WO2017037919A1 - Musical-piece analysis device, musical-piece analysis method, and musical-piece analysis program - Google Patents

Musical-piece analysis device, musical-piece analysis method, and musical-piece analysis program Download PDF

Info

Publication number
WO2017037919A1
WO2017037919A1 PCT/JP2015/075076 JP2015075076W WO2017037919A1 WO 2017037919 A1 WO2017037919 A1 WO 2017037919A1 JP 2015075076 W JP2015075076 W JP 2015075076W WO 2017037919 A1 WO2017037919 A1 WO 2017037919A1
Authority
WO
WIPO (PCT)
Prior art keywords
sound data
analysis
reproduction time
time
musical
Prior art date
Application number
PCT/JP2015/075076
Other languages
French (fr)
Japanese (ja)
Inventor
四郎 鈴木
Original Assignee
Ｐｉｏｎｅｅｒ ＤＪ株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Ｐｉｏｎｅｅｒ ＤＪ株式会社 filed Critical Ｐｉｏｎｅｅｒ ＤＪ株式会社
Priority to PCT/JP2015/075076 priority Critical patent/WO2017037919A1/en
Priority to JP2017537158A priority patent/JPWO2017037919A1/en
Publication of WO2017037919A1 publication Critical patent/WO2017037919A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L21/00—Processing of the speech or voice signal to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
    • G10L21/04—Time compression or expansion
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/45—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of analysis window
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
    • G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination

