| 摘要 |
PROBLEM TO BE SOLVED: To make automatically switchable a conversion mode according to the kinds of music and the sex of a singer when voice conversion between the voices of man and woman is performed by comparing the judging result of a voice analyzing means with the sex of the vocal of the music and specifying the direction of the voice conversion according to the comparison result. SOLUTION: When a user instructs for a voice conversion mode in a KARAOKE performance, the input voice of a male single is analyzed by a voice analysis part 321 and its interval is compared with 1st and 2nd threshold values PH and PL. Here, when the input voice is a low-tone voice typical of a male, its interval is lower than the 2nd threshold value PL, so it is decided that the singer is a male. A parameter part 323 compares the sex decision results of both a sex decision part 322 and the voice analysis part 321 and from this comparison result, it is recognized that the man is singing the music of a female vocal. Consequently, the sex conversion mode from the man to the woman is specified to a voice change part 324. |
