| 摘要 |
PURPOSE:To expand a pitch variation range when a sound is synthesized by varying its pitch by selecting sound source codes which satisfies the range of glottis opening section length in a sound source code book used for analysis and selecting the code having the shortest spectrum distance to a sound source code obtained by analysis among the selected sound source codes. CONSTITUTION:The sound source codes satisfying the range are selected in the sound source code book 106 and used to make a voice analysis using a vocal chord sound source wave model, thereby estimating a sound source code and a glottis parameter. Then a sound source code control circuit 105 selects the code having the shortest spectrum distance to the sound source code obtained by the analysis among the selected sound source codes inputted from a sound code selecting circuit 110 and the code is regarded as a pitch-controlled sound source code. Lastly, the pitch-controlled sound source code and glottis parameter obtained by the analysis are used to synthesize a pitch-controlled voice by a synthesizing circuit 107. |