Definitions

  • the present invention relates to a music analysis device, a music analysis method, and a music analysis program.
  • a technique for automatically analyzing music such as beats, tempos, music keys, and scales from sound data such as music is known (see, for example, Patent Document 1).
  • the sound data is analyzed by sampling sound data of a certain length and analyzing the input waveform using, for example, FFT (Fast Fourier Transform).
  • FFT Fast Fourier Transform
  • BPM Beats Per Minute
  • a DJ-related device may use short-time sound data as a sound effect or a Loop sound source.
  • An object of the present invention is to provide a music analysis apparatus, a music analysis method, and a music analysis program that can accurately analyze even very short time sound data.
  • the music analysis apparatus is Sound data duplicating means for duplicating the input sound data and making it continue so as to be longer than a predetermined reproduction time;
  • a sound data analysis means for analyzing the sound data that has reached a predetermined reproduction time and analyzing the input sound data; It is characterized by having.
  • the music analysis method includes: The procedure to duplicate the input sound data, A procedure for continuing the duplicated sound data so as to be longer than a predetermined reproduction time, Analyzing the sound data that has reached a predetermined playback time and analyzing the input sound data; It is characterized by performing.
  • the music analysis program is: The procedure to duplicate the input sound data, A procedure for continuing the duplicated sound data so as to be longer than a predetermined reproduction time, Analyzing the sound data that has reached a predetermined playback time and analyzing the input sound data; Is executed by a computer.
  • the block diagram which shows the music analysis apparatus which concerns on embodiment of this invention.
  • FIG. 1 shows a music analysis apparatus 1 according to this embodiment.
  • the music analysis device 1 analyzes digitized sound data SD such as input PCM data, determines a key of the sound data SD, and determines the sound data input on the display screen of a display device or the like. This is a device for displaying as a key display KD.
  • the music analysis apparatus 1 is configured as a software application that is activated by a general computer, a portable information terminal equipped with an OS (Operation System), and includes a reproduction time detection means 2, a sound data determination means 3, a sound data duplication means 4, and a sound.
  • Data analysis means 5, execution interval setting means 6, and key determination means 7 are provided.
  • the playback time detection means 2 detects the playback time of the input sound data SD. Specifically, the reproduction time detection means 2 detects the reproduction time by counting the number of samplings from the beginning to the end of the sound data SD. When the reproduction time is detected, the reproduction time detection unit 2 outputs the detected reproduction time to the sound data determination unit 3 and the execution interval setting unit 6.
  • the sound data determination means 3 determines whether or not the input sound data SD has a predetermined reproduction time or more based on the reproduction time detected by the reproduction time detection means 2. Specifically, it is determined whether or not the sound data SD has a reproduction time that can be analyzed by sound data analysis means 5 described later. Whether or not analysis is possible is determined by whether or not the reproduction time of the sound data SD has a time longer than the shortest time necessary for the window function applied to the sound data analysis means 5.
  • the time window length of the window function is determined by the sampling frequency of the sound data SD, the lower limit frequency to be detected, and the frequency resolution. For example, in the case of a regular 4/4 time tune of BPM200, one beat is 300 msec and a sixteenth note is 75 msec. When performing FFT and analyzing the sound data SD of the 27.5 Hz bass corresponding to A0 of the scale, data of 1.2 s at least is required. When the sound data determination unit 3 determines that the sound data SD is less than the predetermined reproduction time, the sound data determination unit 3 outputs a determination result to that effect to the sound data duplication unit 4.
  • the sound data duplicating unit 4 duplicates the input sound data SD based on the determination result of the sound data determining unit 3, and continues the data so that the predetermined time or more is reached. Specifically, as shown in FIG. 2, the sound data duplicating unit 4 duplicates the input sound data SD of the reproduction time t1, and pastes the duplicate data CD behind the sound data SD so as to be continuous. , Continuous sound data SD ′ having a reproduction time t2 is generated.
  • the sound data duplicating unit 4 duplicates the duplicate data CD of the sound data SD, and the reproduction time t2 is 1.2 seconds or longer.
  • the duplication is repeated until the sound data SD ′ is obtained. It should be noted that the duplication only needs to be performed so that the reproduction time t2 can be analyzed by the sound data analysis means 5, and the number N of duplications may not be an integer.
  • the sound data replicating unit 4 outputs to the sound data analyzing unit 5 the sound data SD ′ that has been duplicated for a predetermined time or more.
  • the sound data analysis means 5 performs frequency spectrum analysis of the sound data SD and SD ′.
  • the analysis method is performed by analysis using FFT.
  • the analysis method is not particularly limited, and for example, analysis using DCT (Discrete Cosine transform), analysis using a time axis, analysis using a signal level, volume feeling, and attack feeling may be used. it can.
  • a Hamming window HMW that is a window function is usually applied to the FFT.
  • the Hamming window HMW is applied to weaken the signal intensity amplification at both ends on the time axis during the FFT execution time and reduce the influence of the discontinuous portion of the sampled waveform seam during the FFT execution.
  • the analysis period T0 is set to a 0.7 (70%) value of the Hamming window HMW.
  • the Hamming window HMW is applied.
  • the present invention is not limited to this, and a Hanning window, a flap top window, or the like may be used as the window function.
  • the execution interval setting means 6 sets the FFT execution interval by the sound data analysis means 5 based on the reproduction time detected by the reproduction time detection means 2. Specifically, as shown in FIG. 4, the execution interval setting means 6 sets the execution start time of the second FFT2 as the execution interval TI after the first execution of FFT1. Thereafter, in the present embodiment, the third FFT3 is started after time 2TI, and the FFTs are sequentially executed.
  • the execution interval TI is set according to the reproduction time of the sound data SD and SD ′. For example, in the case of long sound data SD with a reproduction time of 30 seconds or more, the execution interval setting means 6 increases the execution interval TI as shown in FIG. 5, and in the case of short sound data SD with a reproduction time of less than 30 s, As shown in FIG. 6, the execution interval TI is reduced.
  • the lower limit value of the execution interval TI is set to an execution interval TI such that the analysis periods T0 of FFT1, FFT2,.
  • the execution interval setting means 6 outputs the set execution interval TI to the sound data analysis means 5 described above.
  • the sound data analysis unit 5 repeatedly executes FFT based on the execution interval TI, and outputs the analysis result to the key determination unit 7 every time FFT is executed.
  • the key determination unit 7 performs key determination of the sound data SD and SD ′ based on the analysis result output from the sound data analysis unit 5. Specifically, the key determination means 7 stores the reference frequencies of a total of 24 types of keys, that is, two types of minor keys and major keys rooted in each scale for each of the 12 scales per octave. The key determination means 7 adds the analysis results input at every execution interval TI in the time axis direction, and selects a reference frequency close to a frequency with a strong signal strength from the obtained sum, as shown in FIG. Next, the signal strength of each scale is obtained. Next, as shown in FIG.
  • the key determination means 7 rearranges and normalizes the signals in the order of strong signal strength, selects several scales with strong signal strength, and selects the keys of the sound data SD, SD ′. Determine.
  • the key determination means 7 displays the key determination result of the sound data SD, SD ′ as a key display KD on the computer display or the screen of the portable terminal.
  • Step S1 when a user of a computer or a portable terminal selects the music analysis device 1 on the screen, starts the program, and selects the sound data SD to be analyzed, the sound data SD is input to the music analysis device 1 ( Procedure S1).
  • the reproduction time detecting means 2 detects the reproduction time of the sound data SD (step S2).
  • the sound data determination means 3 determines whether or not the reproduction time of the sound data SD is equal to or longer than a predetermined time (step S3).
  • the sound data duplicating means 4 duplicates the sound data SD (step S4) and pastes the duplicate data CD on the sound data SD to obtain continuous sound.
  • Data SD ′ is generated.
  • the execution interval setting means 6 performs the reproduction time of the sound data SD, SD ′. Is set to the execution interval TI of the sound data analysis means 5 (step S6).
  • the sound data analysis means 5 repeats FFT based on the set execution interval TI, and performs frequency spectrum analysis of the sound data SD and SD ′ (step S7).
  • the sound data analysis means 5 determines whether or not the sound data SD, SD ′ has ended (step S8). If it is determined that the sound data SD, SD ′ has ended, the analysis result is output to the key determination means. .
  • the key determination means 7 performs key determination on the sound data SD and SD ′ based on the analysis result (step S9).
  • the key determination means 7 outputs the sound data SD and SD ′ keys as determination results on the display screen of the computer or portable terminal (step S10).
  • the music analysis apparatus 1 includes the sound data duplicating means 4, even if the sound data SD is very short, the sound data SD ′ can be made longer than a predetermined time by duplication. Regardless of the time, the sound data analysis means 5 can perform FFT and perform frequency spectrum analysis to perform key determination of the sound data SD and SD ′. As a result, a variety of sound data SD can be used in a DJ-related device regardless of the length of playback time of the sound data SD, so that a DJ with high performance can be provided.
  • the frequency analysis apparatus 1 includes the reproduction time detection unit 2 and the execution interval setting unit 6, the FFT execution interval TI by the sound data analysis unit 5 can be changed according to the reproduction time of the sound data SD. . Therefore, in the case of the sound data SD having a short reproduction time, the analysis interval of the sound data SD can be improved by shortening the execution interval TI and increasing the number of executions of FFT. On the other hand, in the case of sound data SD having a long reproduction time, the analysis time can be shortened by increasing the execution interval TI of the execution interval TI and reducing the number of FFT executions during reproduction of the sound data SD. In the case of long sound data SD, the number of executions of FFT tends to be relatively small and the analysis tends to be rough. However, since the number of executions sufficient for key determination and the like can be obtained, satisfactory results can be obtained. Can be obtained.
  • the music analysis apparatus 1 performs key determination of the sound data SD, but the present invention is not limited to this, and may be used as a music analysis apparatus that performs key and scale determination.
  • the execution interval setting unit 6 sets the execution interval TI based on the reproduction time of the sound data SD and SD ′.
  • the execution interval may be set based on the data length of the input sound data. In addition, other configurations may be adopted as long as the object of the present invention can be achieved.

