JPH04163486A - Regulating discriminating device - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION Technical Field of the Invention The present invention relates to a tonality determining device that automatically determines the key (tonality) of a given piece of music.

<Background of the Invention> Conventionally, various tonality determination devices for determining tS<tonality) of an input musical piece have been studied.

When these tonality discrimination devices become high-grade, they require the knowledge and know-how of experts to create algorithms for tonality discrimination. Alternatively, if you try to improve the accuracy of discrimination, you will end up having to limit the types of tonality that can be judged.

For example, even if you can distinguish between minor and major, you will not be able to distinguish between other keys, such as Okinawa key or blues.

<Purpose of the Invention> The present invention has been made in view of the above points, and an object of the present invention is to provide a tonality discrimination device that has a relatively simple configuration and is capable of discriminating various keys (tonality). <Structure and operation of the invention> That is, according to the present invention, in order to achieve the above object, a music information supplying means for supplying music information expressing a music;
membership function supply means for supplying a membership function representing the frequency of appearance of each scale for a plurality of keys;
The music information provided from the music information supply means;
similarity calculation means for calculating the degree of similarity with the membership function for a plurality of keys supplied from the membership function supply means; A tonality discrimination device is provided, comprising: output means for outputting information representing the key of a piece of music expressed by piece information.

FIG. 1 shows an example of this configuration, and the music information supply means GS includes, for example, a keyboard, or is connected to the keyboard so that music information can be input from the keyboard. This music information is expressed by pitch and duration.

The membership function supply means MS includes, for example, a storage means that stores in advance membership functions representing the frequency of appearance of each scale for a plurality of tones.

The stored contents of this storage means are based on the results of analyzing a plurality of songs.

The similarity calculation means CA calculates the degree of similarity between the music information from the music information supply means GS and the membership functions for a plurality of keys from the membership function supply means MS. By calculating this degree of similarity, it becomes clear what kind of lyrics the music piece has.

The output of the similarity calculation means CA is given to the output means ○U, and the tonality discrimination result is outputted. This output means ○U includes, for example, a display means, and informs the user of the tonality determination result in various display formats.

Therefore, if membership functions for each key are prepared in advance, the tonality of a piece of music can be determined by simply performing a relatively simple calculation.

<Example> This invention will be described in detail below with reference to an example shown in the drawings.

FIG. 2 shows its overall configuration, and the CPU 1 controls the overall operation of this tonality discrimination device.

This CPUI is connected to a ROM 2, a RAM 3, a MIDI interface 4, and an LCD driver 5 via a common bus 6. In addition, a switch section for inputting control information is provided as necessary (as shown).

ROM2 contains programs for CPU operation as well as several other programs! A membership function for l (tonality) is stored. This membership function will be detailed later.

The RAM 3 is used as a work memory etc. when executing various calculations by the CPU. Further, the MIDI interface 4 is an interface to which performance information is given from an external electronic musical instrument EM.

Furthermore, MIDI stands for MUSICAL INSTRUME.
It is an abbreviation for NT DIGITAL INTERFACE. The performance information provided via the MIDI interface 4 is sequentially stored in the RAM 3 as music information.

The CPU inputs the song information input into RAM3 and the ROM.
Based on the membership functions stored in No. 2, a similarity calculation, which will be described later, is executed. Then, according to the similarity calculation result, a signal indicating the tonality of the input music information is given to the LCD driver 5, and the tonality of the music is displayed according to a predetermined display pattern on the LCD 7.

Here, the CPUI constitutes the similarity calculation means CA explained with reference to FIG. 1, the ROM2 constitutes the membership function supply means MS, and the MIDI interface 4 and R
AM3 constitutes the music information supply means GS, and the LCD
The driver 5 and the LCD 7 constitute an output means OU.

Next, the membership functions stored in advance in the ROM 2 will be explained.

Membership function m of Kaoruzu, C major (C power, J)
. I"laJ will be explained. This membership function m., 1 is obtained by preparing a plurality, preferably a large number, of musical scores of songs that the manufacturer recognizes as C major, and extracting their characteristics.

Specifically, for example, a musical score as shown in FIG. 3 is prepared.

This score is in C major.

Then, assuming a unit time of 1t, which is obtained by dividing one measure into 96, for example, the total number of units of time in which the sound 1c appearing in this song appears is calculated.

Here, let this value be T (C). For example, T
(C)=480t.

