摘要 |
PURPOSE:To synthesize plural kinds of voices and to reduce labor for generating a voice element piece parameter file of a syllable of a voiceless vowel by referring in common to the voice element piece parameter of one kind of voiceless vowel irrespective of a kind of voices to be synthesized and generating a voice parameter train. CONSTITUTION:In a voice element piece parameter file 3, as for other syllable than a voiceless vowel, voice element piece parameters corresponding to each kind of voices to be synthesized such as a man's voice, a woman's voice and a child's voice, etc., are registered in advance, and also, as for the syllable of the voiceless vowel, one kind of voice element piece parameter is registered in advance. Subsequently, a voice parameter train generating means 2 generates a voice parameter train by referring in common to the voice element piece parameter of one kind of voiceless vowel registered in the voice element piece parameter file 3, irrespective of a kind of the voices to be synthesized. In such a way, plural kinds of voices can be synthesized, and the labor for generating the voice element piece parameter file of the syllable of the voiceless vowel is reduced. |