Abstract

A musical-piece analysis device (1) is equipped with: a sound data reproduction means (4) for reproducing inputted sound data so as to continue for a prescribed playback interval or longer; and a sound data analysis means (5) for analyzing the sound data spanning the prescribed playback interval, and also analyzing the inputted sound data.

Description

楽曲解析装置、楽曲解析方法、及び楽曲解析プログラムMusic analysis apparatus, music analysis method, and music analysis program
　本発明は、楽曲解析装置、楽曲解析方法、及び楽曲解析プログラムに関する。 The present invention relates to a music analysis device, a music analysis method, and a music analysis program.
　従来、楽曲等の音データから自動で拍やテンポ、楽曲のキー、スケール等の楽曲を解析する技術が知られている（例えば、特許文献１参照）。
　音データの解析は、ある程度の長さの音データをサンプリングし、入力された波形を、例えば、ＦＦＴ（Fast Fourier Transform）等を用いて解析することにより行われる。
　このような音データの解析は、ＤＪ関連機器の分野においても、ＢＰＭ（Beats Per Minute）楽曲に関連する技術として利用されている。
　ＤＪ関連機器でこのような音データの解析情報を利用すれば、解析されたテンポ、キー、スケール等を用いることにより、曲と曲との間を違和感なくつなぐことができるため、パフォーマンスの高いＤＪを提供することができる。
　ところで、ＤＪ関連機器では、短時間の音データを効果音やＬｏｏｐ音源として利用することがある。 2. Description of the Related Art Conventionally, a technique for automatically analyzing music such as beats, tempos, music keys, and scales from sound data such as music is known (see, for example, Patent Document 1).
The sound data is analyzed by sampling sound data of a certain length and analyzing the input waveform using, for example, FFT (Fast Fourier Transform).
Such analysis of sound data is also used as a technology related to BPM (Beats Per Minute) music in the field of DJ-related equipment.
By using such sound data analysis information on DJ-related equipment, it is possible to connect songs with a sense of incongruity by using the analyzed tempo, key, scale, etc. Can be provided.
By the way, a DJ-related device may use short-time sound data as a sound effect or a Loop sound source.
特開２０１０－９７０８４号公報JP 2010-97084 A
　しかしながら、非常に短時間の音データを解析する場合、音データのサンプリング時間が短いので、ＦＦＴ解析等音データの解析を行うことができないという課題がある。 However, when analyzing sound data for a very short time, there is a problem that the sound data cannot be analyzed such as FFT analysis because the sampling time of the sound data is short.
　本発明の目的は、非常に短時間の音データであっても正確に解析を行うことのできる楽曲解析装置、楽曲解析方法、及び楽曲解析プログラムを提供することにある。 An object of the present invention is to provide a music analysis apparatus, a music analysis method, and a music analysis program that can accurately analyze even very short time sound data.
　本発明に係る楽曲解析装置は、
　入力された音データを複製し、所定の再生時間以上となるように連続させる音データ複製手段と、
　所定の再生時間以上となった音データに対して解析を行い、入力された音データの解析を行う音データ解析手段と、
を備えていることを特徴とする。 The music analysis apparatus according to the present invention is
Sound data duplicating means for duplicating the input sound data and making it continue so as to be longer than a predetermined reproduction time;
A sound data analysis means for analyzing the sound data that has reached a predetermined reproduction time and analyzing the input sound data;
It is characterized by having.
　本発明に係る楽曲解析方法は、
　入力された音データを複製する手順と、
　複製された音データを所定の再生時間以上となるように連続させる手順と、
　所定の再生時間以上となった前記音データに対して解析を行い、入力された音データの解析を行う手順と、
を実行することを特徴とする。 The music analysis method according to the present invention includes:
The procedure to duplicate the input sound data,
A procedure for continuing the duplicated sound data so as to be longer than a predetermined reproduction time,
Analyzing the sound data that has reached a predetermined playback time and analyzing the input sound data;
It is characterized by performing.
　本発明に係る楽曲解析プログラムは、
　入力された音データを複製する手順と、
　複製された音データを所定の再生時間以上となるように連続させる手順と、
　所定の再生時間以上となった前記音データに対して解析を行い、入力された音データの解析を行う手順と、
をコンピュータに実行させることを特徴とする。 The music analysis program according to the present invention is:
The procedure to duplicate the input sound data,
A procedure for continuing the duplicated sound data so as to be longer than a predetermined reproduction time,
Analyzing the sound data that has reached a predetermined playback time and analyzing the input sound data;
Is executed by a computer.
本発明の実施形態に係る楽曲解析装置を示すブロック図。The block diagram which shows the music analysis apparatus which concerns on embodiment of this invention. 前記実施形態における音データの複製を説明するための模式図。The schematic diagram for demonstrating duplication of the sound data in the said embodiment. 前記実施形態における窓関数を説明するための模式図。The schematic diagram for demonstrating the window function in the said embodiment. 前記実施形態におけるＦＦＴの実行間隔を説明するための模式図。The schematic diagram for demonstrating the execution interval of FFT in the said embodiment. 前記実施形態における再生時間の長い音データの実行間隔を説明するための模式図。The schematic diagram for demonstrating the execution interval of the sound data with a long reproduction time in the said embodiment. 前記実施形態における再生時間の短い音データの実行間隔を説明するための模式図。The schematic diagram for demonstrating the execution space | interval of sound data with a short reproduction time in the said embodiment. 前記実施形態におけるＦＦＴ実行後のキー判定を説明するための模式図。The schematic diagram for demonstrating the key determination after FFT execution in the said embodiment. 前記実施形態におけるＦＦＴ実行後のキー判定を説明するための模式図。The schematic diagram for demonstrating the key determination after FFT execution in the said embodiment. 前記実施形態における楽曲解析方法を説明するためのフローチャート。The flowchart for demonstrating the music analysis method in the said embodiment.
　以下、本発明の実施の一形態について説明する。
　図１には、本実施形態に係る楽曲解析装置１が示されている。この楽曲解析装置１は、入力されたＰＣＭデータ等のデジタル化された音データＳＤを解析し、その音データＳＤのキーを判定して、表示装置等の表示画面上に入力された音データのキー表示ＫＤとして表示させる装置である。
　楽曲解析装置１は、一般的なコンピュータ、ＯＳ（Operation System）を搭載した携帯情報端末で起動するソフトウェアアプリケーションとして構成され、再生時間検出手段２、音データ判定手段３、音データ複製手段４、音データ解析手段５、実行間隔設定手段６、及びキー判定手段７を備える。 Hereinafter, an embodiment of the present invention will be described.
