 1.4.3 BACKGROUND AND SPECIALIST SECTIONS ' 7 COMMUNICATION AND LANGUAGE 8 9 10 11 12 13 14 16 17 18 18 19 20 21 22 22 23 24 2.1 2.2 2.3 2.4 2.5 TYPES 2.1.1 2.1.2 2.1.3 2.1.4 2.1.5 OF COMMUNICATION AFFECTIVE COMMUNICATION ICONIC COMMUNICATION SYMBOLIC COMMUNICATION COMBINATIONS OF SYMBOLS MEANING, FORM AND SIGNAL HUMAN COMMUNICATION 2.2.1 2.2.2 2.2.3 2.2.4 VERBAL COMMUNICATION LINGUISTIC LEVELS AFFECTIVE PROSODY AUGMENTATIVE PROSODY COMMUNICATION PROCESSES 2.3.1 2.3.2 2.3.3 2.3.4 2.3.5 COMMUNICATION FACTORS GENERATION ENCODING DECODING UNDERSTANDING DISCUSSION SUMMARY CONTENTS THE TEXT-TO-SPEECH PROBLEM 26 3.1 SPEECH AND WRITING 26 3.1.1 PHYSICAL NATURE 27 3.1.2 SPOKEN FORM AND WRITTEN FORM 28 3.1.3 USE 29 3.1.4 PROSODIC AND VERBAL CONTENT 30 3.1.5 COMPONENT BALANCE 31 3.1.6 NON-LINGUISTIC CONTENT 32 3.1.7 SEMIOTIC SYSTEMS 33 3.1.8 WRITING SYSTEMS 34 3.2 READING ALOUD 35 3.2.1 READING SILENTLY AND READING ALOUD 35 3.2.2 PROSODY IN READING ALOUD 36 3.2.3 VERBAL CONTENT AND STYLE IN READING ALOUD 37 3.3 TEXT-TO-SPEECH SYSTEM ORGANISATION 37 3.3.1 THE COMMON-FORM MODEL 38 3.3.2 OTHER MODELS 39 3.3.3 COMPARISON 40 3.4 SYSTEMS 41 3.4.1 A SIMPLE TEXT-TO-SPEECH SYSTEM 41 3.4.2 CONCEPT TO SPEECH ' 42 3.4.3 CANNED SPEECH AND LIMITED-DOMAIN SYNTHESIS 43 3.5 KEY PROBLEMS IN TEXT-TO-SPEECH 44 3.5.1 TEXT CLASSIFICATION WITH RESPECT TO SEMIOTIC SYSTEMS 44 3.5.2 DECODING NATURAL-LANGUAGE TEXT 46 3.5.3 NATURALNESS 47 3.5.4 INTELLIGIBILITY: ENCODING THE MESSAGE IN SIGNAL 48 3.5.5 AUXILIARY GENERATION FOR PROSODY 49 3.5.6 ADAPTING THE SYSTEM TO THE SITUATION 50 3.6 SUMMARY 50 TEXT SEGMENTATION AND ORGANISATION 52 52 53 53 55 59 60 61 61 62 63 63 4.1 OVERVIEW OF THE PROBLEM 4.2 WORDS 4.2 4.2 4.2 4.2 4.2 4.2 4.2 4.2 .1 .2 .3 .4 .5 .6 .7 .8 4.3 TEXT SE AND SENTENCES WHAT IS A WORD? DEFINING WORDS IN TEXT-TO-SPEECH SCOPE AND MORPHOLOGY CONTRACTIONS AND CLITICS SLANG FORMS HYPHENATED FORMS WHAT IS A SENTENCE? THE LEXICON GMENTATION CONTENTS 4.3.1 TOKENISATION 64 4.3.2 TOKENISATION AND PUNCTUATION 65 4.3.3 TOKENISATION ALGORITHMS 66 4.3.4 SENTENCE SPLITTING 67 4.4 PROCESSING DOCUMENTS 68 4.4.1 MARKUP LANGUAGES 68 4.4.2 INTERPRETING CHARACTERS 70 4.5 TEXT-TO-SPEECH ARCHITECTURES 71 4.6 DISCUSSION 75 4.6.1 FURTHER READING 75 4.6.2 SUMMARY 76 TEXT DECODING: FINDING THE WORDS FROM THE TEXT 78 5.1 OVERVIEW OF TEXT DECODING 78 5.2 TEXT-CLASSIFICATION ALGORITHMS 79 5.2.1 FEATURES AND ALGORITHMS 79 5.2.2 TAGGING AND WORD-SENSE DISAMBIGUATION 82 5.2.3 AD-HOC APPROACHES 83 5.2.4 DETERMINISTIC RULE APPROACHES 83 5.2.5 DECISION LISTS , 85 5.2.6 NAIVE BAYES CLASSIFIER ' 86 5.2.7 DECISION TREES 87 5.2.8 PART-OF-SPEECH TAGGING 88 5.3 NON-NATURAL-LANGUAGE TEXT 92 5.3.1 SEMIOTIC CLASSIFICATION 92 5.3.2 SEMIOTIC DECODING 95 5.3.3 VERBALISATION 95 5.4 NATURAL-LANGUAGE TEXT 97 5.4.1 ACRONYMS AND LETTER SEQUENCES 99 5.4.2 HOMOGRAPH DISAMBIGUATION 99 5.4.3 NON-HOMOGRAPHS 101 5.5 NATURAL-LANGUAGE PARSING 102 5.5.1 CONTEXT-FREE GRAMMARS 102 5.5.2 STATISTICAL PARSING 105 5.6 DISCUSSION 105 5.6.1 FURTHER READING 108 