VOICE SYNTHESIZING METHOD, VOICE SYNTHESIZER AND COMPUTER READABLE MEDIUM RECORDED WITH VOICE SYNTHESIS PROGRAM

摘要

PROBLEM TO BE SOLVED: To enable a natural voice to be synthesized by absorbing the difference between an arbitrarily inputted character string and the character string included in a disctionary to a high degree. SOLUTION: The accent type of an inputted character string is decided (s1) and a rhythm model is selected from a rhythm dictionary in which representative rhythm models among rhythm models expressing rhythmic information with respect to character strings in a word dictionary are included based on the inputted character string and its accent type (s2). When the character string of the selected rhythm model does not coincide with the inputted character string, the rhythmic information of the rhythm model is deformed (s3) and waveform data corresponding to respective characters of the inputted character string are selected from a waveform disctionary based on the rhythm model posterior to the deformation (s4) and selected waveforms with each other are connected (s5). Thus, a natural voice message is produced.