Sixth International Conference on Spoken Language Processing
(ICSLP 2000)

Beijing, China
October 16-20, 2000

A Corpus-Based Chinese Speech Synthesis with Contextual Dependent Unit Selection

Ren-Hua Wang, Zhongke Ma, Wei Li, Donglai Zhu

Department of Electronic Engineering & Information Science, University of Science & Technology of China, Hefei, China

This paper describes the realization of a corpus-based Chinese speech synthesis system, including the corpus design and unit selection procedure. The system selects the synthesis unit according to context similarity between target unit and candidate unit. Neither prosody parameter prediction nor prosody feature modification is needed. The informal test shows that the synthesized speech is quite natural, and the speaking style of original speaker is preserved because units are all from the speakerís utterances.

Full Paper

Bibliographic reference.  Wang, Ren-Hua / Ma, Zhongke / Li, Wei / Zhu, Donglai (2000): "A corpus-based Chinese speech synthesis with contextual dependent unit selection", In ICSLP-2000, vol.2, 391-394.