Automatic musical composition classification device and method

Abstract

An automatic musical composition classification device and method that allow automatic classification of a plurality of musical compositions based on melodic similarity. storing chord progression pattern data representing a chord progression sequence for each of the plurality of musical pieces, extracting a chord progression change feature amount for each of the plurality of musical pieces based on the chord progression pattern data, And according to the chord progression sequence and the chord progression change feature quantity represented by the chord progression pattern data of each of the plurality of music pieces, the multiple pieces of music are grouped.

Description

technical field [0001] The invention relates to an automatic music classification device and method for automatically classifying multiple music pieces. Background technique [0002] Due to the increasing popularity of compressed music data in recent years, and the growing capacity of storage devices, individuals have been able to store and enjoy a large amount of music. On the other hand, it has become extremely difficult for users to select a large number of music pieces and find music pieces they like to listen to. Therefore, there is a need for an efficient music classification and selection method that can solve this problem. [0003] Traditional music classification methods include methods that use information appearing in reference information, such as song titles; singers; names of genres to which the music belongs, such as rock or pop; The stored music pieces are classified as disclosed in Japanese Patent Laid-Open No. 2001-297093. [0004] These methods also inc...

AI Technical Summary

Problems solved by technology

[0008] However, with the method of displaying information such as song title, genre, etc. in reference information described in Japanese Patent Publication No. 2001-297093, many problems have been encountered, namely, that this method requires work on the part of an individual , does not allow network connections, and does not work properly when classification information is difficult to obtain

[0009] In the case of using the classification method of Japanese Patent Publication No. 2002-278547, the image of the listener of the music is subject to personal feelings, and since this image is blurred, it varies even for the same listener, So continuous results cannot be expected when using an image different from the one involved for classification

Therefore, in order to maintain the effectiveness of the subjective image language, it is necessary to continuously obtain feedback from the listener for the classification operation, which raises a problem that a very laborious operation is imposed on the listener

And there is also such a problem: the classification of beats or other rhythm information is limited by the target music

However, in the classification described above, there is such a problem that there is little continuity in the tempo, key, etc. of the actual musical piece, and, for the feature quantity that allows classification by the musical piece unit, the accuracy is limited. very low

[0012] In addition, for the methods disclosed in Japanese Patent Publication Nos. 2001-297093, 2002-278547, 2000-268541, and 2003-58147, it is necessary to use a statically defined language, such as using image words, genres, and major and minor keys, for musical selection, and the impression of the music changes with the mood, so there is a problem that it is not possible to classify the music appropriately

[0013] Although Japanese Patent Laid-Open No. 2002-41059 describes a situation in which a music piece matching the listener's preference is provided when a music piece is selected, because the feature quantity actually used is The result of the extraction is submitted by converting it into a numerical value, which cannot represent changes in melody in the composition, so there is a problem that the accuracy appropriate for classifying the composition according to preference cannot be ensured

Images