| 摘要 |
<P>PROBLEM TO BE SOLVED: To accurately evaluate similarity of a fundamental frequency of a performance sound and a fundamental frequency of a model performance sound. <P>SOLUTION: A probability density function generating unit 115a divides performance sound data which are output from a sound collecting section 104 into a sequence of frames each having a predetermined time length, and generates the probability density function of the fundamental frequency for each frame of the performance sound which is indicated by the performance sound data. A probability density function providing unit 115a divides waveform data in music data which are read from a music storage section 104, into the sequence of frames each having the same time length as that frame, and generates a second probability density function regarding the fundamental frequency of the model performance sound indicated by the waveform data by the same calculation method with the probability density function providing unit 115a in a frame unit. A similarity evaluation unit 116 calculates the similarity of the first probability density function and the second probability density function by frame. <P>COPYRIGHT: (C)2009,JPO&INPIT |