5.6.2 SUMMARY 109 PROSODY PREDICTION FROM TEXT 111 6.1 PROSODIC FORM 111 6.2 PHRASING 112 6.2.1 PHRASING PHENOMENA 112 6.2.2 MODELS OF PHRASING 113 XII CONTENTS 6.3 PROMINENCE 115 6.3.1 SYNTACTIC PROMINENCE PATTERNS 116 6.3.2 DISCOURSE PROMINENCE PATTERNS 118 6.3.3 PROMINENCE SYSTEMS, DATA AND LABELLING 119 6.4 INTONATION AND TUNE 121 6.5 PROSODIC MEANING AND FUNCTION 122 6.5.1 AFFECTIVE PROSODY 123 6.5.2 SUPRASEGMENTALITY 124 6.5.3 AUGMENTATIVE PROSODY 125 6.5.4 SYMBOLIC COMMUNICATION AND PROSODIC STYLE 126 6.6 DETERMINING PROSODY FROM THE TEXT 127 6.6.1 PROSODY AND HUMAN READING 127 6.6.2 CONTROLLING THE-DEGREE OF AUGMENTATIVE PROSODY 128 6.6.3 PROSODY AND SYNTHESIS TECHNIQUES 128 6.7 PHRASING PREDICTION 129 6.7.1 EXPERIMENTAL FORMULATION 129 6.7.2 DETERMINISTIC APPROACHES 130 6.7.3 CLASSIFIER APPROACHES 132 6.7.4 HMM APPROACHES 133 6.7.5 HYBRID APPROACHES 135 6.8 PROMINENCE PREDICTION . 136 6.8.1 COMPOUND-NOUN PHRASES 136 6.8.2 FUNCTION-WORD PROMINENCE 138 6.8.3 DATA-DRIVEN APPROACHES 138 6.9 INTONATIONAL-TUNE PREDICTION 139 6.10 DISCUSSION 139 6.10.1 LABELLING SCHEMES AND LABELLING ACCURACY 139 6.10.2 LINGUISTIC THEORIES AND PROSODY 141 6.10.3 SYNTHESISING SUPRASEGMENTAL AND TRUE PROSODY 142 6.10.4 PROSODY IN REAL DIALOGUES 143 6.10.5 CONCLUSION 144 6.10.6 SUMMARY 144 PHONETICS AND PHONOLOGY 146 7.1 ARTICULATORY PHONETICS AND SPEECH PRODUCTION 146 7.1.1 THE VOCAL ORGANS 147 7.1.2 SOUND SOURCES 147 7.1.3 SOUND OUTPUT 150 7.1.4 THE VOCAL-TRACT FILTER 150 7.1.5 VOWELS 151 7.1.6 CONSONANTS 153 7.1.7 EXAMINING SPEECH PRODUCTION 155 7.2 ACOUSTICS, PHONETICS AND SPEECH PERCEPTION 156 CONTENTS XIII 7.3 7.4 7.5 7.2.1 ACOUSTIC REPRESENTATIONS 7.2.2 ACOUSTIC CHARACTERISTICS THE COMMUNICATIVE USE OF SPEECH 7.3.1 COMMUNICATING DISCRETE INFORMATION WITH A CONTINUOUS CHANNEL 7.3.2 PHONEMES, PHONES AND ALLOPHONES 7.3.3 ALLOPHONIC VARIATION AND PHONETIC CONTEXT 7.3.4 COARTICULATION, TARGETS AND TRANSIENTS 7.3.5 THE CONTINUOUS NATURE OF SPEECH 7.3.6 TRANSCRIPTION 7.3.7 THE DISTINCTIVENESS OF SPEECH IN COMMUNICATION PHONOLOGY: THE LINGUISTIC ORGANISATION OF SPEECH 7.4.1 PHONOTACTICS 7.4.2 WORD FORMATION 7.4.3 DISTINCTIVE FEATURES AND PHONOLOGICAL THEORIES 7.4.4 SYLLABLES 7.4.5 LEXICAL STRESS DISCUSSION 7.5.1 FURTHER READING 7.5.2 SUMMARY 8 PRONUNCIATION 8.1 8.2 8.3 8.4 PRONUNCIATION REPRESENTATIONS 8.1.1 WHY BOTHER? 8.1.2 PHONEMIC AND PHONETIC INPUT 8.1.3 DIFFICULTIES IN DERIVING PHONETIC INPUT 8.1.4 A STRUCTURED APPROACH TO PRONUNCIATION 8.1.5 ABSTRACT PHONOLOGICAL REPRESENTATIONS FORMULATING A PHONOLOGICAL REPRESENTATION SYSTEM 8.2.1 SIMPLE CONSONANTS AND VOWELS 8.2.2 DIFFICULT CONSONANTS 8.2.3 DIPHTHONGS AND AFFRICATES 8.2.4 APPROXIMANT-VOWEL COMBINATIONS 8.2.5 DEFINING THE FULL INVENTORY 8.2.6 PHONEME NAMES 8.2.7 SYLLABIC ISSUES THE LEXICON 8.3.1 LEXICON AND RULES 8.3.2 LEXICON FORMATS 8.3.3 THE OFFLINE LEXICON 8.3.4 THE SYSTEM LEXICON 8.3.5 LEXICON QUALITY 8.3.6 DETERMINING THE PRONUNCIATIONS OF UNKNOWN WORDS GRANHEME-TO-RJHONEME CONVERSION 156 159 160 161 162 166 168 169 170 171 