| 摘要 |
<P>PROBLEM TO BE SOLVED: To provide a technique for regularly synthesizing an appropriate sound waveform by use of a parameter extracted by analyzing an original sound waveform. <P>SOLUTION: In an analysis phase, speech data output from an A/D converter 8 is analyzed in frame units. As a parameter, a PARCOR factor, a voiced speech ratio showing that sound shown by the speech data is a voiced speech, or the like is extracted. In a synthesis phase, a drive sound source waveform is generated in a manner such that a Rosenberg wave generated at a designated pitch and a white noise wave having no pitch are mixed according to the voiced speech ratio. A synthetic filter part 45 synthesizes speech data of one frame by use of the drive sound source waveform and the PARCOR factor. <P>COPYRIGHT: (C)2010,JPO&INPIT |
