臺灣博碩士論文加值系統

本論文乃針對華語語音三個捲舌輔音及其相對之非捲舌輔音，使用電腦語音訊號就時域及頻域的參數做比較性的研究。此三個捲舌輔音為ㄓ ( /zh/ )、ㄔ ( /ch/ )、ㄕ ( /sh/ )，其相對之非捲舌輔音分別為ㄗ ( /z/ )、ㄘ ( /c/ )、ㄙ ( /s/ ) 。研究對象為30位講正常華語的男女，他們按照華語音標所發的音被錄入電腦中，經過一般語音訊號處理後，擷取出語音參數做分析，本研究所使用的語音參數為: 音長、跨零率、反射係數、共振頻等。實驗顯示，在時域方面: 一般而言捲舌輔音的音長較短、跨零率較小，且反射係數前後值之差的絕對值之總和較大。在頻域方面: 由頻譜圖我們發現，捲舌輔音的第一共振頻 (F1)，頻率較低，但能量較非捲舌輔音者大。本實驗結果將有助於日後華語語音之分析及辨識。

In this paper, a comparative study of three pairs of the retroflex versus non-retroflex consonants of Chinese speech is conducted. They are /zh/ vs. /z/, /ch / vs. /c/, and /sh/ vs. /s/. The subjects are 30 female or male speakers who can speak normal Chinese. Their speech according to the Chinese phonetic alphabets is recorded into a computer and processed by ordinary speech signal processing to extract various speech parameters. These parameters include speech duration､zero-crossing rate､reflection coefficients and formants. Experiments show that the retroflex consonants generally possess, in time domain, a shorter speech duration﹑a lower zero-crossing rate､and a larger sum of the absolute difference of the reflection coefficients. In frequency domain, the retroflex consonants possess a lower first formant (F1) but larger energy at that frequency. This study may be beneficial to the Chinese speech analysis and/or recognition.

中文摘要 i
英文摘要 ii
致謝 iii
目錄 iv
圖表目錄 vi
符號說明 viii
第一章 緒論 1
1.1 電腦作語音辨識 1
1.2 電腦用於中文語音之辨識 1
1.3 華語語音結構 2
1.4 電腦語音訊號處理 4
1.4.1 語音訊號輸入處理 5
1.4.2 語音參數擷取 8
1.5 研究動機 9
1.6 研究方法 9
1.7 論文架構 10
第二章 語音參數簡介 11
2.1 音長 11
2.2 跨零率 11
2.3 頻譜和共振頻 12
2.3.1 頻譜 12
2.3.2 共振頻 13
2.4 線性預測係數 14
2.4.1 發音模型 14
2.4.2 最小均方誤差 17
2.4.3 求取線性預測係數之方法 20
2.5 反射係數 23
2.5.1 無損失均勻管子的模型 23
2.5.2 多節管子的模型 30
第三章 實驗程序及實驗結果的分析 37
3.1實驗程序 37
3.2實驗一: 音長 37
3.2.1 結果及分析 37
3.3 實驗二: 跨零率 38
3.3.1 結果及分析 39
3.4 實驗三: 反射係數 40
3.4.1 結果與分析 40
3.5 實驗四: 頻譜 43
3.5.1結果及分析 43
第四章 結論 46
參考文獻 47
發表目錄 50
自傳 51

[1] 王駿發，1996，“電腦中文語音辨認系統”，資訊與教育，52期，頁3-9，七月。

[2] Ren-Yuan Lan, Lee-Feng Chien, Shiao-hong Hwang, Hung-Yun Hsieh, Rung-Chiuan Yang, Bo-Ren Bai, Jia-Chi Weng, Ten-Ju Yang, Shi-Wei Lin, Ken-Jiann Chen, Chiu-Yu Tseng, Lin-Shan Lee, 1995, “Golden Mandarin (Ⅲ)-A User-Adaptive Prosodic-Segment-Based Mandarin Dictation Machine For Chinese Language With Very Large Vocabulary”, Proc.ICASSP-95, USA, pp.57-60.

[3] 陳榮貴，褚芳達，陳建宏，黃英峰，1994，“TL語音查號系統”，交通部電信電信研究所論文。

[4] 許志興，王駿發，1992，“不特定語者之連續中文數字辨認系統”，國立成功大學，資訊工程研究所，碩士論文。

[5] 詹嘩晴，王駿發，1993，“詞彙量無關之中文詞組辨認系統”，國立成功大學，資訊工程研究所，碩士論文。

[6] 吳富政，吳宗憲，1994，“以單音節為樣本之不特定語者少量國語詞彙辨認”，國立成功大學，資訊工程研究所，碩士論文。

[7] Jyh-Shing Shyuu, Jhing-Fa Wang, and Chung-Hsien Wu, 1994, “ A Robust Vocabulary Mandarin Speech Recognition System With A Friendly Human-Interface Design”, Proceedings of 1994 Global Cooperative Software Development Conference, pp.45-55.