172 172 179 181 184 186 189 189 190 192 192 192 193 194 195 196 197 197 199 201 201 203 204 206 207 208 210 213 214 215 216 218 XIV CONTENTS 8.4.1 8.4.2 8.4.3 8.4.4 8.4.5 8.4.6 RULE-BASED TECHNIQUES GRAPHEME-TO-PHONEME ALIGNMENT NEURAL NETWORKS PRONUNCIATION BY ANALOGY OTHER DATA-DRIVEN TECHNIQUES STATISTICAL TECHNIQUES 8.5 FURTHER ISSUES 8.5.1 8.5.2 8.5.3 MORPHOLOGY LANGUAGE ORIGIN AND NAMES POST-LEXICAL PROCESSING 8.6 SUMMARY 218 219 219 220 221 221 222 222 223 223 224 SYNTHESIS OF PROSODY 225 9.1 INTONATION OVERVIEW 225 9.1.1 F0 AND PITCH 226 9.1.2 INTONATIONAL FORM 226 9.1.3 MODELS OF F0 CONTOURS 227 9.1.4 MICRO-PROSODY 229 9.2 INTONATIONAL BEHAVIOUR 229 9.2.1 INTONATIONAL TUNE 229 9.2.2 DOWNDRIFT ' 230 9.2.3 PITCH RANGE 233 9.2.4 PITCH ACCENTS AND BOUNDARY TONES 234 9.3 INTONATION THEORIES AND MODELS 236 9.3.1 TRADITIONAL MODELS AND THE BRITISH SCHOOL 236 9.3.2 THE DUTCH SCHOOL 237 9.3.3 AUTOSEGMENTAL-METRICAL AND TOBI MODELS 237 9.3.4 THE INTSINT MODEL 239 9.3.5 THE FUJISAKI MODEL AND SUPERIMPOSITIONAL MODELS 239 9.3.6 THE TILT MODEL 242 9.3.7 COMPARISON 244 9.4 INTONATION SYNTHESIS WITH AM MODELS 245 9.4.1 PREDICTION OF AM LABELS FROM TEXT 246 9.4.2 DETERMINISTIC SYNTHESIS METHODS 246 9.4.3 DATA-DRIVEN SYNTHESIS METHODS 247 9.4.4 ANALYSIS WITH AUTOSEGMENTAL MODELS 248 9.5 INTONATION SYNTHESIS WITH DETERMINISTIC ACOUSTIC MODELS 248 9.5.1 SYNTHESIS WITH SUPERIMPOSITIONAL MODELS 249 9.5.2 SYNTHESIS WITH THE TILT MODEL 249 9.5.3 ANALYSIS WITH FUJISAKI AND TILT MODELS 250 9.6 DATA-DRIVEN INTONATION MODELS 250 9.6.1 UNIT-SELECTION-STYLE APPROACHES 251 9.6.2 DYNAMIC-SYSTEM MODELS 252 CONTENTS XV 9.7 9.8 9.6.3 9.6.4 HIDDEN MARKOV MODELS FUNCTIONAL MODELS TIMING 9.7.1 9.7.2 9.7.3 9.7.4 9.7.5 9.7.6 FORMULATION OF THE TIMING PROBLEM THE NATURE OF TIMING KLATT RULES THE SUMS-OF-PRODUCTS MODEL THE CAMPBELL MODEL OTHER REGRESSION TECHNIQUES DISCUSSION 9.8.1 9.8.2 SIGNALS AND FURTHER READING SUMMARY FILTERS 253 254 254 255 255 256 257 258 259 259 260 260 10 SIGNALS AND FILTERS 262 10.1 ANALOGUE SIGNALS 262 10.1.1 SIMPLE PERIODIC SIGNALS: SINUSOIDS 263 10.1.2 GENERAL PERIODIC SIGNALS 265 10.1.3 SINUSOIDS AS COMPLEX EXPONENTIALS 266 10.1.4 FOURIER ANALYSIS 269 10.1.5 THE FREQUENCY DOMAIN 270 10.1.6 THE FOURIER TRANSFORM ' 275 10.2 DIGITAL SIGNALS 278 10.2.1 DIGITAL WAVEFORMS 279 10.2.2 DIGITAL REPRESENTATIONS 280 10.2.3 THE DISCRETE-TIME FOURIER TRANSFORM 280 10.2.4 THE DISCRETE FOURIER TRANSFORM 281 10.2.5 THE Z-TRANSFORM 282 10.2.6 THE FREQUENCY DOMAIN FOR DIGITAL SIGNALS 283 10.3 PROPERTIES OF TRANSFORMS 284 10.3.1 LINEARITY 284 10.3.2 TIME AND FREQUENCY DUALITY 284 10.3.3 SCALING 285 10.3.4 IMPULSE PROPERTIES 285 10.3.5 TIME DELAY 286 10.3.6 FREQUENCY SHIFT 286 10.3.7 CONVOLUTION 287 10.3.8 ANALYTICAL AND NUMERICAL ANALYSIS 287 10.3.9 STOCHASTIC SIGNALS 288 10.4 DIGITAL FILTERS 288 10.4.1 DIFFERENCE EQUATIONS 289 10.4.2 THE IMPULSE RESPONSE 289 10.4.3 THE FILTER CONVOLUTION SUM 292 10.4.4 THE FILTER TRANSFER FUNCTION 293 XVI CONTENTS 10.4.5 THE TRANSFER FUNCTION AND THE