FIG. 1 shows a music analysis apparatus 1 according to this embodiment. The music analysis device 1 analyzes digitized sound data SD such as input PCM data, determines a key of the sound data SD, and determines the sound data input on the display screen of a display device or the like. This is a device for displaying as a key display KD.
The music analysis apparatus 1 is configured as a software application that is activated by a general computer, a portable information terminal equipped with an OS (Operation System), and includes a reproduction time detection means 2, a sound data determination means 3, a sound data duplication means 4, and a sound. Data analysis means 5, execution interval setting means 6, and key determination means 7 are provided.
　再生時間検出手段２は、入力された音データＳＤの再生時間を検出する。具体的には、再生時間検出手段２は、音データＳＤの始めから最後までのサンプリングの回数をカウントすることにより、再生時間を検出する。再生時間を検出したら、再生時間検出手段２は、音データ判定手段３及び実行間隔設定手段６に、検出した再生時間を出力する。 The playback time detection means 2 detects the playback time of the input sound data SD. Specifically, the reproduction time detection means 2 detects the reproduction time by counting the number of samplings from the beginning to the end of the sound data SD. When the reproduction time is detected, the reproduction time detection unit 2 outputs the detected reproduction time to the sound data determination unit 3 and the execution interval setting unit 6.
　音データ判定手段３は、再生時間検出手段２で検出された再生時間に基づいて、入力された音データＳＤが所定の再生時間以上を有するか否かを判定する。具体的には、音データＳＤが、後述する音データ解析手段５で解析できるだけの再生時間を有するか否かを判定する。
　解析の可否の判定は、音データ解析手段５に適用される窓関数に必要な最短の時間以上の時間を、音データＳＤの再生時間が有しているか否かで判定される。 The sound data determination means 3 determines whether or not the input sound data SD has a predetermined reproduction time or more based on the reproduction time detected by the reproduction time detection means 2. Specifically, it is determined whether or not the sound data SD has a reproduction time that can be analyzed by sound data analysis means 5 described later.
Whether or not analysis is possible is determined by whether or not the reproduction time of the sound data SD has a time longer than the shortest time necessary for the window function applied to the sound data analysis means 5.
　窓関数の時間窓長は、音データＳＤのサンプリング周波数、検出する下限周波数、周波数分解能により決定される。
　例えば、ＢＰＭ２００の通常の４／４拍子の楽曲の場合、１拍は３００ｍｓｅｃとなり、１６分音符は７５ｍｓｅｃとなる。ＦＦＴを実行して、音階のＡ０にあたる２７．５Ｈｚの低音の音データＳＤの解析を行う場合、最低でも１．２ｓのデータが必要である。
　音データ判定手段３は、音データＳＤが所定の再生時間未満であると判定したら、その旨の判定結果を、音データ複製手段４に出力する。 The time window length of the window function is determined by the sampling frequency of the sound data SD, the lower limit frequency to be detected, and the frequency resolution.
For example, in the case of a regular 4/4 time tune of BPM200, one beat is 300 msec and a sixteenth note is 75 msec. When performing FFT and analyzing the sound data SD of the 27.5 Hz bass corresponding to A0 of the scale, data of 1.2 s at least is required.
When the sound data determination unit 3 determines that the sound data SD is less than the predetermined reproduction time, the sound data determination unit 3 outputs a determination result to that effect to the sound data duplication unit 4.
　音データ複製手段４は、音データ判定手段３の判定結果に基づいて、入力された音データＳＤを複製し、前述した所定の時間以上となるように連続させる。具体的には、音データ複製手段４は、図２に示されるように、入力された再生時間ｔ１の音データＳＤを複製し、複製データＣＤを、音データＳＤの後ろに貼り付けて連続させ、再生時間ｔ２の連続した音データＳＤ’を生成する。 The sound data duplicating unit 4 duplicates the input sound data SD based on the determination result of the sound data determining unit 3, and continues the data so that the predetermined time or more is reached. Specifically, as shown in FIG. 2, the sound data duplicating unit 4 duplicates the input sound data SD of the reproduction time t1, and pastes the duplicate data CD behind the sound data SD so as to be continuous. , Continuous sound data SD ′ having a reproduction time t2 is generated.
　例えば、音データＳＤの再生時間ｔ１が前述した例における１．２ｓｅｃより短い場合、音データ複製手段４は、音データＳＤの複製データＣＤを複製して、再生時間ｔ２が１．２ｓｅｃ以上の連続した音データＳＤ’となるまで複製を繰り返す。
　なお、複製は、再生時間ｔ２が音データ解析手段５で解析できるだけの再生時間であればよく、複製の回数Ｎは整数でなくてもよい。
　音データ複製手段４は、複製により、所定の時間以上とされた音データＳＤ’を音データ解析手段５に出力する。 For example, when the reproduction time t1 of the sound data SD is shorter than 1.2 seconds in the above-described example, the sound data duplicating unit 4 duplicates the duplicate data CD of the sound data SD, and the reproduction time t2 is 1.2 seconds or longer. The duplication is repeated until the sound data SD ′ is obtained.
It should be noted that the duplication only needs to be performed so that the reproduction time t2 can be analyzed by the sound data analysis means 5, and the number N of duplications may not be an integer.
The sound data replicating unit 4 outputs to the sound data analyzing unit 5 the sound data SD ′ that has been duplicated for a predetermined time or more.
　音データ解析手段５は、音データＳＤ、ＳＤ’の周波数スペクトル解析を行う。解析方法は、本実施形態では、ＦＦＴを用いて解析で行っている。但し、解析方法は、特に限定されるわけでなく、例えば、ＤＣＴ（Discrete Cosine transform）を用いた解析や、時間軸による解析や、信号レベルによる解析、音量感、アタック感による解析を用いることができる。
　ＦＦＴには、図３に示すように、通常、窓関数であるハミング窓ＨＭＷが適用されている。ハミング窓ＨＭＷは、ＦＦＴの実行時間中の時間軸上の両端の信号強度の増幅を弱め、ＦＦＴ実行時に、サンプリングされた波形の継ぎ目の不連続部分の影響を少なくするために適用されている。 The sound data analysis means 5 performs frequency spectrum analysis of the sound data SD and SD ′. In this embodiment, the analysis method is performed by analysis using FFT. However, the analysis method is not particularly limited, and for example, analysis using DCT (Discrete Cosine transform), analysis using a time axis, analysis using a signal level, volume feeling, and attack feeling may be used. it can.
As shown in FIG. 3, a Hamming window HMW that is a window function is usually applied to the FFT. The Hamming window HMW is applied to weaken the signal intensity amplification at both ends on the time axis during the FFT execution time and reduce the influence of the discontinuous portion of the sampled waveform seam during the FFT execution.
　このため、ＦＦＴを実行したデータの時間軸上の両端部は、信号強度が弱すぎるので、解析データとして使用することができない。
