Sixth International Conference on Spoken Language Processing
(ICSLP 2000)

Beijing, China
October 16-20, 2000

Analysis of Fundamental Frequency Contours of Standard Chinese in Terms of the Command-Response Model and its Application to Synthesis by Rule of Intonation

Changfu Wang (1), Hiroya Fujisaki (2), Ryou Tomana (2 ), Sumio Ohno (3)

(1) University of Science and Technology of China, Hefei, Anhui, China
(2) Science University of Tokyo, Noda, Japan
(3) Tokyo University of Technology, Hachioji, Japan

Previous studies by the authors have shown that the commandresponse model applies quite well to F0 contours of Standard Chinese, and can be used to extract parameters as well as to derive rules for speech synthesis. The present paper describes results of analysis of F0 contours of utterances by 6 native speakers at three speech rates, conducted to examine the effects of speech rate and speaker differences. It also describes results of analysis of syllable durations at the three speech rates. These results are then applied to derive rules and quantization levels of parameters for F0 contour generation in speech synthesis of Standard Chinese. In particular, it shows that a three-level quantization of syllable duration is sufficient for the naturalness of prosody when combined with rules for F0 contour parameters.

Full Paper

Bibliographic reference.  Wang, Changfu / Fujisaki, Hiroya / Tomana, Ryou / Ohno, Sumio (2000): "Analysis of fundamental frequency contours of standard Chinese in terms of the command-response model and its application to synthesis by rule of intonation", In ICSLP-2000, vol.3, 326-329.