IMPULSE RESPONSE 293 10.5 DIGITAL FILTER ANALYSIS AND DESIGN 294 10.5.1 POLYNOMIAL ANALYSIS: POLES AND ZEROS 294 10.5.2 FREQUENCY INTERPRETATION OF THE Z-DOMAIN TRANSFER FUNCTION 297 10.5.3 FILTER CHARACTERISTICS 298 10.5.4 PUTTING IT ALL TOGETHER 304 10.6 SUMMARY 305 11 ACOUSTIC MODELS OF SPEECH PRODUCTION 309 11.1 THE ACOUSTIC THEORY OF SPEECH PRODUCTION 309 11.1.1 COMPONENTS IN THE MODEL 309 11.2 THE PHYSICS OF SOUND 311 11.2.1 RESONANT SYSTEMS 311 11.2.2 TRAVELLING WAVES 313 11.2.3 ACOUSTIC WAVES 315 11.2.4 ACOUSTIC REFLECTION 317 11.3 THE VOWEL-TUBE MODEL 318 11.3.1 DISCRETE TIME AND DISTANCE 319 11.3.2 JUNCTION OF TWO TUBES 320 11.3.3 SPECIAL CASES OF JUNCTIONS 322 11.3.4 THE TWO-TUBE VOCAL-TRACT MODEL ' 323 11.3.5 THE SINGLE-TUBE MODEL 325 11.3.6 THE MULTI-TUBE VOCAL-TRACT MODEL 327 11.3.7 THE ALL-POLE RESONATOR MODEL 329 11.4 SOURCE AND RADIATION MODELS 330 11.4.1 RADIATION 330 11.4.2 THE GLOTTAL SOURCE 330 11.5 MODEL REFINEMENTS 333 11.5.1 MODELLING THE NASAL CAVITY 333 11.5.2 SOURCE POSITIONS IN THE ORAL CAVITY 334 11.5.3 MODELS WITH VOCAL-TRACT LOSSES 335 11.5.4 SOURCE AND RADIATION EFFECTS 336 11.6 DISCUSSION 336 11.6.1 FURTHER READING 339 11.6.2 SUMMARY 339 12 ANALYSIS OF SPEECH SIGNALS 341 12.1 SHORT-TERM SPEECH ANALYSIS 341 12.1.1 WINDOWING 342 12.1.2 SHORT-TERM SPECTRAL REPRESENTATIONS 343 12.1.3 FRAME LENGTHS AND SHIFTS 345 12.1.4 THE SPECTROGRAM 348 12.1.5 AUDITORY SCALES 351 CONTENTS XVII 12.2 FILTER-BANK ANALYSIS 352 12.3 THE CEPSTRUM 353 12.3.1 CEPSTRUM DEFINITION 353 12.3.2 TREATING THE MAGNITUDE SPECTRUM AS A SIGNAL 353 12.3.3 CEPSTRAL ANALYSIS AS DECONVOLUTION 355 12.3.4 CEPSTRAL ANALYSIS DISCUSSION 356 12.4 LINEAR-PREDICTION ANALYSIS 357 12.4.1 FINDING THE COEFFICIENTS: THE COVARIANCE METHOD 358 12.4.2 THE AUTOCORRELATION METHOD 360 12.4.3 LEVINSON-DURBIN RECURSION 361 12.5 SPECTRAL-ENVELOPE AND VOCAL-TRACT REPRESENTATIONS 362 12.5.1 LINEAR-PREDICTION SPECTRA 362 12.5.2 TRANSFER-FUNCTION POLES 364 12.5.3 REFLECTION COEFFICIENTS 366 12.5.4 LOG AREA RATIOS 367 12.5.5 LINE-SPECTRUM FREQUENCIES 367 12.5.6 LINEAR-PREDICTION CEPSTRA 369 12.5.7 MEL-SCALED CEPSTRA 370 12.5.8 PERCEPTUAL LINEAR PREDICTION 370 12.5.9 FORMANT TRACKING 370 12.6 SOURCE REPRESENTATIONS , 372 12.6.1 RESIDUAL SIGNALS 372 12.6.2 CLOSED-PHASE ANALYSIS 374 12.6.3 OPEN-PHASE ANALYSIS 377 12.6.4 IMPULSE/NOISE MODELS 378 12.6.5 PARAMETERISATION OF GLOTTAL-FLOW SIGNALS 379 12.7 PITCH AND EPOCH DETECTION 379 12.7.1 PITCH DETECTION 379 12.7.2 EPOCH DETECTION: FINDING THE INSTANT OF GLOTTAL CLOSURE 381 12.8 DISCUSSION 384 12.8.1 FURTHER READING 385 12.8.2 SUMMARY 386 13 SYNTHESIS TECHNIQUES BASED ON VOCAL-TRACT MODELS 387 13.1 SYNTHESIS SPECIFICATION: THE INPUT TO THE SYNTHESISER 387 13.2 FORMANT SYNTHESIS 388 13.2.1 SOUND SOURCES 389 13.2.2 SYNTHESISING A SINGLE FORMANT 390 13.2.3 RESONATORS IN SERIES AND PARALLEL 391 13.2.4 SYNTHESISING CONSONANTS 392 13.2.5 A COMPLETE SYNTHESISER 394 13.2.6 THE PHONETIC INPUT TO THE SYNTHESISER 394 13.2.7 FORMANT-SYNTHESIS QUALITY 397 XVIII CONTENTS 13.3 