摘要 |
<P>PROBLEM TO BE SOLVED: To provide a speech synthesizer etc., for generating a synthesized speech wherein a speech that a person utters and an artificial speech are mixed by a sound recording and editing system. <P>SOLUTION: When data representing a fixed form message are supplied, a speech segment editing section 5 searches a speech segment database 7 for speech segment data of a speech segment whose reading matches that of a speech segment in the fixed form message and converts its voice uttering speed. In the speech segment database 7, speech segment data of artificial sounds are made to correspond to words or speech segments that a person does not use during utterance. A speech segment editing section 5, on the other hand, predicts the meter of the fixed form message and selects searched phrase data which match respective speech segments in the fixed form message most one by one according to the meter prediction result. For speech segments that can not be selected, waveform data representing waveforms by unit speech are supplied to an acoustic processing section 41. Then the selected speech segment data and waveform data supplied to the acoustic processing section 41 are combined with each other to generate data representing a synthesized speech. <P>COPYRIGHT: (C)2006,JPO&NCIPI |