Chan Wai Nang. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2006. / Includes bibliographical references (leaves 77-82). / Abstracts in English and Chinese. / Chapter Chapter1 --- Introduction --- p.1 / Chapter 1.1 --- Speaker recognition --- p.1 / Chapter 1.2 --- State of the art of speaker recognition techniques --- p.2 / Chapter 1.3 --- Motivations --- p.5 / Chapter 1.4 --- Thesis outline --- p.6 / Chapter Chapter2 --- Acoustic Features --- p.8 / Chapter 2.1 --- Speech production --- p.8 / Chapter 2.1.1 --- Physiology of speech production --- p.8 / Chapter 2.1.2 --- Source-filter model --- p.11 / Chapter 2.2 --- Vocal tract and vocal source related acoustic features --- p.14 / Chapter 2.3 --- Linear predictive analysis of speech --- p.15 / Chapter 2.4 --- Features for speaker recognition --- p.16 / Chapter 2.4.1 --- Vocal tract related features --- p.17 / Chapter 2.4.2 --- Vocal source related features --- p.19 / Chapter 2.5 --- Wavelet octave coefficients of residues (WOCOR) --- p.20 / Chapter Chapter3 --- Statistical approaches to speaker recognition --- p.24 / Chapter 3.1 --- Statistical modeling --- p.24 / Chapter 3.1.1 --- Classification and modeling --- p.24 / Chapter 3.1.2 --- Parametric vs non-parametric --- p.25 / Chapter 3.1.3 --- Gaussian mixture model (GMM) --- p.25 / Chapter 3.1.4 --- Model estimation --- p.27 / Chapter 3.2 --- Classification --- p.28 / Chapter 3.2.1 --- Multi-class classification for speaker identification --- p.28 / Chapter 3.2.2 --- Two-speaker recognition --- p.29 / Chapter 3.2.3 --- Model selection by statistical model --- p.30 / Chapter 3.2.4 --- Performance evaluation metric --- p.31 / Chapter Chapter4 --- Content dependency study of WOCOR and MFCC --- p.32 / Chapter 4.1 --- Database: CU2C --- p.32 / Chapter 4.2 --- Methods and procedures --- p.33 / Chapter 4.3 --- Experimental results --- p.35 / Chapter 4.4 --- Discussion --- p.36 / Chapter 4.5 --- Detailed analysis --- p.39 / Summary --- p.41 / Chapter Chapter5 --- Speaker Segmentation --- p.43 / Chapter 5.1 --- Feature extraction --- p.43 / Chapter 5.2 --- Statistical methods for segmentation and clustering --- p.44 / Chapter 5.2.1 --- Segmentation by spectral difference --- p.44 / Chapter 5.2.2 --- Segmentation by Bayesian information criterion (BIC) --- p.47 / Chapter 5.2.3 --- Segment clustering by BIC --- p.49 / Chapter 5.3 --- Baseline system --- p.50 / Chapter 5.3.1 --- Algorithm --- p.50 / Chapter 5.3.2 --- Speech database --- p.52 / Chapter 5.3.3 --- Performance metric --- p.53 / Chapter 5.3.4 --- Results --- p.58 / Summary --- p.60 / Chapter Chapter6 --- Application of vocal source features in speaker segmentation --- p.61 / Chapter 6.1 --- Discrimination power of WOCOR against MFCC --- p.61 / Chapter 6.1.1 --- Experimental set-up --- p.62 / Chapter 6.1.2 --- Results --- p.63 / Chapter 6.2 --- Speaker segmentation using vocal source features --- p.67 / Chapter 6.2.1 --- The construction of new proposed system --- p.67 / Summary --- p.72 / Chapter Chapter7 --- Conclusions --- p.74 / Reference --- p.77

