Prosody and style controls in CU VOCAL using SSML and SAPI XML tags

Author

Fung, Tien-Ying ; Li, Yuk-Chi ; Meng, Helen ; Ching, P.C.

Author_Institution

Human-Comput. Commun. Lab., Chinese Univ. of Hong Kong, Shatin, China

fYear

2004

fDate

15-18 Dec. 2004

Firstpage

209

Lastpage

212

Abstract

CU VOCAL is a Cantonese text-to-speech (TTS) engine. We use a syllable-based concatenative synthesis approach to generate intelligible and natural synthetic speech in Cantonese. The paper reports on our recent enhancements in CU VOCAL to support user adjustments in prosody and style with the use of the Speech Synthesis Markup Language (SSML) in the input text. CU VOCAL was previously developed as a SAPI-compliant engine to enable easy integration with other applications. The paper also reports on our enhancements in the CU VOCAL SAPI (speech API) engine to support the SAPI 5 XML tags.

Keywords

XML; application program interfaces; natural language interfaces; speech synthesis; Cantonese text-to-speech engine; SAPI XML tags; SSML; Speech Synthesis Markup Language; prosody; style; syllable-based concatenative synthesis; Communication system control; Digital signal processing; Engines; Laboratories; Markup languages; Research and development management; Speech synthesis; Systems engineering and theory; Technical Activities Guide -TAG; XML;

fLanguage

English

Publisher

ieee

Conference_Titel

Chinese Spoken Language Processing, 2004 International Symposium on

Print_ISBN

0-7803-8678-7

Type

conf

DOI

10.1109/CHINSL.2004.1409623

Filename

1409623