CLASSICAL LINEAR-PREDICTION SYNTHESIS 13.3.1 COMPARISON WITH FORMANT SYNTHESIS 13.3.2 THE IMPULSE/NOISE SOURCE MODEL 13.3.3 LINEAR-PREDICTION DIPHONE-CONCATENATIVE SYNTHESIS 13.3.4 A COMPLETE SYNTHESISER 13.3.5 PROBLEMS WITH THE SOURCE 13.4 ARTICULATORY SYNTHESIS 13.5 DISCUSSION 13.5.1 FURTHER READING 13.5.2 SUMMARY 14 SYNTHESIS BY CONCATENATION AND SIGNAL-PROCESSING MODIFICATION 14.1 SPEECH UNITS IN SECOND-GENERATION SYSTEMS 14.1.1 CREATING A DIPHONE INVENTORY 14.1.2 OBTAINING DIPHONES FROM SPEECH 14.2 PITCH-SYNCHRONOUS OVERLAP AND ADD (PSOLA) 14.2.1 TIME-DOMAIN PSOLA 14.2.2 EPOCH MANIPULATION 14.2.3 HOW DOES PSOLA WORK? 14.3 RESIDUAL-EXCITED LINEAR PREDICTION (RELP) 14.3.1 RESIDUAL MANIPULATION 14.3.2 LINEAR-PREDICTION PSOLA 14.4 SINUSOIDAL MODELS 14.4.1 PURE SINUSOIDAL MODELS 14.4.2 HARMONIC/NOISE MODELS 14.5 MBROLA' 14.6 SYNTHESIS FROM CEPSTRAL COEFFICIENTS 14.7 CONCATENATION ISSUES 14.8 DISCUSSION 14.8.1 FURTHER READING 14.8.2 SUMMARY 15 HIDDEN-MARKOV-MODEL SYNTHESIS 15.1 THE HMM FORMALISM 15.1.1 OBSERVATION PROBABILITIES 15.1.2 DELTA COEFFICIENTS 15.1.3 ACOUSTIC REPRESENTATIONS AND COVARIANCE 15. 15. 15. 15. 15. .4 STATES AND TRANSITIONS .5 RECOGNISING WITH HMMS .6 LANGUAGE MODELS .7 THE VITERBI ALGORITHM .8 TRAINING HMMS 15.1.9 CONTEXT-SENSITIVE MODELLING 399 399 400 401 403 404 405 407 409 410 412 413 414 414 415 416 417 421 421 423 423 424 425 426 429 429 431 433 433 433 435 435 436 438 439 440 440 443 444 447 451 CONTENTS XIX 15.1.10 ARE HMMS A GOOD MODEL OF SPEECH? 454 15.2 SYNTHESIS FROM HIDDEN MARKOV MODELS 456 15.2.1 FINDING THE LIKELIEST OBSERVATIONS GIVEN THE STATE SEQUENCE 457 15.2.2 FINDING THE LIKELIEST OBSERVATIONS AND STATE SEQUENCE 459 15.2.3 ACOUSTIC REPRESENTATIONS 459 15.2.4 CONTEXT-SENSITIVE SYNTHESIS MODELS 463 15.2.5 DURATION MODELLING 464 15.2.6 SIGNAL PROCESSING IN HMM SYNTHESIS 464 15.2.7 HMM SYNTHESIS SYSTEMS 465 15.3 LABELLING DATABASES WITH HMMS 467 15.3.1 DETERMINING THE WORD SEQUENCE 467 15.3.2 DETERMINING THE PHONE SEQUENCE 468 15.3.3 DETERMINING THE PHONE BOUNDARIES 468 15.3.4 MEASURING THE QUALITY OF THE ALIGNMENTS 470 15.4 OTHER DATA-DRIVEN SYNTHESIS TECHNIQUES 471 15.5 DISCUSSION 471 15.5.1 FURTHER READING 471 15.5.2 SUMMARY 472 16 UNIT-SELECTION SYNTHESIS 474 16.1 FROM CONCATENATIVE SYNTHESIS TO UNIT SELECTION 474 16.1.1 EXTENDING CONCATENATIVE SYNTHESIS 475 16.1.2 THE HUNT AND BLACK ALGORITHM 477 16.2 FEATURES, 479 16.2.1 BASE TYPES 479 16.2.2 LINGUISTIC AND ACOUSTIC FEATURES 480 16.2.3 CHOICE OF FEATURES 481 16.2.4 TYPES OF FEATURES 482 16.3 THE INDEPENDENT-FEATURE TARGET-FUNCTION FORMULATION 484 16.3.1 THE PURPOSE OF THE TARGET FUNCTION 484 16.3.2 DEFINING A PERCEPTUAL SPACE 485 16.3.3 PERCEPTUAL SPACES DEFINED BY INDEPENDENT FEATURES 486 16.3.4 SETTING THE TARGET WEIGHTS USING ACOUSTIC DISTANCES 488 16.3.5 LIMITATIONS OF THE INDEPENDENT-FEATURE FORMULATION 491 16.4 THE ACOUSTIC-SPACE TARGET-FUNCTION FORMULATION 493 16.4.1 DECISION-TREE CLUSTERING 494 16.4.2 GENERAL PARTIAL-SYNTHESIS FUNCTIONS 496 16.5 JOIN FUNCTIONS 497 