| 摘要 |
PROBLEM TO BE SOLVED: To realize a flexible rhythm control for improving the nature of intonation of the whole of a sentence. SOLUTION: In a voice synthesizer including a language processing part 12, a phoneme processing part 13, an acoustic parameter generating part 14 and a voice waveform generating part 15, the language processing part 12 is constituted of a rhythmic word forming part 121, an inter-rhythmic word dependency intensity determining part 122, an inter-rhythmic word modification analyzing part 123, a rhythmic word receiving number determining part 124 and an inter-adjacent rhythmic word connection intensity determining part 125 and the phoneme process part 13 is constituted of a rhythmic phrase boundary intensity determining part 131, a rhythm information generating part 132 and a phoneme sign and rhythm sign generating part 133. Thus, the connection intensity between adjacent rhythmic words is determined by using at least one of the depth of boundary and the number of modifications to be obtained from dependency intensitie and modification relation between rhythmic words and the boundary intensity of a rhythmic phrase is determined by the obtained connection intensity between the adjacent rhythmic words to generate rhythm information. |
