MINISTRY OF EDUCATION AND TRAINING HANOT UNIVERSITY OF SCIENCE AND TECHNOLOGY NVTTIL KHÃĐN Nguyen Thi Lan NOLLYWAMOANT UDYING THE PHONE’ CHARACTERISTICS OF GLOTTALIZED TONES IN VIETNAMESE EXPRESSIVE SPEECH ADO'IONHOAL MASTER THESIS OF SCIENCE INFORMATION TECHNOLOGY. HE LOZ Hanoi — 2015 MINISTRY OF EDUCATION AND TRAINING ILANOI UNIVERSITY OF SCIENCE AND TECIINOLOGY Nguyen Thi Lan STUDYING THK PHONETIC CHARACTERISTICS OF GLOTTALIZED TONES IN VIETNAMESE EXPRESSIVE SPEECH Department INFORMATION TECHNOLOGY MAS'TER ‘THESIS OF SCIENCE INFORMATION TECHNOLOGY SUPERVISOR: Dr Tran Do Dat Hanoi 2015 COMMITMENT T commit myself to be the person who was responsible for conducting this study All reference figures were extracted with clear derivation. Ihe presented results are ‘truthful and have not published in any other person’s work. NGUYEN Thi Lan LIST OF FIGURES Vigure 1-1 Schematic diagram of Llanoi Vietnamese tones (Michaud, 2004a).
8 Figure 2-1 Speaker F7 (lel) and M10 (right) in the recording booth. Figure 2-2 Sentence and Syllable Level Annotation with SoundForge (above) oui Praat (below) of the corpus. 31 Figure 3-1 Visualization of closing instant synchronized with EGG (above) and DEGG (below) signals (enrich, 2001) .csccessccssssessssessseesssasveee 34 Figure 3-2 Visualization of opening instant synchronized with EGG (above) and DGG (balow) signals (1lenrich, 2001). 35 Figure 3-3 Fxample of KGG and DEGG sigruls with indication of glottis closure and opening 36 Figure 3-4 lwo realizations of glottalization on SKP /aGa/ with two attitudes.
37 Figure 3-5 Average curves of I'0 and Oq over 6 tokens of /a6a/, speaker M7 .38 Figure 3-6 Two realizations of glolativalion on SFP /da3/ of two attitudes, (a) declarative/neutral; (b): irritation. sone dD Figure 3-7 Average curves of FO and Og over 6 tokens of Ad33/, speaker M6. 40 Figure 3-8 Determining mechanisms of voice based on DECPA and FO parameters (each point of !U&Oq contour corresponds with a cycle on DEGG signal).42 Figure 3-9 Determining the duration of pressed voice based on local dipping of Oq (each point of !'080q contour corresponds with a cycle on DUOG signal).44 Figure 3-10 The tool for detection integrated three analysis modules. 45 Figure 3-11 Some visually illustrative figures of creaky voice from the detection tool.
47 Figure 3-12 Some visually illustrative figures of pressed voice from the detection tool. Figure 3-13 Averaged FO contours of SFP /a6a/ of 10 male speakers’ surprise and declaration. ee) Figure 3-1 Averaged Oq contours of SFP /a6a/ of 10 male speakers: surprise (left) and declaration (right).40 Figure 3-15 Averaged FO contours of SFP /ci33/ of 10 male speakers. irritation and declaration.
Tigure 3-16 Avereged Òq contours of SEP /da3/ of 10 male speakers.51 Figure 3-17 Averaged Oq contours of SFP /da3/ of 10 male speakers: irritation.53 Figure 3-18 Proposed model for combination of speaker altitude, voice auaity and glottalized tone in Vietnamese expressive speech processing. 54 ACKNOWLEDGEMENT ‘This is the second time that 1 sit here, at Hanoi University of Science and Technology, with a great honor to write these grateful wards to people who have been supporting me since the first, moment. T erfered the university. The first acknowledgement was written in my graduation thesis 2.5 years ago and today, this one just awakes a special emotion in me.
I wish to thank all my professors and colleagues at School of Information and Communication Technology and MICA International Research Institute, who have helped me with generous supports. Their advice and knowledge they imparted to me are gratefully appreciated, inspiring me a lot to finish this thesis Special thanks to my supervisor Dr. Tran Do Dat and colleagues of Speech Communication Department, MICA Tnstitute, including Dr. Do Thi Ngoe Diep, Nguyen Thi Thu Trang, Nguyen Tuan Ninh, Tran Thi Anh Xuan, Dr.
Nguyen Viet Son, Dr. Nguyen Cong Phuong, Nguyen Dục Anh and Nguyen Tie Thanh, for their advice and encouragement they gave to me, especially Dr. Mac Dang Khoa and Dr. Alexis Michaud for their thorough review and invaluable suggestions Another thanks for two thesis reviewers including Assoc Prof.
