| 摘要 |
PURPOSE: To adjust power of an unvoiced vowel to make it easy to hear by making a parameter of voice power different by the position (the head, the middle, or the end of a word) of the vowel in synthesizing a voice including the unvoiced vowel. CONSTITUTION: A phonemic symbol string obtained by a character string analysis part 1 is inputted to a voice parameter string generator 2, and a voice phoneme parameter file 3 is referred to generate a voice parameter string. Meanwhile, prosody information obtained by the analysis part 1 is inputted to a prosody parameter string generator 4 to generate a prosody parameter string. A voice synthesizer 5 generates a synthesized voice corresponding to an input character string based on prescribed synthesis rules in accordance with the voice parameter string and the prosody parameter string. A certain value is added to the parameter of a voice power preliminarily set in a phoneme parameter file in accordance with the position of an unvoiced vowel, thereby adjusting power of the unvoiced vowel to make it easy to hear. |
