KR20120109763A

KR20120109763A - Apparatus and method for analyzing information of polyphonic sound source using neural computer

Info

Publication number: KR20120109763A
Application number: KR1020110027308A
Authority: KR
Inventors: 김웅겸
Original assignee: 후퍼소프트 주식회사
Priority date: 2011-03-28
Filing date: 2011-03-28
Publication date: 2012-10-09

Abstract

PURPOSE: A music information analysis apparatus and a method thereof which uses a neural network computing are provided to obtain basic information which consists of music by using music. CONSTITUTION: Neural network learning is performed for voice extraction. Neural network learning is performed for harmony pattern recognition. A polyphonic harmony is extracted to monophonic harmony. The pitch of the extracted voice is analyzed(5). Harmony is analyzed from a polyphonic harmony sound source(9). The analyzed data is re-configured as a music file(11). [Reference numerals] (1) Bit analysis; (10) Finishing; (11) Music file; (2) Key analysis; (3) Normalizing a data C key standard; (4) Allocating voice; (5) Analyzing a single harmony note; (6) Correlation; (7) Pitch class profile; (8) Comparison; (9) Polyphonic harmony sound source; (AA) Music file; (BB,CC) Window(block) determination; (DD) Reference pitch class profile; (EE,FF) Nerve network usage

Description

Apparatus and method for analyzing information of polyphonic sound source using neural computer}

The present invention relates to music information retrieval. More particularly, the present invention analyzes music information on multiple sound sources (hereinafter referred to as music) including a single source sound source (monophonic) to generate a note. The present invention relates to an apparatus and method for analyzing music information of multiple sound sources using neural network computing.

In general, music information analysis refers to an analysis of digitized music data through algorithms to obtain information on elements (pitch, chord, voice, etc.) that make up music. Most MIR techniques have been mainly studied on key analysis, chord analysis, and source partitioning through rule-based algorithms, but it is very difficult to commercialize them because of their low accuracy when applied to off-sample music. It is.

In addition, conventional music information analysis systems are often configured for a single purpose only, and there is no system for analyzing overall information on music.

Therefore, the problem to be solved by the present invention in order to solve the above-described problems, by applying a neural network (Neural-network) technology to machine (computer) by itself to improve the accuracy of music information analysis, the overall information (music) It is to provide an apparatus and method for analyzing music information of multiple sound sources using neural network computing, which can analyze voice-multiplexed sound as a single chord-extraction, rhythm, pitch, and chord.

In addition, another technical problem to be solved by the present invention is to provide an apparatus and method for analyzing music information of multiple sound sources using neural network computing, which enables composition of music scores from unspecified music.

Problems to be solved by the present invention are not limited to the above-mentioned problems, and other problems not mentioned will be clearly understood by those skilled in the art from the following description.

According to an aspect of the present invention, there is provided an apparatus and method for analyzing music information of multiple sound sources using neural network computing, comprising: a first process of performing neural network learning for voice extraction; A second process of neural network learning for chord pattern recognition; A third step of extracting a polyphonic into a monophonic; A fourth step of analyzing a pitch of each extracted voice; A fifth step of analyzing the chord from the multiple chord sound source; And a sixth step of reconstructing the analyzed data into the music score file.

Specific details of other embodiments are included in the detailed description and the drawings.

According to the apparatus and method for analyzing music information of multiple sound sources using neural network computing according to an embodiment of the present invention as described above, one or more of the following effects exist.

First, since the basic information for composing music using music can be obtained, it can be used in various applications.

Second, the rhythm information may be applied to a rhythm action game or the chord information may be applied to a performance program.

Third, it can be used in music education programs using the generated sheet music.

The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned can be clearly understood by those skilled in the art from the description of the claims.

1 is a block diagram of an apparatus for analyzing music information according to an embodiment of the present invention.

Advantages and features of the present invention, and methods of achieving the same will become apparent with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, it is to be understood that the present invention is not limited to the disclosed embodiments, but may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. It is intended that the disclosure of the present invention be limited only by the terms of the appended claims.

Also, the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. In the present specification, the singular form includes plural forms unless otherwise specified in the specification. As used herein, "comprises" and / or "comprising" does not exclude the presence or addition of components other than the mentioned components. Unless otherwise defined, all terms (including technical and scientific terms) used in the present specification may be used in a sense that can be commonly understood by those skilled in the art.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings in order to describe the present invention in more detail. Like reference numerals refer to like elements throughout.

1 is a block diagram illustrating an apparatus for analyzing music information of the present invention, wherein the apparatus for analyzing music information of the present invention is mounted on a terminal capable of driving a software, for example, a PC, a PDA, a mobile phone, or the like in hardware or software. .

The music information analyzer is composed of voice extractors (1, 3, 4, 5), chord extractors (2, 6, 7, 8, 9), and an output unit (10, 11) composed of scores using analyzed data. It consists of

Each neural network must be trained using sample data to determine neural network coefficients. This requires more learning for higher accuracy.

When a music file (MOV / MP3 / MP4, etc.) is input, the music information analyzer first determines a block size of PCM data through bit analysis (1), and then finds a key used for each music section (2).

After each voice is normalized (3) with data using the C-major as the main key, each voice is extracted using the learned neural network (4).

The pitch of the extracted single chord voice is analyzed (5).

In the chord extraction section, autocorrelation coefficient data is determined from a given multiple chord sound source (6), and a pitch class profile is constructed (7).

Obtaining the PCP Data Configured as Key Key Values The continuous chord (9) data is generated using the comparison with the reference PCP (8) and the neural network.

The single voice pitch data 5 from the voice extracting unit and the chord data 9 from the chord extracting unit are collected (10) and output as a score file (11).

As described above, the present invention has been described with reference to the embodiments shown in the drawings, but it is only for the purpose of describing the present invention, and those skilled in the art to which the present invention pertains various modifications or equivalents from the detailed description of the invention. It will be appreciated that one embodiment is possible. Therefore, the true scope of the present invention should be determined by the technical spirit of the claims.

Claims

A first process of neural network learning for voice extraction;
A second process of neural network learning for chord pattern recognition;
A third step of extracting a polyphonic into a monophonic;
A fourth step of analyzing a pitch of each extracted voice;
A fifth step of analyzing the chord from the multiple chord sound source;
And a sixth process of reconstructing the analyzed data into the music score file, wherein the music information analysis method of the multiple sound sources using neural network computing.