Unvoiced and Voiced Phoneme Independent Component Analysis

被引:0
作者
Wei, Hong [1 ]
Shi, Xinling [1 ]
Pu, Yuanyuan [1 ]
Yang, Jian [1 ]
机构
[1] Yunnan Univ, Sch Informat Sci & Technol, Kunming 650091, Yunnan, Peoples R China
来源
2010 6TH INTERNATIONAL CONFERENCE ON WIRELESS COMMUNICATIONS NETWORKING AND MOBILE COMPUTING (WICOM) | 2010年
关键词
speech synthesis; formants; F0; power spectrum; independent components ordering;
D O I
暂无
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Speech is assumed that its production come from some different vocal sources. Regarding the unvoiced and voiced phoneme, based on the independent component analysis, we separate its independent components from a specific phoneme and discuss the formants, fundamental frequency F0 and power spectrum of its every independent component. With the help of different independent component acoustic feature, the independent components of the specific phoneme are ordered. The different timbre phoneme is synthesized by different feature processing. Our analysis is proved to be effective by the experimental results.
引用
收藏
页数:4
相关论文
共 5 条
[1]  
Hyvarinen A., 2007, INDEPENDENT COMPONEN, P6
[2]   Phoneme recognition using ICA-based feature extraction and transformation [J].
Kwon, OW ;
Lee, TW .
SIGNAL PROCESSING, 2004, 84 (06) :1005-1019
[3]  
Yang F.S., 2006, PRINCIPLE APPL INDEP
[4]  
Young S., HTK BOOK VERSION 3 0
[5]   Statistical parametric speech synthesis [J].
Zen, Heiga ;
Tokuda, Keiichi ;
Black, Alan W. .
SPEECH COMMUNICATION, 2009, 51 (11) :1039-1064