本文通過研究同卵雙胞胎之間的共振峰曲線異同，旨在發掘「個體選擇」對各人共振峰曲線的影響，並檢驗共振峰的動態特徵是否能夠用來鑑定同卵雙胞胎的聲音。本文的實驗對象為四男四女，共八對同卵雙胞胎，皆為普通話－上海話雙語使用者。其中有一男一女兩對雙胞胎自小分開撫養，在不同的語言環境中長大。 / 在第一個實驗(第二章)中，我們採集了普通話與上海話中的複合元音/ua/，並分析了其共振峰。結果顯示，雖然同卵雙胞胎擁有相同的生理構造及類似的語言環境，他們的共振峰曲線卻存在著差異。利用共振峰曲線，分析判別法可以正確鑑別同卵雙胞胎的聲音，盡管在「跨錄音時段」的分析中，正確判別率會有所下降。我們比較了受試者的強勢語言與弱勢語言之後，發現同卵雙胞胎在他們的強勢語言中擁有更高的相似程度。而通過比較一起長大與分開撫養的雙胞胎，我們發現一起長大的雙胞胎未必比分開撫養的雙胞胎更相似。我們認為，對於同卵雙胞胎來說，兩人聲音的相似程度可能受到是否願意擁有雙胞胎兄弟姐妹這一心理因素所影響。 / 在第二個實驗(第三章)中，我們分析了上海話中韻母/ue?/的共振峰曲線。該元音在新派上海話中正在產生音變，而這一變化造成了很大的「個體內差異」，可能影響到聲音鑑別。 然而，分析判別法的結果顯示，/ue?/的共振峰曲線也能夠很好地鑑別同卵雙胞胎的聲音。但不同於前一個實驗，我們發現在發韻母/ue?/的時候，同卵雙胞胎之間可能比無關聯的話說人之間差異更大。 / This thesis investigated the formant dynamics in the speech of Shanghainese-Mandarin bilingual identical twins. The aim of this study is to find out the role “individual choices play in shaping the formant trajectories and whether the between-speaker differences in the formant trajectories are able to discriminate identical twins. The subjects were four pairs of male and four pairs of female Shanghainese-Mandarin bilingual identical twins, among which one pair of male and one pair of female twins were separated since birth, and grew up in different language environments. / In Chapter 2, the Mandarin and Shanghainese /ua/ in the read speech collected from two recording sessions were examined. The formant trajectories were extracted. Both visual inspections and statistical analyses were conducted. The results showed that despite the same organic structures and similar language environments, differences could be found between identical twins in the formant dynamics. Discriminant Analysis suggested that such differences were large enough to discriminate them, though the classification rates dropped when cross-sessional analysis was conducted. A comparison between the subjects’ dominant and non-dominant language showed that the twins were more similar in their dominant language. The data also suggested that the separated twins were not necessarily more different than those non-separated twins. Their willingness to be twins could affect the similarity of their voices. / Chapter 3 analysed the formant dynamics of the /u/ rhyme in Shanghainese in the speech of the twins. This rhyme is undergoing change in the Shanghainese spoken by the young generation. Discriminant Analysis showed that although the varied pronunciations caused large within-speaker differences, the speaker-specific features in the formant dynamics still yielded high correct classification rates. However, due to the large scope of variations, identical twins were found to be not necessarily more similar than unrelated speakers. / Finally, in Chapter 4, the implications of the results on the understanding of speech perception are discussed and the contributions to forensic speaker identification are outlined. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Zuo, Donghui. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2012. / Includes bibliographical references (leaves 90-95). