摘要 |
PROBLEM TO BE SOLVED: To provide method and device for synthesizing a voice scalablly changing the voice data and synthetic quality according to a use from a synthesized voice of quality of an extent obtained with a voice synthetic method based on the voice data of low capacity using a voice elemental piece of a length of a phoneme or a syllable etc., until the synthesized voice of high quality similar to a natural voice based on a voice data base of large capacity. SOLUTION: When a phoneme line and rhythm information are inputted to an elemental piece selection part 101, the elemental piece selection part 101 refers to the inputted phoneme line and rhythm information, and selects the optimum voice elemental piece data from the voice data base 12 to send them to a rhythm deformation part 102. A voice waveform, the rhythm information and rhythm boundary information, etc., are stored in the voice data base 12. The rhythm deformation part 102 deforms the voice elemental piece data selected by the elemental piece selection part 101 so as to be suited to the inputted rhythm information to send them to an elemental piece connection part 103. The elemental piece connection part 103 connects successively the elemental piece data deformed by the rhythm deformation part 102 to generate the synthesized voice. |