臺灣博碩士論文加值系統

語音轉換的效能取決於其對映函數是否能充分對映兩語者之特徵參數。前人研究乃基於向量碼書對映之技術，但因其存在量化失真，使其語音轉換效能降低。本論文則以連續機率模型來描述語音特徵參數，分別基於統計分析以及高斯混合模型之觀點，以求取最佳對映函數。鑑於對映函數的訓練需要大量的語料量，而語音特徵參數彼此間之相關性極高，因此引入主成分分析作維度降低的預處理，有效地降低對映函數訓練所需之語料量，同時提昇模型訓練之收斂速度。本論文更進一步分析國語基本音節的發聲特徵，對不同音類的語音設計其最佳對映函數，並將其結果應用在聽障者之發聲矯正上。根據實驗結果顯示，在發聲缺陷最為嚴重的擦音及塞擦音之矯正上，其效果斐然。

The performance of voice conversion depends on the mapping function with the aim to convert the characteristic features from the source speaker to the target speaker. Previous research is based on vector cookbook mapping, but the converter’s performance is degraded due to the quantization noise. To overcome this limitation, we proposed two mapping functions based on continuous probabilistic models. One is based on a statistical model, and the other is based on a Gaussian mixture model. To save that the training data, we exploit the high correlation of speech characteristic features, and employ the principal component analysis to reduce the dimension of characteristic features. Simulation results indicate that the proposed mapping function helps to enhance the hearing-impaired speech, especially the fricatives and affricates.

第一章 緒論 ………………………………………………………...1
1.1動機與方向 ………………………………………………….1
1.2章節概要 …………………………………………………….3
第二章 語音模型分析與預處理…………………………………….4
2.1 語音的分析與合成 ..…………………………………………4
2.1.1 諧波正弦分析 ….………………………………………5
2.1.2 諧波正弦合成 ……………….…………………………7
2.2 訓練資料的預處理 …..………………………………………9
2.2.1 主成分分析 …….………………………………………9
2.2.1.1 基本原理 …………………………………………10
2.2.1.2 演算法則 …………………………………………12
2.2.1.3 效能評估 …………………………………………13
2.2.2 動態時間校準 ...………………………………………15
第三章 基於統計分析之轉換函數 …………….…………………20
3.1 模型描述 ……………………………………………………20
3.1 參數訓練 ……………………………………………………22
第四章 基於高斯混合模型之轉換函數 …….……………………26
4.1 高斯混合模型 ………………………………………………26
4.2 對映函數及其訓練 …………………………………………28
第五章 語音轉換之系統模擬 ……………….……………………32
5.1 國語基本音節 ………………………………………………32
5.2 對映函數之比較 ……………………………………………35
5.3 主成分分析之於對映函數效能評估 ………………………37
5.4 聽障語者之發音矯正實驗 …………………………………39
第六章 結論與未來展望 …………………….……………………46
6.1結論 …………….……………………………………………46
6.2未來展望 ……….……………………………………………47
參考文獻 ……………………………………….……………………48

[1] M. Abe, S. Nakamura, K. Shikano, and H. Kuwabara, “Voice conversion through vector quantization,” Proc. IEEE ICASSP-88 Acoustics, Speech, and Signal Processing, vol. 1, pp. 655-658, 1988.
[2] Y. M. Cheng, D. O’Shaughnessy, and P. Mermelstein, “Statistical Recovery of Wideband Speech from Narrowband Speech,” IEEE Trans.
[3] Y. Stylianou, O. Cappe, and E. Moulines, “Continuous probabilistic transform for voice conversion,” IEEE Trans. Speech and Audio Processing, vol. 2, pp. 544-548, October 1994.
[4]李承龍, 「語音轉換及其在異常發聲矯正之應用」, 國立交通大學碩士論文，民國九十年。
[5] R. J. McAulay and T. F. Quatieri, “Speech analysis-synthesis based on a sinusoidal representation,” IEEE Trans. Acoust., Speech and Signal Proc., ASSP-34(4):744-754, 1986.
[6] R. J. McAulay and T. F. Quatieri, “ Speech Transformations Based on a Sinusoidal Representation,” IEEE Trans. Acoustics, Speech, and Signal Processing, vol. ASSP-34, No.6, pp 1449-1464, December 1986.
[7] I.T. Jolliffe, “Principal Component Analysis, ” Springer-Verlag, 1986.
[8] Hallowell Davis and S. Richard Silverman, “Hearing and Deafness”, Edition, Holt, Rineheart and Winston.
[9] H. Sskoe and S. Chiba, “Dynamic programming optimization for spoken word recognition,” IEEE Trans. Acoustic, Speech, Signal Proc., ASSP-26(1): 43-49, February 1978.
[10] R. J. McAulay and T. F. Quatieri, “Pitch estimation and voicing detection based on a sinusoidal model,” Proc. IEEE Int. Conf. Acoust., Speech and Signal Proc., pages 249-252, Albuquerque, NM, Apr. 3-6 1990.
[11] Lawrence Rabiner, Biing-Hwang Juang, “Fundamentals of Speech Recognition”, Prentice Hall, New Jersey, 1993.
[12] R. J. McAulay and T. F. Quatieri. “Low rate speech coding based on a sinusoidal model,” In S. Furui and M. M. Sondhi, editors, Advances in Speech Signal Processing, chapter 6. Marcel Dekker, New York, 1992
[13] R. J. McAulay and T. F. Quatieri. “Computationally efficient sine-wane synthesis and its application to sinusoidal transform coding,” In Int. Conf. Acoust., Speech Proc., pages 370-373, 1988.
[14] Stephen A. Zahorian, Martin Rothenberg, “Principal-components analysis for low-redundancy encoding of speech spectra”, J. Acoust. Soc. Am. 69(3), March 1981.
[15] A. Dempster, N. Laird, and D. Rubin, “Maximum likelihood from incomplete data via the EM algorithm,” J. Royal Stat. Soc., vol. 39, pp. 1-38, 1977.
[16] S. M. Kay, Fundamentals of Statistical Signal Processing: Estimation Theory. Englewood Cliffs, NJ: Prentice-Hall, 1993.
[17] Douglas A. Reynolds, Richard C. Rose, “Robust Text-Independent Speaker Identification Using Gaussian Mixture Speaker Models”, IEEE Trans. Speech and Audio Processing, vol. 3, pp. 72-83, January 1995.
[18] L. R. Rabiner, J. G. Wilpon, B. H. Juang, “A segmental K-means training procedure for connected word recognition based on whole word reference patterns,” AT&T Technical Journal, Vol. 65, No. 3, pp. 21-31, May 1986.
[19]謝國平, “語言學概論”, 三民書局, pp71-80, 民85.
[20] Lawrence Rabiner, Biing-Hwang Juang, “Fundamentals of Speech Recognition”, p163-171, Prentice Hall, New Jersey, 1993.
[21] B. Massen and D. Provel, “The effect of segmental and suprasegmental corrections on the intelligibility of deaf speech,” J. Acoust. Soc. Am., 78, pp. 877-886, 1985.
[22] R. Nickerson, “Characteristics of the speech of deaf persons,” The Volta Review, 77, pp. 342-362, 1975.
[23] W. Kinsner, A. Langi, “Speech and image signal compression with wavelets”, IEEE Communications, Computers and Power in the Modern Environment, p368-375, 1993.
[24] S. Wang, A. Sekey, and A. Gersho. “An objective measure for predicting subjective quality of speech coders.” IEEE J. Select. Areas Commun., 10(5): 819-829, June 1992.