Truong Ninh ‘Thuan (VNU) and Dr. Vu Thi Luong Giang (SOICT, MUST) for their worth comments which helped the thesis’s presentation become much better. Special thanks to my family and friends who always stand by me, lifting me up when I was down. Without them, my life would be nonsense! NGUYEN Thị Lan tà LIST OF TABLES Table 1-1 Vietnamese consonants 12 ‘Table 1-2 Vietnamese vowels/diphthongs.
13 Table 1-3 Phonetic characlerislies of Viclnamese inilial consonants 14 Table 1-4 Phonetic characteristics of Vietnamese final consonants. 14 'Table 1-5 Phonetic characteristics of Vietnamese vowels/diphthongs. 18 Table 1-6 Summarized description of & tenes of Viotnamese. 16 ‘Table 2-1 Intended attitudes.
Table 2-2 List of speakers 28 ‘Table 3-1 Statistics of Mechanism I-A/Pressed Voiee/Mechanism 1-13 of tone 6a will attitude surprise 50 Table 3-2 Statistics of Mechanism I-A/Creaky Voice/Mechanism I-B of tone 3 with “.nnẽẽố 52 Table 3-3 Statistics of Mechanism I. ‘AlPressed Voice/Mechanism IBof tone 3 with attitude irritation. 53 LIST OF ABBREVIATIONS SEP — Sentence-final particle KGG Hlectroglottography DEGG — The derivative of the electroglottography signal IPA — International phonetic association DbCPA Derivative-Hlectroglottographic Closure Peak Amplitude X-SAMPA ~The Extended Speech Assessment Methods Phonetic Alphabet FO Fundamental frequency CONTENTS COMMITMENT. LIST OF ABBREVIATIONS: LIST OF TABLES.
& LIST OF FIGURES. Lt Background knowledge. Vietnamese phonetics and phonology.42 The phonetic characteristics of complex lexical tone system in Vietnamese.2 CHottalized tones in the context of expressive speech: raising issues. aoe 1S 143 The scope of the thesis.
BUILDING VIETNAMESE ATFEITUDINAL SPEECTI CORPUS FOR SENTENCE-FINAL PARTICLE: 2.1 Method of using expressive morphemes caying lexical lones — 2 2.2 Designing sample corpus.3 _ The pogress of buljding the saniple coTpS. Blicitation method and speakers.3 Past-processing and annotation - - _. ANALYSING VARIATION IN REALIZATION OF GLOTTALIZED TONES BY VARIOUS ATTITUDE 33 31 Analysis Method.2 A pilot dala analysis: importa new discoveries and an.sn awieh inlo the use& of expesve morphemes and giollalized Lonea. Comparison of attinudes: Surprise qnà Declaralion.2 Comparison of attitudes: Irritation and Declaration.
LIST OF ABBREVIATIONS SEP — Sentence-final particle KGG Hlectroglottography DEGG — The derivative of the electroglottography signal IPA — International phonetic association DbCPA Derivative-Hlectroglottographic Closure Peak Amplitude X-SAMPA ~The Extended Speech Assessment Methods Phonetic Alphabet FO Fundamental frequency INTRODUCTION Nowadays, using speech in human-machine interaction is gradually becoming the major rend which promises Lo replace traditional commumicalion methods: mouse, keyboard, sercen, for example. However, a high-quality human-machine interaction system that can completely behave as a human being, currently, is still just beyond our reach. One of the primary reasons is because of the lack of advanced techniques that enable precisely processing (either synthesis or recognition) the expression of human utterances. ‘The expression, in other words, refers to attitudinal or emotional aspects when someone speaks, which hereby can convey much linguistic information.
In this perspective, the attitudinal aspects in speaker utterances, also called speaker altiludes are of no sinall importance. If' speaker altiludes play such an iraportant role in the interactions between humans, they need to be taken into account in the imleraction between humans and machines (Picard, 1997). Atitudial mformalion in a spoken utterance can be lexically oncoded but can also be conveyed by intonation, including modifications of voice quality (Seibert, 2003) However, the modification of those features in Vietnamese is quite complex since ít has the interplay between intonation and tones: especially, the complexity even becomes much more complicated when dealing with glottalized tones which are tone ngd and tone nding. lurthermore, in expressive speech, how the interplay can be expressed, what its realization will be and with which mechanisms, are several among many questions sel oul.
Among eight tones in Vietnamese, tone ngd and tone ngng are considered the most complicated since they have glotlalizaion phenomenon accompanied. Ta mos! cascs, with simpler tones, the interaction between intonation and tone simplifies to be described by the changing in fundamental frequency, intensity or duration pararicters, whercas wilh these two glollalived tones, these parameters are exactly LIST OF TABLES Table 1-1 Vietnamese consonants 12 ‘Table 1-2 Vietnamese vowels/diphthongs. 13 Table 1-3 Phonetic characlerislies of Viclnamese inilial consonants 14 Table 1-4 Phonetic characteristics of Vietnamese final consonants. 14 'Table 1-5 Phonetic characteristics of Vietnamese vowels/diphthongs.