　そこで、本実施形態では、解析可能なデータとして、ＦＦＴ実行時間のうち信号強度が弱められない解析期間Ｔ０の信号強度を用いて、周波数スペクトル解析を行っている。解析期間Ｔ０は、任意に定めうるが、本実施形態では、ハミング窓ＨＭＷの０．７（７０％）値で設定している。
　なお、本実施形態では、ハミング窓ＨＭＷを適用しているが、これに限らず、ハニング窓、フラップトップ窓等を窓関数として用いてもよい。 For this reason, since the signal intensity is too weak at both ends on the time axis of the data on which FFT has been executed, it cannot be used as analysis data.
Therefore, in the present embodiment, frequency spectrum analysis is performed using the signal intensity of the analysis period T0 during which the signal intensity is not weakened in the FFT execution time as the data that can be analyzed. Although the analysis period T0 can be arbitrarily determined, in this embodiment, the analysis period T0 is set to a 0.7 (70%) value of the Hamming window HMW.
In the present embodiment, the Hamming window HMW is applied. However, the present invention is not limited to this, and a Hanning window, a flap top window, or the like may be used as the window function.
　実行間隔設定手段６は、再生時間検出手段２で検出された再生時間に基づいて、音データ解析手段５によるＦＦＴの実行間隔を設定する。
　具体的には、実行間隔設定手段６は、図４に示されるように、１回目にＦＦＴ１が実行された後、２回目のＦＦＴ２の実行開始の時間を、実行間隔ＴＩとして設定する。以後、本実施形態では、３回目のＦＦＴ３は、時間２ＴＩ後に開始され、順次ＦＦＴが実行される。 The execution interval setting means 6 sets the FFT execution interval by the sound data analysis means 5 based on the reproduction time detected by the reproduction time detection means 2.
Specifically, as shown in FIG. 4, the execution interval setting means 6 sets the execution start time of the second FFT2 as the execution interval TI after the first execution of FFT1. Thereafter, in the present embodiment, the third FFT3 is started after time 2TI, and the FFTs are sequentially executed.
　実行間隔ＴＩの設定は、音データＳＤ、ＳＤ’の再生時間によって設定される。
　実行間隔設定手段６は、例えば、再生時間が３０ｓｅｃ以上の長い音データＳＤの場合、図５に示されるように、実行間隔ＴＩを大きくし、再生時間が３０ｓ未満の短い音データＳＤの場合、図６に示されるように、実行間隔ＴＩを小さくする。実行間隔ＴＩの下限値は、ＦＦＴ１、ＦＦＴ２・・・の解析期間Ｔ０が連続するような実行間隔ＴＩとする。
　実行間隔設定手段６は、設定した実行間隔ＴＩを前述した音データ解析手段５に出力する。
　音データ解析手段５は、実行間隔ＴＩに基づいて、ＦＦＴを繰り返し実行し、ＦＦＴの実行の度に解析結果をキー判定手段７に出力する。 The execution interval TI is set according to the reproduction time of the sound data SD and SD ′.
For example, in the case of long sound data SD with a reproduction time of 30 seconds or more, the execution interval setting means 6 increases the execution interval TI as shown in FIG. 5, and in the case of short sound data SD with a reproduction time of less than 30 s, As shown in FIG. 6, the execution interval TI is reduced. The lower limit value of the execution interval TI is set to an execution interval TI such that the analysis periods T0 of FFT1, FFT2,.
The execution interval setting means 6 outputs the set execution interval TI to the sound data analysis means 5 described above.
The sound data analysis unit 5 repeatedly executes FFT based on the execution interval TI, and outputs the analysis result to the key determination unit 7 every time FFT is executed.
　キー判定手段７は、音データ解析手段５から出力された解析結果に基づいて、音データＳＤ、ＳＤ’のキー判定を行う。
　具体的には、キー判定手段７は、オクターブあたり１２の音階それぞれについて、各音階をルートとするマイナーキー及びメジャーキーの２種類、計２４種類のキーの基準周波数を記憶している。
　キー判定手段７は、実行間隔ＴＩ毎に入力される解析結果を時間軸方向に合算し、得られた合算値から、信号強度の強い周波数に近い基準周波数を選択し、図７に示されるように、各音階の信号強度を求める。
　次に、キー判定手段７は、図８に示されるように、信号強度の強い順番に並べ替え、正規化し、信号強度の強いいくつかの音階を選択して、音データＳＤ、ＳＤ’のキーを判定する。
　キー判定手段７は、音データＳＤ、ＳＤ’のキー判定結果をキー表示ＫＤとして、コンピュータのディスプレイや、携帯端末の画面上に表示する。 The key determination unit 7 performs key determination of the sound data SD and SD ′ based on the analysis result output from the sound data analysis unit 5.
Specifically, the key determination means 7 stores the reference frequencies of a total of 24 types of keys, that is, two types of minor keys and major keys rooted in each scale for each of the 12 scales per octave.
The key determination means 7 adds the analysis results input at every execution interval TI in the time axis direction, and selects a reference frequency close to a frequency with a strong signal strength from the obtained sum, as shown in FIG. Next, the signal strength of each scale is obtained.
Next, as shown in FIG. 8, the key determination means 7 rearranges and normalizes the signals in the order of strong signal strength, selects several scales with strong signal strength, and selects the keys of the sound data SD, SD ′. Determine.
The key determination means 7 displays the key determination result of the sound data SD, SD ′ as a key display KD on the computer display or the screen of the portable terminal.
　次に、前述した構成を備えた楽曲解析装置１による音データＳＤのキー判定を、図９に示されるフローチャートに基づいて、説明する。
　まず、コンピュータや携帯端末のユーザーが、楽曲解析装置１を画面上で選択し、プログラムを起動し、解析対象である音データＳＤを選択すると、楽曲解析装置１に音データＳＤが入力される（手順Ｓ１）。
　音データＳＤが入力されたら、再生時間検出手段２は、音データＳＤの再生時間を検出する（手順Ｓ２）。
　音データ判定手段３は、音データＳＤの再生時間が所定の時間以上であるかどうかを判定する（手順Ｓ３）。 Next, key determination of the sound data SD by the music analysis apparatus 1 having the above-described configuration will be described based on the flowchart shown in FIG.
First, when a user of a computer or a portable terminal selects the music analysis device 1 on the screen, starts the program, and selects the sound data SD to be analyzed, the sound data SD is input to the music analysis device 1 ( Procedure S1).
When the sound data SD is input, the reproduction time detecting means 2 detects the reproduction time of the sound data SD (step S2).
The sound data determination means 3 determines whether or not the reproduction time of the sound data SD is equal to or longer than a predetermined time (step S3).
　音データＳＤの再生時間が所定の時間未満であると判定されたら、音データ複製手段４は、音データＳＤを複製し（手順Ｓ４）複製データＣＤを音データＳＤに貼り付けて、連続した音データＳＤ’を生成する。
　はじめから音データＳＤの再生時間が所定の時間以上であるか、音データＳＤ’の再生時間が所定の時間以上となった場合、実行間隔設定手段６は、音データＳＤ、ＳＤ’の再生時間に基づいて、音データ解析手段５の実行間隔ＴＩを設定する（手順Ｓ６）。 If it is determined that the reproduction time of the sound data SD is less than the predetermined time, the sound data duplicating means 4 duplicates the sound data SD (step S4) and pastes the duplicate data CD on the sound data SD to obtain continuous sound. Data SD ′ is generated.
