TEXT VOICE SYNTHESIZER, TEXT VOICE SYNTHESIZING METHOD AND RECORDING MEDIUM WITH THE METHOD RECORDED THEREON

摘要

PROBLEM TO BE SOLVED: To clearly pronounce any words while an increase in the amount of beforehand registered voice element data is suppressed to an absolute minimum. SOLUTION: Respective voice element data corresponding to each fundamental voice unit are beforehand registered in a voice element database 14. Exceptional voice element data corresponding to a specific voice unit included in a specific word are beforehand registered in an exceptional voice element database 15. When voice data for a specific word are to be synthesized, the exceptional voice element data corresponding to the specific voice unit included in the specific word are retrieved from the database 15 and used. Thus, when voice data indicating the specific word are converted into voice, the specific word is clearly heard.