| 摘要 |
<P>PROBLEM TO BE SOLVED: To enable natural singing composition approximate to human singing based on the performance data input in real time. <P>SOLUTION: Transition state information representing the attack transition Attack corresponding to phoneme transition from a consonant (s) to a vowel (a), transition state information representing note transition NtN corresponding to phoneme transition from a consonant (a) to a vowel (i), and transition state information representing the release transition Release corresponding to phoneme transition from a consonant (a) to a silence Sil are generated. During the generation of singing voice, sound source control is so performed as to add a minute change of pitch, amplitude, etc., to the phoneme transition part corresponding to the singing voice according to various pieces of the transition state information. A vibrato effect may be added to the singing voice according to the control information of vibrato tracks. <P>COPYRIGHT: (C)2006,JPO&NCIPI |