When the reproduction time of the sound data SD from the beginning is equal to or longer than the predetermined time or when the reproduction time of the sound data SD ′ is equal to or longer than the predetermined time, the execution interval setting means 6 performs the reproduction time of the sound data SD, SD ′. Is set to the execution interval TI of the sound data analysis means 5 (step S6).
　音データ解析手段５は、設定された実行間隔ＴＩに基づいて、ＦＦＴを繰り返し、音データＳＤ、ＳＤ’の周波数スペクトル解析を行う（手順Ｓ７）。
　音データ解析手段５は、音データＳＤ、ＳＤ’が終了したか否かを判定し（手順Ｓ８）、音データＳＤ、ＳＤ’が終了したと判定されたら、解析結果をキー判定手段に出力する。
　キー判定手段７は、解析結果に基づいて、音データＳＤ、ＳＤ’のキー判定を行う（手順Ｓ９）。
　キー判定手段７は、判定結果である音データＳＤ、ＳＤ’のキーを、コンピュータや携帯端末の表示画面上に出力する（手順Ｓ１０）。 The sound data analysis means 5 repeats FFT based on the set execution interval TI, and performs frequency spectrum analysis of the sound data SD and SD ′ (step S7).
The sound data analysis means 5 determines whether or not the sound data SD, SD ′ has ended (step S8). If it is determined that the sound data SD, SD ′ has ended, the analysis result is output to the key determination means. .
The key determination means 7 performs key determination on the sound data SD and SD ′ based on the analysis result (step S9).
The key determination means 7 outputs the sound data SD and SD ′ keys as determination results on the display screen of the computer or portable terminal (step S10).
　このような本実施形態によれば、以下のような効果がある。
　楽曲解析装置１が音データ複製手段４を備えていることにより、非常に短い音データＳＤであっても、複製により所定時間以上の音データＳＤ’とすることができるため、音データＳＤの再生時間によらず、音データ解析手段５により、ＦＦＴを実行して周波数スペクトル解析を行って、音データＳＤ、ＳＤ’のキー判定を行うことができる。
　これにより、音データＳＤの再生時間の長短によらず、ＤＪ関連機器で様々な音データＳＤを利用できるので、パフォーマンスの高いＤＪを提供することができる。 According to this embodiment, there are the following effects.
Since the music analysis apparatus 1 includes the sound data duplicating means 4, even if the sound data SD is very short, the sound data SD ′ can be made longer than a predetermined time by duplication. Regardless of the time, the sound data analysis means 5 can perform FFT and perform frequency spectrum analysis to perform key determination of the sound data SD and SD ′.
As a result, a variety of sound data SD can be used in a DJ-related device regardless of the length of playback time of the sound data SD, so that a DJ with high performance can be provided.
　楽曲解析装置１が再生時間検出手段２及び実行間隔設定手段６を備えていることにより、音データＳＤの再生時間に応じて、音データ解析手段５によるＦＦＴの実行間隔ＴＩを変更することができる。従って、再生時間の短い音データＳＤの場合は、実行間隔ＴＩを短くして、ＦＦＴの実行回数を増やして、音データＳＤの解析精度を向上させることができる。
　一方、再生時間の長い音データＳＤの場合、実行間隔ＴＩの実行間隔ＴＩを長くして、音データＳＤの再生中のＦＦＴの実行回数を減らして、解析時間を短縮することができる。なお、長い音データＳＤの場合、ＦＦＴの実行回数が相対的に少なくなり、解析が粗くなる傾向があるが、キー判定等の用途には充分な実行回数が得られるため、支障なく良好な結果を得ることができる。 Since the music analysis apparatus 1 includes the reproduction time detection unit 2 and the execution interval setting unit 6, the FFT execution interval TI by the sound data analysis unit 5 can be changed according to the reproduction time of the sound data SD. . Therefore, in the case of the sound data SD having a short reproduction time, the analysis interval of the sound data SD can be improved by shortening the execution interval TI and increasing the number of executions of FFT.
On the other hand, in the case of sound data SD having a long reproduction time, the analysis time can be shortened by increasing the execution interval TI of the execution interval TI and reducing the number of FFT executions during reproduction of the sound data SD. In the case of long sound data SD, the number of executions of FFT tends to be relatively small and the analysis tends to be rough. However, since the number of executions sufficient for key determination and the like can be obtained, satisfactory results can be obtained. Can be obtained.
　なお、本発明は、前述の実施形態に限定されるものではなく、以下に示すような変形をも含むものである。
　前述した実施形態では、楽曲解析装置１は、音データＳＤのキー判定を行うものであったが、本発明はこれに限らず、キー及びスケール判定を行う楽曲解析装置として利用してもよい。
　また、前述した実施形態では、実行間隔設定手段６は、音データＳＤ、ＳＤ’の再生時間に基づいて、実行間隔ＴＩを設定していたが、本発明はこれに限られない。入力される音データのデータ長に基づいて、実行間隔を設定してもよい。
　その他、本発明の目的を達成できる範囲で他の構成を採用してもよい。 In addition, this invention is not limited to the above-mentioned embodiment, The modification as shown below is also included.
In the embodiment described above, the music analysis apparatus 1 performs key determination of the sound data SD, but the present invention is not limited to this, and may be used as a music analysis apparatus that performs key and scale determination.
In the above-described embodiment, the execution interval setting unit 6 sets the execution interval TI based on the reproduction time of the sound data SD and SD ′. However, the present invention is not limited to this. The execution interval may be set based on the data length of the input sound data.
In addition, other configurations may be adopted as long as the object of the present invention can be achieved.
　１…楽曲解析装置、２…再生時間検出手段、３…音データ判定手段、４…音データ複製手段、５…音データ解析手段、６…実行間隔設定手段、７…キー判定手段、ＣＤ…複製データ、ＨＭＷ…ハミング窓、ＫＤ…キー表示、Ｓ１…手順、Ｓ２…手順、Ｓ３…手順、Ｓ４…手順、Ｓ６…手順、Ｓ７…手順、Ｓ８…手順、Ｓ９…手順、Ｓ１０…手順、ＳＤ…音データ、Ｔ０…解析期間、ｔ１…再生時間、ｔ２…再生時間、ＴＩ…実行間隔
  DESCRIPTION OF SYMBOLS 1 ... Music analysis apparatus, 2 ... Reproduction time detection means, 3 ... Sound data determination means, 4 ... Sound data duplication means, 5 ... Sound data analysis means, 6 ... Execution interval setting means, 7 ... Key judgment means, CD ... Duplication Data, HMW ... Hamming window, KD ... Key display, S1 ... Procedure, S2 ... Procedure, S3 ... Procedure, S4 ... Procedure, S6 ... Procedure, S7 ... Procedure, S8 ... Procedure, S9 ... Procedure, S10 ... Procedure, SD ... Sound data, T0 ... analysis period, t1 ... reproduction time, t2 ... reproduction time, TI ... execution interval