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstracts also in Chinese; appendixes includes Chinese. / List of Tables --- p.vii / List of Figures --- p.viii / Acknowledgements --- p.ix / Chapter 1Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Speaker identification and speaker-specificity --- p.1 / Chapter 1.2 --- Between-speaker differences --- p.2 / Chapter 1.2.1 --- Sources of between-speaker differences --- p.2 / Chapter 1.2.2 --- Types of between-speaker differences --- p.6 / Chapter 1.3 --- Within-speaker differences --- p.8 / Chapter 1.4 --- Identical twins as ideal subjects --- p.9 / Chapter 1.4.1 --- Physiological similarities between identical twins --- p.10 / Chapter 1.4.2 --- Previous studies on the speech patterns of identical twins --- p.11 / Chapter 1.5 --- Formant dynamics as ideal parameters in speaker identification --- p.18 / Chapter 1.5.1 --- The discriminatory power of formant dynamics --- p.19 / Chapter 1.5.2 --- Previous studies on individual differences in formant dynamics --- p.21 / Chapter 1.5.3 --- Differences between this study and previous studies on formant dynamics --- p.24 / Chapter 1.6 --- Bilingualism in speaker identification --- p.24 / Chapter 1.7 --- Summary of Chapter One --- p.26 / Chapter 1.8 --- Implications of this study --- p.26 / Chapter 1.9 --- Research questions --- p.27 / Chapter 1.10 --- Organisation of this thesis --- p.27 / Chapter 2Chapter 2 --- Formant Dynamics of /ua/ in Mandarin and Shanghainese --- p.29 / Chapter 2.1 --- Introduction --- p.29 / Chapter 2.2 --- Method --- p.29 / Chapter 2.2.1 --- Subjects --- p.29 / Chapter 2.2.2 --- Questionnaire --- p.31 / Chapter 2.2.3 --- Materials --- p.33 / Chapter 2.2.4 --- Procedures --- p.34 / Chapter 2.2.5 --- Measurements --- p.35 / Chapter 2.3 --- Results --- p.35 / Chapter 2.3.1 --- FT1 --- p.36 / Chapter 2.3.2 --- FT2 --- p.37 / Chapter 2.3.3 --- FT3 --- p.38 / Chapter 2.3.4 --- FT4 --- p.39 / Chapter 2.3.5 --- MT1 --- p.41 / Chapter 2.3.6 --- MT2 --- p.42 / Chapter 2.3.7 --- MT3 --- p.43 / Chapter 2.3.8 --- MT4 --- p.43 / Chapter 2.4 --- Discriminant Analysis --- p.45 / Chapter 2.4.1 --- Collapsed-session analysis --- p.46 / Chapter 2.4.2 --- Cross-session analysis --- p.52 / Chapter 2.5 --- Summary --- p.54 / Chapter 3Chapter 3 --- Formant Dynamics of /ue?/ in Shanghainese --- p.56 / Chapter 3.1 --- Introduction --- p.56 / Chapter 3.2 --- Method --- p.58 / Chapter 3.2.1 --- Subjects --- p.58 / Chapter 3.2.2 --- Materials --- p.58 / Chapter 3.2.3 --- Procedures --- p.61 / Chapter 3.2.4 --- Auditory transcription --- p.61 / Chapter 3.2.5 --- Acoustic measurement --- p.61 / Chapter 3.3 --- Results --- p.61 / Chapter 3.3.1 --- Auditory transcriptions --- p.62 / Chapter 3.3.2 --- Formant trajectories --- p.64 / Chapter 3.4 --- Discriminant Analysis --- p.68 / Chapter 3.4.1 --- Overall correct classification rates --- p.68 / Chapter 3.4.2 --- Similarities between identical twins --- p.71 / Chapter 3.5 --- Summary --- p.73 / Chapter 4Chapter 4 --- Conclusion --- p.74 / Chapter 4.1 --- Summary of findings --- p.74 / Chapter 4.2 --- Implications --- p.76 / Chapter 4.2.1 --- Speech production --- p.76 / Chapter 4.2.2 --- Forensic speaker identification --- p.80 / Chapter 4.3 --- Limitations --- p.81 / Chapter 4.4 --- Future research --- p.82 / Chapter 4.5 --- Final conclusions --- p.83 / Appendix A --- p.85 / Appendix B --- p.87 / Appendix C --- p.89 / References --- p.90

Thesis (Ed.M.)--Boston University

Thesis (Ed.M.)--Boston University

Digitized by Kansas Correctional Industries

Digitized by Kansas Correctional Industries

Digitized by Kansas Correctional Industries

Digitized by Kansas Correctional Industries

Digitized by Kansas Correctional Industries

Digitized by Kansas Correctional Industries