VOICE PIECE MAKING DEVICE, VOICE SYNTHESIZING DEVICE, VOICE PIECE MAKING METHOD, VOICE SYNTHESIZING METHOD, AND RECORDING MEDIUM

摘要

PROBLEM TO BE SOLVED: To improve a natural characteristic of a synthesized voice by suppressing noise feeling caused in connection points of voice pieces. ;SOLUTION: A reproducing section 112 of a voice piece generating section 11 delivers a voice waveform previously prepared to a generating section 113 dividing into each phoneme, and the generating section 113 generates voice pieces for voice synthesizing. When voice waveforms of a vowel and voiceless friction consonant succeeding to the vowel are detected, the reproducing section 112 delivers this voice waveforms to the generating section 113 and a filter 115. The filter 115 returns the waveform in which a high frequency component is eliminated in a transient section from the vowel to the voiceless friction consonant to the generating section 113. The generating section 113 makes the waveform in which the high frequency component is eliminated as the voice waveform of the vowel in the transient section, also, makes difference between the waveform before elimination of the high frequency component and the waveform after elimination as the waveform of the voiceless friction consonant in the transient section, each voice piece is generated.;COPYRIGHT: (C)2000,JPO