Claims (5)

  1. 　入力された音データを複製し、所定の再生時間以上となるように連続させる音データ複製手段と、
    　所定の再生時間となった音データに対して解析を行い、入力された音データの解析を行う音データ解析手段と、
    　を備えていることを特徴とする楽曲解析装置。     Sound data duplicating means for duplicating the input sound data and making it continue so as to be longer than a predetermined reproduction time;
    A sound data analysis means for analyzing the sound data having a predetermined playback time and analyzing the input sound data;
    A music analysis apparatus comprising:
  2. 　請求項１に記載の楽曲解析装置において、
    　入力された音データの再生時間の長さが、所定の再生時間以上であるか否かを判定する音データ判定手段を備え、
    　前記音データ複製手段は、前記音データ判定手段により、入力された音データが前記所定の再生時間未満であると判定されたときに、入力された音データを複製することを特徴とする楽曲解析装置。     The music analysis apparatus according to claim 1,
    Sound data determination means for determining whether or not the length of the reproduction time of the input sound data is equal to or longer than a predetermined reproduction time;
    The sound data duplicating means duplicates the input sound data when the sound data judging means judges that the inputted sound data is less than the predetermined reproduction time. apparatus.
  3. 　請求項１又は請求項２に記載の楽曲解析装置において、
    　前記音データ複製手段は、入力された音データのすべてを複製することを特徴とする楽曲解析装置。     In the music analysis device according to claim 1 or 2,
    The sound data duplicating unit duplicates all of the input sound data.
  4. 　入力された音データを複製する手順と、
    　複製された音データを所定の再生時間以上となるように連続させる手順と、
    　所定の再生時間以上となった音データに対して解析を行い、入力された音データの解析を行う手順と
    を実行することを特徴とする楽曲解析方法。     The procedure to duplicate the input sound data,
    A procedure for continuing the duplicated sound data so as to be longer than a predetermined reproduction time,
    A music analysis method comprising: performing analysis on sound data having a predetermined reproduction time or longer and performing analysis of input sound data.
  5. 　入力された音データを複製する手順と、
    　複製された音データを所定の再生時間となるように連続させる手順と、
    　所定の再生時間以上となった前記音データに対して解析を行い、入力された音データの解析を行う手順と、
    をコンピュータに実行させることを特徴とする楽曲解析プログラム。
          The procedure to duplicate the input sound data,
    A procedure for continuing the replicated sound data so as to have a predetermined playback time;
    Analyzing the sound data that has reached a predetermined playback time and analyzing the input sound data;
    The music analysis program which makes a computer execute.
PCT/JP2015/075076 2015-09-03 2015-09-03 Musical-piece analysis device, musical-piece analysis method, and musical-piece analysis program WO2017037919A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/JP2015/075076 WO2017037919A1 (en) 2015-09-03 2015-09-03 Musical-piece analysis device, musical-piece analysis method, and musical-piece analysis program
JP2017537158A JPWO2017037919A1 (en) 2015-09-03 2015-09-03 Music analysis apparatus, music analysis method, and music analysis program