Similarly, find the total appearance time of the sound 11WC'' that appears in this song. Let this value be T(C''). For example, T (CF2)=o. In the same manner, the total appearance time for 12 sounds 1 (C-B) is calculated. For example, in the case of this song, it would be as follows. In addition, t
is omitted.

T (C) = 480 T　(Cヰ)=O T(D)=192 T(D”)=O T (E) = 336 T (F) = 32 T(F”)=16 T (G) = 608 T (G) = 48 T (A) = 848 T (A”) = 48 T (B) = 432 Here, the scale vector is defined as follows.

Th= (T (C), T (C Nishiki), T (
D), T(D"), T(E), T(E"), T(F
), T (F”), T (G), T (G”)
, T (A), T (eye A), T (B)
) Therefore, in this case, Th= (480,0,192,0,336,32,1
6,608,48,848,48.

Then, it is standardized so that the sum of squares of Th becomes 1, and this is defined as a standard scale vector Th'.

Th”” (T (C)/D, T (C”)/D, T
(D)/D, T (D”)/D, T (E)/D, T
(E”)/D, T (F)/D, T (F”)/D, T
(G)/D, T (GI'I)/D, T (A)/D
, T (A″)/D, T (B)/D)” Here, D=v′I: (T (α))2. E is the sum of the scale α, that is, C−B.

In the above example:

Th'= (0,372,0,0,148°0,0.2
61.0.025゜0.012.0.472.0.037゜0.659.0
．． 037.0.335) For other songs, find Th' using the method explained above. For example, for 100 songs,
By performing the above processing, T h'o'' T h', 9 is obtained.Then, the sum vector ST of the standard scale vectors for the 100 songs is obtained.

5T=Th'o+Th', +・-・-+Th'98+T
h99'= (nTh'+ (c), nTh"i(C"
), ......, nTh': (A”), I:Th
': (B)) E is the sum for song l, and therefore for 100 songs.

As an example, ST is as follows.

ST= (52,8,1,1,15,5,0°22.2
．． 2.4. 1.2.51.9. '3.1.54.5
．． 3.9.35.2) Now standardize this vector.

S T' ” T h'o/D' + T
h't/D'＋・・・・・・＋T h'9
B/D'+Th'oe/D'= (I:
Th'i (C) /D'.

nTh'+(C")/D',...

nTh’t　(A”)/D’.

I: Th'+ (B) / D') Here, D'=Jn(Th'+ (α))2, which is the sum of the scale α, that is, C−B.

In the above example, it takes the following values:

S T'= (0,516,0,011,0,152,
0゜0.217.0.023.0.012.0.508
゜0.030.0.533.0.038.0.344)
Each component of this vector ST' is defined as a membership function m.

, J (TH). That is, the membership function mar+aJ(TH) is a function having the sound [TH as a variable, and the above example is as shown in FIG.

In this way, the root note is C. Membership function m for Msj. When haj(TH) is found, Ct1
The membership function of M8J is m C”M8 J (
T H)" mCMaj (T H1/2Z ). Here, 1/2z represents a semitone.

As will be understood with reference to FIG. 5, for example, C-1
/2Z=C+11/2Z=B. Therefore, the relationship is as follows.

m C"l'laj (C) = m Cff1aj
(B)m C”Maj (C”) = m CFIa
j (C)mo″r+aj (B) =mCriaj
(AI'I) This results in a membership function m. +tka,
j(TH) is as shown in Figure 116.

Similarly, D1. laj~B1. The membership function mQm,j (TH)−mBM,J(
TH) is found. In the above, the membership function of the major system has been found based on the 12 root notes.

Next is the minor membership function m. ,(TH)
~meff, (TH) is obtained by preparing the musical scores of, for example, 100 songs and using the same procedure as above. Examples of mc, , (TH) and me'', n (TH) among these membership functions are shown in FIGS. 7 and 8.

Select, for example, 100 blues songs, ie, major blues and minor blues songs,
By performing the same processing as above, each membership function m CMajBS (TH) -m Br, 1aJ
es ('r H) and m(, nQs(TH) -me
Find +nes (T H). 9 and 10 respectively show m of the above membership functions. m-J
An example of Bs (TH) and mcmss (TH) is shown.

The membership functions for major, minor, major blues, and minor blues have been found above, but it goes without saying that membership functions for other keys that require judgment can also be found in the same manner as above. stomach.

