| 摘要 |
PURPOSE:To easily synthesize a natural speech by determining a mora interval by using a consonant and a speaking speed between two moras as parameters. CONSTITUTION:A phoneme continuance time length generation part 2 section a phoneme symbol sequence generated by a text analytic part 1 into section units of VCV(vowel/consonat/vowel). The continuance time length between vowel part energy gravity centers is calculated from a rhythm rule containing a consonant and speaking speed SR is section units sectioned by the section sectioning part as parameters by a vowel part energy-gravity center time length setting part and set to each section sectioning part. After the continuance time length between the vowel part energy gravity centers at a desired speaking speed is obtained, the continuance time length of each vowel and consonant present in the continuance time length between the vowel part energy gravity centers is determined and outputted to an acoustic amplitude pattern generation part 3, a pitch pattern generation part 4, and a sound source generation part 5. |