16.5.1 BASIC ISSUES IN JOINING UNITS 497 16.5.2 PHONE-CLASS JOIN COSTS 498 16.5.3 ACOUSTIC-DISTANCE JOIN COSTS 499 16.5.4 COMBINING CATEGORICAL AND AND ACOUSTIC JOIN COSTS 500 16.5.5 PROBABILISTIC AND SEQUENCE JOIN COSTS 501 XX CONTENTS 16.5.6 JOIN CLASSIFIERS 502 16.6 SEARCHING 504 16.6.1 BASE TYPES AND SEARCHING 505 16.6.2 PRUNING 508 16.6.3 PRE-SELECTION 508 16.6.4 BEAM PRUNING 509 16.6.5 MULTI-PASS SEARCHING 509 16.7 DISCUSSION 510 16.7.1 UNIT SELECTION AND SIGNAL PROCESSING 511 16.7.2 FEATURES, COSTS AND PERCEPTION 511 16.7.3 EXAMPLE UNIT-SELECTION SYSTEMS 512 16.7.4 FURTHER READING 514 16.7.5 SUMMARY 515 17 FURTHER ISSUES 517 17.1 DATABASES 517 17.1.1 UNIT-SELECTION DATABASES 517 17.1.2 TEXT MATERIALS 518 17.1.3 PROSODY DATABASES 518 17.1.4 LABELLING 519 17.1.5 WHAT EXACTLY IS HAND LABELLING? ' 519 17.1.6 AUTOMATIC LABELLING 521 17.1.7 AVOIDING EXPLICIT LABELS 521 17.2 EVALUATION 522 17.2.1 SYSTEM TESTING: INTELLIGIBILITY AND NATURALNESS 523 17.2.2 WORD-RECOGNITION TESTS 523 17.2.3 NATURALNESS TESTS 524 17.2.4 TEST DATA 525 17.2.5 UNIT OR COMPONENT TESTING 525 17.2.6 COMPETITIVE EVALUATIONS 526 17.3 AUDIO-VISUAL SPEECH SYNTHESIS 527 17.3.1 SPEECH CONTROL 528 17.4 SYNTHESIS OF EMOTIONAL AND EXPRESSIVE SPEECH 529 17.4.1 DESCRIBING EMOTION 529 17.4.2 SYNTHESISING EMOTION WITH PROSODY CONTROL 529 17.4.3 SYNTHESISING EMOTION WITH VOICE TRANSFORMATION 530 17.4.4 UNIT SELECTION AND HMM TECHNIQUES 531 17.5 SUMMARY 531 18 CONCLUSION 533 18.1 SPEECH TECHNOLOGY AND LINGUISTICS 533 18.2 FUTURE DIRECTIONS 536 18.3 CONCLUSION 539 CONTENTS XXI APPENDIX A PROBABILITY 540 A. 1 DISCRETE PROBABILITIES 540 A. 1.1 DISCRETE RANDOM VARIABLES 540 A. 1.2 PROBABILITY MASS FUNCTIONS 541 A. 1.3 EXPECTED VALUES 541 A. 1.4 MOMENTS OF A PMF 542 A.2 PAIRS OF DISCRETE RANDOM VARIABLES 542 A.2.1 MARGINAL DISTRIBUTIONS 543 A.2.2 INDEPENDENCE 543 A.2.3 EXPECTED VALUES 543 A.2.4 MOMENTS OF A JOINT DISTRIBUTION 544 A.2.5 HIGHER-ORDER MOMENTS AND COVARIANCE 544 A.2.6 CORRELATION 544 A.2.7 CONDITIONAL PROBABILITY 545 A.2.8 BAYES'RULE 545 A.2.9 SUM OF RANDOM VARIABLES 545 A.2.10 THE CHAIN RULE 546 A.2.11 ENTROPY 546 A.3 CONTINUOUS RANDOM VARIABLES 547 A.3.1 CONTINUOUS RANDOM VARIABLES 548 A.3.2 EXPECTED VALUES ' 548 A.3.3 THE GAUSSIAN DISTRIBUTION 549 A.3.4 THE UNIFORM DISTRIBUTION 549 - A.3.5 CUMULATIVE DENSITY FUNCTIONS 549 A.4 PAIRS OF CONTINUOUS RANDOM VARIABLES 550 A.4.1 INDEPENDENT VERSUS UNCORRELATED 551 A.4.2 THE SUM OF TWO RANDOM VARIABLES 551 A.4.3 ENTROPY 552 A.4.4 KULLBACK-LEIBLER DISTANCE 552 APPENDIX B PHONE DEFINITIONS 553 REFERENCES 556 INDEX 583
any_adam_object	1
any_adam_object_boolean	1
author	Taylor, Paul
author_facet	Taylor, Paul
author_role	aut
author_sort	Taylor, Paul
author_variant	p t pt
building	Verbundindex
bvnumber	BV035039831
callnumber-first	T - Technology
callnumber-label	TK7882
callnumber-raw	TK7882.S65
callnumber-search	TK7882.S65
callnumber-sort	TK 47882 S65
callnumber-subject	TK - Electrical and Nuclear Engineering
classification_rvk	ES 950
ctrlnum	(OCoLC)221147648 (DE-599)BVBBV035039831