Then, the membership function obtained in advance in this way is
It is stored in ROM2 shown in FIG. As understood from the above explanation, membership functions for roots other than C can be found by shifting the designation of TH of the membership function whose root is C, so R
The following four membership functions whose root is C are stored in OM2.

m effla = (T H) See Figure 4, mc, .
, (TH) See Figure 7, m eras jBs (TH) See Figure 9, m c
+n8s (T H) jl! See Figure 10.

Next, the operation of this embodiment will be explained.

In this embodiment, when a predetermined number of notes are input, key determination is started in real time, and the resulting information regarding the tonality of the piece of music is displayed in graphic ink.

FIG. 11 shows the operation flow of the CPU 1. First, at 100, initialization processing is performed. Thereafter, input/output processing 200 is performed on performance information input via the MIDI interface 4. The next 30
At 0, the presence or absence of the note 0N10FF is detected, and the MID
New Note 0N10F via I interface 4
If information indicating F is input, the key determination process of 400 is executed; otherwise, the key determination process is changed from 300 to 2.
Return to 00.

Specifically, in the key determination process 400, first, in step 401, it is detected whether the note is ON or OFF, and if a new input is OFF.
If N, proceed to 402 and note name of the note ON (note t
ltW)n is detected. Then, in step 403, flag register O(n) is set to 1. This flag register (n) is provided on RAMa. Subsequently, after clearing the ON time measurement counter T (n) in 404, the ON time measurement for this new input scale is started. This ON time measurement counter T(n) is constituted by the CPUI and the RAM 3, and automatically performs time measurement processing when a count start is set. Then, the process returns from 404 to 200.

If OFF is detected in 401, the process proceeds to 405, where the note name (scale) n of the note OFF is detected. In the following step 406, the flag register ○(n) is set to 0, and in 407, the ON time measurement counter T is set.
The counting operation (n) is stopped, and in step 408, the contents of the ON time measurement counter T(n) are added to the total ON time register 5TI(n). This total ON time register STI (n) has elements for each tone, and can be thought of as a vector.

In 409, it is determined whether the number of notes input so far exceeds a predetermined number, that is, register P(OFF)=20, and if P(OFF)<20, in 409,
After incrementing register P (OFF) by 1, 2
Returns to 00.

If YES is determined in 409, the process proceeds to 411, where the value of the register P (OFF) is decreased by 10, and then the process proceeds to 412. In other words, as can be understood from the operations of 409, 410, and 411, in the initial state, after more than 20 notes are input, key determination processing is performed, and after that, 10 new notes are input. Attitude judgment processing will now be performed each time.

412, the total time register STI (n
) Calculate the square root F of the sum of squares of each component.

F=(n (STI (n)) 2In this formula, =
means to calculate the sum for each scale n, that is, C-B.

In the following 413, the total ON time register STI
In order to standardize each element of (n), that is, to obtain a standardized total ON time register STI' (n) (which can also be considered as a vector), each component of the total ON time register STI (n) of,
Divide by F.

After performing this processing (preprocessing), similarity calculation is performed between each membership function stored in the ROM 2 and the standardized total ON time register S TI ' (n). 4140 to 414B are similarity calculations with major membership functions, and 415
C to 415B are the similarity calculations with the minor type membership function, 4160 to 416B are the similarity calculations with the major blues type membership functions, and 4170 to 417B are the similarity calculations with the minor blues type membership functions. This is a similarity calculation with a function.

- To illustrate an example, the membership function m of CMa.

The similarity calculation between ++aJ(n) (n and the above-mentioned TH are the same and refer to each sound qc-B) and each component of the standardized total ON time register STI' (n) is as follows. become that way.

Scr+aJ=nST I' (n) *mCff1
aj (n) In this formula, E is the sum of each sound Win(CB).

Each operation as shown in FIG. 11 is performed using each membership function and the standardized total ON time.
), the operation result is S CMa j
""' S BMa j, S Cm-S B+n+ S
CMaj8S-S 8M5jBs, S CmB5~
12*4=48 values of 58mB5 are found.

After this series of processing, the process returns to 200, and the CPUI sends the similarity calculation result S to the LCD driver 5. Maj～
S BI”la, ++ S C+n~S B+n-S
Cff1aj9S~SgaajBS-S CmB5-S
8mB5 is sent to drive LCD7. Figure 12 is L
One display form of CD7 is shown. That is, LCD7
is displayed in four pie charts, major, minor,
It is possible to show 48 calculation results for each root note of major blues and minor blues. Therefore, the user can visually confirm the key of the piece of music that is currently being input for performance from the display of these four pie charts.

