| 摘要 |
<P>PROBLEM TO BE SOLVED: To provide an automatic music transcription apparatus which can automatically transcript a musical notation from acoustic signals played on a single musical instrument even when two or more sounds are played simultaneously. <P>SOLUTION: The automatic musical transcription apparatus comprises: an input part 1 for inputting an acoustic signal and a harmonic power ratio detecting part 2 for detecting a harmonic power ratio of a sample input acoustic signal of a musical instrument to a fundamental tone; a storage part 3 for storing the ratio; a scale note detecting part 4 for detecting power of each scale note from the above acoustic signal; a harmonic eliminating part 5 which subtracts, about each scale note, the product of the fundamental tone power multiplied by a power ratio of each harmonic to a scale note of the fundamental tone from power of the scale tone assuming that each scale note is a fundamental tone and performs the subtraction to all the scale notes sequentially from a lowest scale note; a note information detecting part 6 which extracts scale notes having power of a threshold value or higher after processing by the harmonic eliminating means and detects note information; and a detection result output part 7 for outputting a detection result of the above note information to a file or the like. <P>COPYRIGHT: (C)2007,JPO&INPIT |
