发明名称 VOICE SYNTHESIS METHOD
摘要 PROBLEM TO BE SOLVED: To eliminate the occurrence of the distortion at the connecting point of voice waveforms by selecting the pitch of the start point which is close to the pitch of the end point of the unit voice waveform selected from the unit phonemic symbol sequence that is one before during a unit voice waveform selection. SOLUTION: A control section 1 selects a unit voice waveform in the unit phoneme symbol sequence, that constitutes a phonemic symbol sequence, from a waveform dictionary 3. During this operation, a selection is made for the pitch of the start point which is close to the pitch of the end point of the unit voice waveform selected from the previous unit phonemic symbol sequence among uinit voice waveforms each having a phoneme matching with a certain unit phonemic symbol sequence. However, if the unit phonemic symbol sequence, which is before the unit phonemic symbol sequence, does not exist, a selection is made for the unit voice waveform having the pitch of the start point which is closer to the pitch pattern of the voice to be synthesized as much as possible. If plural unit waveforms which meet the condition above exist, a selection is made for the pitch of the end point which is close to the pitch pattern corresponding to inputted character string.
申请公布号 JPH09146577(A) 申请公布日期 1997.06.06
申请号 JP19950307899 申请日期 1995.11.27
申请人 SANYO ELECTRIC CO LTD 发明人 HIRAI HIROYUKI
分类号 G10L13/08;G10L13/06;(IPC1-7):G10L3/00;G10L5/04 主分类号 G10L13/08
代理机构 代理人
主权项
地址