International Symposium on Chinese Spoken Language Processing (ISCSLP 2002)

Taipei, Taiwan
August 23-24, 2002

Voice Quality Analysis under the Pitch Effect

Dan-Ning Jiang, Jian-Hua Tao, Lian-Hong Cai

Department of Computer Science and Technology, Tsinghua University, Beijing, China

Voice quality is the perceived timbre of speech. Considering the interaction between voice quality and prosody could improve the quality of acoustic processing in speech synthesis. This paper explores voice quality in accordance with various pitch conditions, and opens out the relationship between them. To ensure the accuracy of analysis, a rich experiment data set is prepared, which contains isolated vowels produced in modal quality, and statistic methods are applied. Acoustic parameter H1-A3 is used as the measurement of voice quality, since it has been proven to be a good indicator of perceived breathiness. The study found that, generally, the amount of breathiness increases with pitch in low pitch stage, and it decreases or changes little in median and high pitch stage. The interaction between voice quality and pitch is also relevant to vowel identities and individual characteristics. Quantitative changing function is estimated through regression analysis. Experiments prove the efficiency of the estimated changing function.

Full Paper

Bibliographic reference.  Jiang, Dan-Ning / Tao, Jian-Hua / Cai, Lian-Hong (2002): "Voice quality analysis under the pitch effect", In ISCSLP 2002, paper 109.