18 Table 1-6 Summarized description of & tenes of Viotnamese. 16 ‘Table 2-1 Intended attitudes. Table 2-2 List of speakers 28 ‘Table 3-1 Statistics of Mechanism I-A/Pressed Voiee/Mechanism 1-13 of tone 6a will attitude surprise 50 Table 3-2 Statistics of Mechanism I-A/Creaky Voice/Mechanism I-B of tone 3 with “.nnẽẽố 52 Table 3-3 Statistics of Mechanism I. ‘AlPressed Voice/Mechanism IBof tone 3 with attitude irritation.
53 LIST OF ABBREVIATIONS SEP — Sentence-final particle KGG Hlectroglottography DEGG — The derivative of the electroglottography signal IPA — International phonetic association DbCPA Derivative-Hlectroglottographic Closure Peak Amplitude X-SAMPA ~The Extended Speech Assessment Methods Phonetic Alphabet FO Fundamental frequency LIST OF ABBREVIATIONS SEP — Sentence-final particle KGG Hlectroglottography DEGG — The derivative of the electroglottography signal IPA — International phonetic association DbCPA Derivative-Hlectroglottographic Closure Peak Amplitude X-SAMPA ~The Extended Speech Assessment Methods Phonetic Alphabet FO Fundamental frequency LIST OF FIGURES Vigure 1-1 Schematic diagram of Llanoi Vietnamese tones (Michaud, 2004a). 8 Figure 2-1 Speaker F7 (lel) and M10 (right) in the recording booth. Figure 2-2 Sentence and Syllable Level Annotation with SoundForge (above) oui Praat (below) of the corpus. 31 Figure 3-1 Visualization of closing instant synchronized with EGG (above) and DEGG (below) signals (enrich, 2001) .csccessccssssessssessseesssasveee 34 Figure 3-2 Visualization of opening instant synchronized with EGG (above) and DGG (balow) signals (1lenrich, 2001).
35 Figure 3-3 Fxample of KGG and DEGG sigruls with indication of glottis closure and opening 36 Figure 3-4 lwo realizations of glottalization on SKP /aGa/ with two attitudes. 37 Figure 3-5 Average curves of I'0 and Oq over 6 tokens of /a6a/, speaker M7 .38 Figure 3-6 Two realizations of glolativalion on SFP /da3/ of two attitudes, (a) declarative/neutral; (b): irritation. sone dD Figure 3-7 Average curves of FO and Og over 6 tokens of Ad33/, speaker M6. 40 Figure 3-8 Determining mechanisms of voice based on DECPA and FO parameters (each point of !U&Oq contour corresponds with a cycle on DEGG signal).42 Figure 3-9 Determining the duration of pressed voice based on local dipping of Oq (each point of !'080q contour corresponds with a cycle on DUOG signal).44 Figure 3-10 The tool for detection integrated three analysis modules.
45 Figure 3-11 Some visually illustrative figures of creaky voice from the detection tool. 47 Figure 3-12 Some visually illustrative figures of pressed voice from the detection tool. Figure 3-13 Averaged FO contours of SFP /a6a/ of 10 male speakers’ surprise and declaration. ee) Figure 3-1 Averaged Oq contours of SFP /a6a/ of 10 male speakers: surprise (left) and declaration (right).40 Figure 3-15 Averaged FO contours of SFP /ci33/ of 10 male speakers.
irritation and declaration. Tigure 3-16 Avereged Òq contours of SEP /da3/ of 10 male speakers.51 Figure 3-17 Averaged Oq contours of SFP /da3/ of 10 male speakers: irritation.53 Figure 3-18 Proposed model for combination of speaker altitude, voice auaity and glottalized tone in Vietnamese expressive speech processing. Proposalz lora fillscalo zludy and snusticalsans of[Hhomlie Bạn based on EGG and DEGG signal.1 Observation about the irvegulantties of DECPA sou 33.2 Building « tool for detection of creaky and pressed voice in Matlab -# 3.4 A full-scale study: đetaileđ analysis nesuls.- LAT B41 Surprise and Declaration. 342 Irritation and Declaration - - - 50 3.
CONCLUSION AND PERSPECTIVES REFERENCES. PUBLICATION: APPENDIX A: COFNTEXTS OF EACH COLLECTED SENTENCE. APPENDIX B: FIGURES OF AVERAGED FU&0q CONTOURS OF EACH SPEAKER WITH STANDARD DEVIATION FOR THE USED ATTITUDES CONTENTS COMMITMENT. LIST OF ABBREVIATIONS: LIST OF TABLES.
& LIST OF FIGURES. Lt Background knowledge. Vietnamese phonetics and phonology.42 The phonetic characteristics of complex lexical tone system in Vietnamese.2 CHottalized tones in the context of expressive speech: raising issues. aoe 1S 143 The scope of the thesis.
BUILDING VIETNAMESE ATFEITUDINAL SPEECTI CORPUS FOR SENTENCE-FINAL PARTICLE: 2.1 Method of using expressive morphemes caying lexical lones — 2 2.2 Designing sample corpus.3 _ The pogress of buljding the saniple coTpS. Blicitation method and speakers.