摘要 |
A character string analyzing means (2) receives a character string and performs the morpheme analysis to generate a phoneme symbol string, accent instructions, descent instructions, and the durations of syllables. A sound source generating parameter calculating means (8) calculates the frequency Fo, the voiced sound source intensity Av and the voiceless sound source intensity Af in accordance with the accent instructions, descent instructions, and the durations of syllables. A sound source generating means (10) generates and outputs sound source waveforms in accordance with the Fo, Av and Af. The filter coefficient control means (13) obtains filter coefficients in accordance with a phoneme symbol string given by the character string analyzing means (2) referring to a phoneme dictionary. The voice synthesizing filter means (15) outputs synthesized voice waveforms in accordance with given sound source waveforms and the filter coefficients.
|