| 摘要 |
PURPOSE:To facilitate the synthesis of a sound similar to a human voice and to improve the naturalness of the sound by controlling the sectional area of an acoustic pipe before the energy of syllable data by supplying the syllable data to the acoustic pipe and varying its sectional area. CONSTITUTION:An input character is processed by a Japanese processing part 1 by using a language registered in a preliminarily provided dictionary registration part 2 and then the processed language is sectioned by a syllable processing part 3 by pitch, energy, and continuance to generate the syllable data, which is supplied to the acoustic pipe to vary the sectional area. At this time, the sectional area of the acoustic pipe is controlled a constant time before the energy to obtain a voice synthesis output. Consequently, the sound similar to the human voice can be synthesized and the naturalness is improved. |
