Sixth International Conference on Spoken Language Processing
(ICSLP 2000)

Beijing, China
October 16-20, 2000

PureTalk: A High Quality Japanese Text-to-Speech System

Masayuki Yamada, Yasuo Okutani, Toshiaki Fukada, Takashi Aso, Yasuhiro Komori

Platform Technology Development Center, Canon Inc., Kawasaki-shi, Kanagawa, Japan

This paper describes a high quality Japanese text to speech (TTS) system, PureTalk. This system is similar to the conventional diphone-based TTS using PSOLA except that PureTalk employs the following novel tech- niques which enable to produce more intelligible and nat- ural-sounding speech: 1) two-stage duration modeling based on a linear regression technique, 2) F0 contour modeling using polynomial segment models, 3) sophisti- cated waveform unit selection, and 4) ecient waveform compression designed for TTS system. The result of the subjective hearing test shows that PureTalk achieves high quality under practical computation and memory requirement.


Full Paper

Acoustic Example

Bibliographic reference.  Yamada, Masayuki / Okutani, Yasuo / Fukada, Toshiaki / Aso, Takashi / Komori, Yasuhiro (2000): "Puretalk: a high quality Japanese text-to-speech system", In ICSLP-2000, vol.2, 403-406.