International Symposium on Chinese Spoken Language Processing (ISCSLP 2002)

Taipei, Taiwan
August 23-24, 2002

Design of Embedded Application Oriented Distributed Speech Synthesis System with High Naturalness

Hao Tang, Bo Yin, Ren-Hua Wang

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

In this paper, a unique design scheme of embedded application oriented distributed speech synthesis system with high naturalness is presented in detail. Based on the client/server model, text is firstly converted into parameter sequence (PS) by the front-end tool at the server side. To complete the text-tospeech process, the back-end speech synthesizer at the client side converts the PS into speech upon receival of it through a certain data transmission channel. This design scheme is distinctive in that it is able to obtain highly natural speech with extremely low cost of computing and storage at the client side. Therefore, it is ideally suited for embedded devices with limited performance and storage. The feasible realization of this design scheme on the TI MSP50C614 device with capability of adjusting the speaking speed dynamically in real time is discussed in the end of the paper.


Full Paper

Bibliographic reference.  Tang, Hao / Yin, Bo / Wang, Ren-Hua (2002): "Design of embedded application oriented distributed speech synthesis system with high naturalness", In ISCSLP 2002, paper 76.