dewey-full	006.54
dewey-hundreds	000 - Computer science, information, general works
dewey-ones	006 - Special computer methods
dewey-raw	006.54
dewey-search	006.54
dewey-sort	16.54
dewey-tens	000 - Computer science, information, general works
discipline	Informatik Sprachwissenschaft Literaturwissenschaft
discipline_str_mv	Informatik Sprachwissenschaft Literaturwissenschaft
edition	1. publ.
format	Book
fullrecord	<?xml version="1.0" encoding="UTF-8"?><collection xmlns="http://www.loc.gov/MARC21/slim"><record><leader>01261nam a2200361 c 4500</leader><controlfield tag="001">BV035039831</controlfield><controlfield tag="003">DE-604</controlfield><controlfield tag="005">20090212 </controlfield><controlfield tag="007">t</controlfield><controlfield tag="008">080905s2009 d\|\|\| \|\|\|\| 00\|\|\| ger d</controlfield><datafield tag="020" ind1=" " ind2=" "><subfield code="a">9780521899277</subfield><subfield code="9">978-0-521-89927-7</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(OCoLC)221147648</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(DE-599)BVBBV035039831</subfield></datafield><datafield tag="040" ind1=" " ind2=" "><subfield code="a">DE-604</subfield><subfield code="b">ger</subfield><subfield code="e">rakwb</subfield></datafield><datafield tag="041" ind1="0" ind2=" "><subfield code="a">ger</subfield></datafield><datafield tag="049" ind1=" " ind2=" "><subfield code="a">DE-19</subfield><subfield code="a">DE-11</subfield></datafield><datafield tag="050" ind1=" " ind2="0"><subfield code="a">TK7882.S65</subfield></datafield><datafield tag="082" ind1="0" ind2=" "><subfield code="a">006.54</subfield><subfield code="2">22</subfield></datafield><datafield tag="084" ind1=" " ind2=" "><subfield code="a">ES 950</subfield><subfield code="0">(DE-625)27936:</subfield><subfield code="2">rvk</subfield></datafield><datafield tag="100" ind1="1" ind2=" "><subfield code="a">Taylor, Paul</subfield><subfield code="e">Verfasser</subfield><subfield code="4">aut</subfield></datafield><datafield tag="245" ind1="1" ind2="0"><subfield code="a">Text-to-speech synthesis</subfield><subfield code="c">Paul Taylor</subfield></datafield><datafield tag="250" ind1=" " ind2=" "><subfield code="a">1. publ.</subfield></datafield><datafield tag="264" ind1=" " ind2="1"><subfield code="a">Cambridge [u.a.]</subfield><subfield code="b">Cambridge Univ. Pr.</subfield><subfield code="c">2009</subfield></datafield><datafield tag="300" ind1=" " ind2=" "><subfield code="a">XXVIII, 597 S.</subfield><subfield code="b">graph. Darst.</subfield></datafield><datafield tag="336" ind1=" " ind2=" "><subfield code="b">txt</subfield><subfield code="2">rdacontent</subfield></datafield><datafield tag="337" ind1=" " ind2=" "><subfield code="b">n</subfield><subfield code="2">rdamedia</subfield></datafield><datafield tag="338" ind1=" " ind2=" "><subfield code="b">nc</subfield><subfield code="2">rdacarrier</subfield></datafield><datafield tag="500" ind1=" " ind2=" "><subfield code="a">Literaturverz. S. [556] - 