In the example in Figure 12, F
Map is the most likely candidate, and the next candidate is A”
Examples include ff++ CMB 2, G, and n.

An embodiment of the present invention has been described above, and according to this embodiment, m Cm of music input from a keyboard etc.
can be detected and displayed in real time,
Moreover, even if there is a change in key during the song, the display format changes sequentially, which is preferable. In this case, for the modulation, the total ON time register STI' (n)
Even better results can be obtained if the ON time data cumulative values of each scale stored in 1 are stored only for the latest predetermined number of notes, and the moving cumulative values are performed.

Furthermore, in the above embodiment, since the probability of each leg is displayed graphically, the relationship between the legs can be understood, which is convenient for music learning.

In addition, according to the above embodiment, membership functions expressed in semitone units were prepared, but even finer tones @ widths were prepared.
A membership function may also be expressed. For example, 1!1311iW is a CM that is taken in small pitch units so that similarity calculation can be performed even for bend information.
j8S membership function m. h-*ss (TH
). By doing this, it is possible to improve the accuracy of the discrimination results for a specific key such as blues.

Further, in the above embodiment, as described in relation to FIG. 12, the probability that the music corresponds to each leg is displayed in a graph, but it is not necessary to take such a display form; for example, , As a result of calculating the degree of similarity,
Only the name of the key determined to have the highest degree of similarity may be displayed in text. Alternatively, a plurality of key candidates and their probabilities may be displayed numerically.

Further, according to the embodiment described above, the processing for each key is performed in real time, but it is not necessary to do so. - After the performance information of the oil is given, the same calculation as above is executed and the result of key discrimination is shown. You can do it like this. At this time, in addition to inputting music information by actually playing the music on the keyboard, performance information can be supplied in various forms. For example, it is possible to input voice from a microphone or to provide pre-encoded automatic performance information to the tonality discrimination device.

Further, by supplying the discrimination result output of the tonality discrimination device to the automatic accompaniment device and controlling the key of the automatic accompaniment, it is also possible to perform optimal automatic accompaniment. In this way, the key discrimination result output can be used for various purposes (singer identification, karaoke imitation level identification, etc.) in addition to simply displaying the result.

<Effects of the Invention> As explained above, according to the present invention, if a membership function for each leg is prepared in advance, the tonality of a piece of music can be determined by simply performing a relatively simple calculation. There is no limit to the tonality to be used, and there is no need for a complicated algorithm to determine the tonality, making it effective.

[Brief explanation of drawings]

The drawings illustrate the present invention; FIG. 1 is a block diagram of the invention;
115 is a diagram showing the configuration of one embodiment, FIG. 3 is a diagram showing a musical score for explaining the membership function, FIG. 4 is a diagram showing the membership function of C major, and FIG. 115 is a diagram showing each root note. Figure 6 is a diagram showing the membership function of Cu major, Figure 7 is a diagram showing the membership function of C minor, and Figure 8 is a diagram showing the membership function of CI'I minor. 9 is a diagram showing the membership function of C major blues, and FIG. 1o is a diagram showing the membership function of C major blues.
Diagram showing the membership function of minor blues, 1st
1 is a diagram showing the processing flow of the CPU, FIG. 12 is a diagram showing the display format, and FIG. 13 is a diagram showing the display format. FIG. 3 is a diagram showing a state in which membership functions are expressed in pitch units smaller than a semitone. GS... Music information supply means, MS...
Membership function supply means, CA...similarity calculation means, OU...output means, 1...
CPU, 2...ROM, 3...RAM
, 4...MIDI interface, 5...
...LCD driver, 7...LCD. Patent applicant Casio Computer Co., Ltd. Figure 1 4eYJAli. 1-v-. i tt*a '7= L-轡慟----Yang-'' l@2■ W44■ C Mejimer's Metsubarge, Soft Function No. 5 @ No. 6tm C''j Diq- Member 1-no, Noah Function TLC, L No. 71! I C minor part 1, 7 function diagram 8 C# minor part 1, 7 function, 9 @ C major blues met part 27 function M2C diagram

Claims (1)

[Claims] (1) a music information supply means for supplying music information expressing a music; a membership function supply means for supplying a membership function representing the appearance frequency of each scale for a plurality of keys; The above song information,
similarity calculation means for calculating the degree of similarity with the membership function for a plurality of keys supplied from the membership function supply means; 1. A tonality discriminating device comprising: output means for outputting information representing the key of a piece of music expressed by music piece information.