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/JP2015/075076 WO2017037919A1 (en) 2015-09-03 2015-09-03 Musical-piece analysis device, musical-piece analysis method, and musical-piece analysis program

Publications (1)

Publication Number Publication Date
WO2017037919A1 true WO2017037919A1 (en) 2017-03-09

Family

ID=58188510

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2015/075076 WO2017037919A1 (en) 2015-09-03 2015-09-03 Musical-piece analysis device, musical-piece analysis method, and musical-piece analysis program

Country Status (2)

Country Link
JP (1) JPWO2017037919A1 (en)
WO (1) WO2017037919A1 (en)

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006106535A (en) * 2004-10-08 2006-04-20 Nippon Telegr & Teleph Corp <Ntt> Sound signal storage and retrieval device and program

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006106535A (en) * 2004-10-08 2006-04-20 Nippon Telegr & Teleph Corp <Ntt> Sound signal storage and retrieval device and program

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
"Kosoku Fourier Henkan Excel VBA-yo FFT Program", 29 March 2007 (2007-03-29), Retrieved from the Internet <URL:http://tsuyu.cocolog-nifty.com/ blog/2007/03/publi.html> [retrieved on 20151008] *

Also Published As

Publication number Publication date
JPWO2017037919A1 (en) 2018-07-12

Similar Documents

Publication Publication Date Title
US8901406B1 (en) Selecting audio samples based on excitation state
JP5229998B2 (en) Code name detection device and code name detection program
CN108292499A (en) Skill determining device and recording medium
WO2019043797A1 (en) Song analysis device and song analysis program
JP2015079151A (en) Music discrimination device, discrimination method of music discrimination device, and program
JP5395399B2 (en) Mobile terminal, beat position estimating method and beat position estimating program
US20050217461A1 (en) Method for music analysis
JP6920445B2 (en) Music analysis device and music analysis program
WO2017037919A1 (en) Musical-piece analysis device, musical-piece analysis method, and musical-piece analysis program
Pang et al. Automatic detection of vibrato in monophonic music
WO2017037920A1 (en) Musical-piece analysis device, musical-piece analysis method, and musical-piece analysis program
JP6263383B2 (en) Audio signal processing apparatus, audio signal processing apparatus control method, and program
JP6263382B2 (en) Audio signal processing apparatus, audio signal processing apparatus control method, and program
JP6286933B2 (en) Apparatus, method, and program for estimating measure interval and extracting feature amount for the estimation
JP2007272118A (en) Method, device, and program for extracting musical piece tempo
JP6946442B2 (en) Music analysis device and music analysis program
JP6168649B2 (en) Code detection apparatus and program
WO2024034115A1 (en) Audio signal processing device, audio signal processing method, and program
WO2024034118A1 (en) Audio signal processing device, audio signal processing method, and program
JP6847242B2 (en) Music analysis device and music analysis program
JP6854350B2 (en) Music analysis device and music analysis program
JP7176114B2 (en) MUSIC ANALYSIS DEVICE, PROGRAM AND MUSIC ANALYSIS METHOD
EP2043089B1 (en) Method and device for humanizing music sequences
Brent A perceptually based onset detector for real-time and offline audio parsing
WO2020255213A1 (en) Musical composition structural analysis device and musical composition structural analysis program

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15903038

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2017537158

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 15903038

Country of ref document: EP

Kind code of ref document: A1