582</subfield></datafield><datafield tag="650" ind1=" " ind2="7"><subfield code="a">Sprachcodierung</subfield><subfield code="2">swd</subfield></datafield><datafield tag="650" ind1=" " ind2="7"><subfield code="a">Synthèse automatique de la parole</subfield><subfield code="2">ram</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">Speech synthesis</subfield></datafield><datafield tag="856" ind1="4" ind2="2"><subfield code="m">HEBIS Datenaustausch</subfield><subfield code="q">application/pdf</subfield><subfield code="u">http://bvbr.bib-bvb.de:8991/F?func=service&doc_library=BVB01&local_base=BVB01&doc_number=016708674&sequence=000001&line_number=0001&func_code=DB_RECORDS&service_type=MEDIA</subfield><subfield code="3">Inhaltsverzeichnis</subfield></datafield><datafield tag="999" ind1=" " ind2=" "><subfield code="a">oai:aleph.bib-bvb.de:BVB01-016708674</subfield></datafield></record></collection>
id	DE-604.BV035039831
illustrated	Illustrated
index_date	2024-07-02T21:52:49Z
indexdate	2024-07-09T21:20:49Z
institution	BVB
isbn	9780521899277
language	German
oai_aleph_id	oai:aleph.bib-bvb.de:BVB01-016708674
oclc_num	221147648
open_access_boolean
owner	DE-19 DE-BY-UBM DE-11
owner_facet	DE-19 DE-BY-UBM DE-11
physical	XXVIII, 597 S. graph. Darst.
publishDate	2009
publishDateSearch	2009
publishDateSort	2009
publisher	Cambridge Univ. Pr.
record_format	marc
spelling	Taylor, Paul Verfasser aut Text-to-speech synthesis Paul Taylor 1. publ. Cambridge [u.a.] Cambridge Univ. Pr. 2009 XXVIII, 597 S. graph. Darst. txt rdacontent n rdamedia nc rdacarrier Literaturverz. S. [556] - 582 Sprachcodierung swd Synthèse automatique de la parole ram Speech synthesis HEBIS Datenaustausch application/pdf http://bvbr.bib-bvb.de:8991/F?func=service&doc_library=BVB01&local_base=BVB01&doc_number=016708674&sequence=000001&line_number=0001&func_code=DB_RECORDS&service_type=MEDIA Inhaltsverzeichnis
spellingShingle	Taylor, Paul Text-to-speech synthesis Sprachcodierung swd Synthèse automatique de la parole ram Speech synthesis
title	Text-to-speech synthesis
title_auth	Text-to-speech synthesis
title_exact_search	Text-to-speech synthesis
title_exact_search_txtP	Text-to-speech synthesis
title_full	Text-to-speech synthesis Paul Taylor
title_fullStr	Text-to-speech synthesis Paul Taylor
title_full_unstemmed	Text-to-speech synthesis Paul Taylor
title_short	Text-to-speech synthesis
title_sort	text to speech synthesis
topic	Sprachcodierung swd Synthèse automatique de la parole ram Speech synthesis
topic_facet	Sprachcodierung Synthèse automatique de la parole Speech synthesis
url	http://bvbr.bib-bvb.de:8991/F?func=service&doc_library=BVB01&local_base=BVB01&doc_number=016708674&sequence=000001&line_number=0001&func_code=DB_RECORDS&service_type=MEDIA
work_keys_str_mv	AT taylorpaul texttospeechsynthesis

Verfügbarkeit

Es ist kein Print-Exemplar vorhanden.

Fernleihe Bestellen Achtung: Nicht im THWS-Bestand! Inhaltsverzeichnis

MARC

Datensatz im Suchindex

Es ist kein Print-Exemplar vorhanden.

Ähnliche Einträge