Unit selection and waveform concatenation strategies in Cantonese text-to-speech.

Oey Sai Lok.Thesis (M.Phil.)--Chinese University of Hong Kong, 2005.Includes bibliographical references.Abstracts in English and Chinese.Chapter 1. --- Introduction --- p.1Chapter 1.1 --- An overview of Text-to-Speech technology --- p.2Chapter 1.1.1 --- Text processing --- p.2Chapter 1.1.2 --- Acoustic synthesis --- p.3Chapter 1.1.3 --- Prosody modification --- p.4Chapter 1.2 --- Trends in Text-to-Speech technologies --- p.5Chapter 1.3 --- Objectives of this thesis --- p.7Chapter 1.4 --- Outline of the thesis --- p.9References --- p.11Chapter 2. --- Cantonese Speech --- p.13Chapter 2.1 --- The Cantonese dialect --- p.13Chapter 2.2 --- Phonology of Cantonese --- p.14Chapter 2.2.1 --- Initials --- p.15Chapter 2.2.2 --- Finals --- p.16Chapter 2.2.3 --- Tones --- p.18Chapter 2.3 --- Acoustic-phonetic properties of Cantonese syllables --- p.19References --- p.24Chapter 3. --- Cantonese Text-to-Speech --- p.25Chapter 3.1 --- General overview --- p.25Chapter 3.1.1 --- Text processing --- p.25Chapter 3.1.2 --- Corpus based acoustic synthesis --- p.26Chapter 3.1.3 --- Prosodic control --- p.27Chapter 3.2 --- Syllable based Cantonese Text-to-Speech system --- p.28Chapter 3.3 --- Sub-syllable based Cantonese Text-to-Speech system --- p.29Chapter 3.3.1 --- Definition of sub-syllable units --- p.29Chapter 3.3.2 --- Acoustic inventory --- p.31Chapter 3.3.3 --- Determination of the concatenation points --- p.33Chapter 3.4 --- Problems --- p.34References --- p.36Chapter 4. --- Waveform Concatenation for Sub-syllable Units --- p.37Chapter 4.1 --- Previous work in concatenation methods --- p.37Chapter 4.1.1 --- Determination of concatenation point --- p.38Chapter 4.1.2 --- Waveform concatenation --- p.38Chapter 4.2 --- Problems and difficulties in concatenating sub-syllable units --- p.39Chapter 4.2.1 --- Mismatch of acoustic properties --- p.40Chapter 4.2.2 --- "Allophone problem of Initials /z/, Id and /s/" --- p.42Chapter 4.3 --- General procedures in concatenation strategies --- p.44Chapter 4.3.1 --- Concatenation of unvoiced segments --- p.45Chapter 4.3.2 --- Concatenation of voiced segments --- p.45Chapter 4.3.3 --- Measurement of spectral distance --- p.48Chapter 4.4 --- Detailed procedures in concatenation points determination --- p.50Chapter 4.4.1 --- Unvoiced segments --- p.50Chapter 4.4.2 --- Voiced segments --- p.53Chapter 4.5 --- Selected examples in concatenation strategies --- p.58Chapter 4.5.1 --- Concatenation at Initial segments --- p.58Chapter 4.5.1.1 --- Plosives --- p.58Chapter 4.5.1.2 --- Fricatives --- p.59Chapter 4.5.2 --- Concatenation at Final segments --- p.60Chapter 4.5.2.1 --- V group (long vowel) --- p.60Chapter 4.5.2.2 --- D group (diphthong) --- p.61References --- p.63Chapter 5. --- Unit Selection for Sub-syllable Units --- p.65Chapter 5.1 --- Basic requirements in unit selection process --- p.65Chapter 5.1.1 --- Availability of multiple copies of sub-syllable units --- p.65Chapter 5.1.1.1 --- "Levels of ""identical""" --- p.66Chapter 5.1.1.2 --- Statistics on the availability --- p.67Chapter 5.1.2 --- Variations in acoustic parameters --- p.70Chapter 5.1.2.1 --- Pitch level --- p.71Chapter 5.1.2.2 --- Duration --- p.74Chapter 5.1.2.3 --- Intensity level --- p.75Chapter 5.2 --- Selection process: availability check on sub-syllable units --- p.77Chapter 5.2.1 --- Multiple copies found --- p.79Chapter 5.2.2 --- Unique copy found --- p.79Chapter 5.2.3 --- No matched copy found --- p.80Chapter 5.2.4 --- Illustrative examples --- p.80Chapter 5.3 --- Selection process: acoustic analysis on candidate units --- p.81References --- p.88Chapter 6. --- Performance Evaluation --- p.89Chapter 6.1 --- General information --- p.90Chapter 6.1.1 --- Objective test --- p.90Chapter 6.1.2 --- Subjective test --- p.90Chapter 6.1.3 --- Test materials --- p.91Chapter 6.2 --- Details of the objective test --- p.92Chapter 6.2.1 --- Testing method --- p.92Chapter 6.2.2 --- Results --- p.93Chapter 6.2.3 --- Analysis --- p.96Chapter 6.3 --- Details of the subjective test --- p.98Chapter 6.3.1 --- Testing method --- p.98Chapter 6.3.2 --- Results --- p.99Chapter 6.3.3 --- Analysis --- p.101Chapter 6.4 --- Summary --- p.107References --- p.108Chapter 7. --- Conclusions and Future Works --- p.109Chapter 7.1 --- Conclusions --- p.109Chapter 7.2 --- Suggested future works --- p.111References --- p.113Appendix 1 Mean pitch level of Initials and Finals stored in the inventory --- p.114Appendix 2 Mean durations of Initials and Finals stored in the inventory --- p.121Appendix 3 Mean intensity level of Initials and Finals stored in the inventory --- p.124Appendix 4 Test word used in performance evaluation --- p.127Appendix 5 Test paragraph used in performance evaluation --- p.128Appendix 6 Pitch profile used in the Text-to-Speech system --- p.131Appendix 7 Duration model used in Text-to-Speech system --- p.13