[8] Nanning Zheng, Guorong Xuan, 1983,“A Chinese Speech Recognition System”, Proceedings of ICASSP’83, Vol. 8, pp.308-311.

[9] Yong Qin , Fu-Yuang Mo, Chang-Li Li, Ding-Hua Guan, 1996, “ Chinese Speech Recognition With Very Large Vocabulary”, Processing of ICSLP’96,Vol.1, pp817-820.

[10] Jim Jian-Xiong Wu , Li Deng , Jacky Chan, 1996, “Modeling Context-Dependent Phonetic Units In A Continuous Speech Recognition System for Mandarin Chinese”, Processing of ICSLP’96, pp.2281-2284.

[11] Lin-Shan Lee, Chiu-yu Tseng and Ming ouhyoung, 1989, “The Synthesis Rules In A Chinese Text-To-Speech System”, IEEE Trans. On Acoustics,Speech and Signal Processingm, Vol.37, pp. 1309-1320.

[12] Bin Ma, Taiyi Huang, Bo Xu, and Xijun Zhang, 1996, “Context-Dependent Acoustic Models For Chinese Speech Recognition”, Proceedings of ICASSP’96, Vol.1, pp. 455-458.

[13] 國立台灣師範大學國音教材編輯委員會，2003，國音學，六版，正中書局，台北，頁113-202。

[14] Hsin-Min Wang et al, 1997, “Complete Recognition Of Continuous Mandarin Speech For Chinese Language With Very Large Vocabulary Using Limited Training Data”, IEEE Trans. On speech and audio processing, Vol.5, No2, pp.195-200.

[15] Y.W. Wong et al, 1999, “Acoustic Modeling And Language Modeling For Cantonese LVCSR”, Proc. EUROSPEECH’99, Vol.3, pp.1091-1094.

[16] Thomas F. Quatieri, 2002, “Discrete-Time Speech Signal Processing Principles and Practice”, Prentice Hall Signal Processing Series, NJ, USA,PP. 72-73.

[17] Oppenheim A.V. and Schafer R.W., 1989, “Discrete-Time Signal Processing”, Prentice Hall, Englewood Cliffs, NJ, pp. 468-471.

[18] Ligang Zhou, Hideaki Seki, and Ken’iti Kido, 1998 “An Investigation Of The Chinese Rolling Tongue Consonants By Time-Frequency Analysis”, Proceedings of ICSP’98, PP. 698-701.

[19] Fang Zheng, 1999, “A Syllable-Synchronous Network Search Algorithm For Word Decoding In Chinese Speech Recognition”, Proceedings of ICASSP’99, pp. 601-604.

[20] Jialu Zhang, Shinan Lu, and Shiqian Qi, 1982, “A Cluster Analysis Of The Perceptual Features Of Chinese Speech Sounds”, Journal of Chinese Linguistics, Vol. 10, pp.190-206.

[21] Rabiner L.R. and Schafer R.W., 1978, “Digital Processing of Speech Signals”, Prentice Hall, Englewood Cliffs, NJ.

[22] Yiu-kie Lau and Chok-ki Chan, 1985, “Speech Recognition Based On Zero Crossing Rate and energy”, IEEE Trans. On speech and signal processing, Vol.33, No1, pp.320-323.

[23] Peterson G. E. and Barney H. L., 1952, “Control Methods Used In A Study of the Vowels”, J. Acoust . Soc. Am., Vol.24, NO2, pp. 175-184.

[24] Yung-Sheng Hsiao and Childers D. G., 1996, “A New Approach To Formant Estimation And Modification Based On Pole Interaction”, Signals, Systems and Computers, 1996. 1996 Conference Record of the Thirtieth Asilomar Conference on , Vol.1 , No 3-6, pp. 783-787.

[25] Atal B.S. and Hanauer S.L., 1971, “ Speech Analysis And Synthesis By Linear Prediction Of The Speech Waveform”, . Acoustical Society of America, Vol.50, pp. 637-655.

[26] Makhoul J., 1973, “Spectral Analysis Of Speech By Linear Prediction”, IEEE Trans., On Audio and Electroacoustics, Vol.21, No.3, pp.140-148.

[27] Makhoul J., 1975, “Linear Prediction: A Tutorial Review”, Proc. IEEE, Vol.63, pp. 561-580.

[28] Markel J. D. and Gray Jr. A. H., 1973, “On Autocorrelation Equations As Applied To Speech Analysis”, IEEE Trans. On Audio and Electroacoustics, Vol.21, No2, pp.69-79.

[29] Ackenhusen J., 1987, “Regular form of Durbin's Recursion For Programmable Signal Processors”, IEEE Trans. On Acoustics, Speech and Signal Processing , Vol.35, No11, pp. 1628-1629.

[30] 王小川，2004，語音訊號處理，一版，全華書局。

[31] Portnoff M. R., 1973, “Quasi-One-Dimensional Digital Simulation For The Time-Varying Vocal Tract”, Masters Thesis, Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science.