JPH049893A - Melody analyzer - Google Patents

Abstract

PURPOSE:To take a melody analysis corresponding to a music style efficiently by simple constitution by providing a means which retrieves a matching melody pattern rule in a set selected by a rule set selecting means and analyzes the melody according to the specified style. CONSTITUTION:A melody pattern rule data base F12 is stored with sets of melody pattern rules defining respective different melody patterns as to respectively music styles. A pattern inspection part F13 (analyzing means) receives a sequence of intervals regarding the melody from an interval evaluation part F3 and a sequence of the kinds of sounds from a sound kind classification part F11 and retrieves the melody pattern rule matching those sequences in the melody pattern rule data base F12 selected by a data base selection part F15 to analyze the melody. Consequently, the melody is analyzed by the simple relation in relation with the style.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a music device, and more particularly to a melody analyzer that automatically analyzes a given melody.

[Background] In Japanese Patent Application No. 1-126705, the applicant has proposed a melody analyzer that automatically analyzes melodies. This melody analyzer receives tonality and chord progression as input information in addition to the melody, and can identify the function and character of each note in the melody based on the musical knowledge built into the device.

Unfortunately, this type of melody analyzer is basically a "program-driven" structure, and the musical knowledge for classifying each note of the two melodies is expressed as part of the program. For this reason, it is not easy to express musical knowledge (device), and development requires time and cost, processing is complicated, and the configuration becomes large. Also, this type of melody analyzer cannot indicate what musical style fits the melody.

[Object of the Invention J] Therefore, the object of the present invention is to improve the melody analyzer as described above, and to provide a melody analyzer that can efficiently perform melody analysis according to music styles with a simpler configuration.

[Structure and operation of the invention] According to one aspect of the present invention, a melody imparting means for imparting data representing a melody, a chord imparting means for imparting data representing a chord, and a tonality imparting means for imparting data representing a tonality. means for generating a plurality of different pitch class sets for each of the plurality of note types based on the chord and the tonality; melody sound classification means for identifying the sound type of each note of the melody and forming a sequence of note types by determining the relationship with each of the different pitch class sets; and a melody pattern rule that defines characteristics of each melody pattern composed of at least one sequence of notes for each of a plurality of different music styles. melody pattern rule database means for storing the set;
a style specifying means for variably specifying one style from among the plurality of different music styles; and a rule for selecting a set of melody pattern rules related to the style specified by the style specifying means from the melody pattern rule database means. set selection means;
Searching for a melody pattern rule that matches the arrangement of note types from the melody sound classification means and the arrangement of intervals from the interval evaluation means from the set of melody pattern rules selected by the rule set selection means. This provides a melody analyzer characterized by having an analysis means for analyzing the melody based on a specified style.

According to this configuration, the musical knowledge necessary for melody analysis is converted into a database by the melody pattern rule database means and incorporated into the device, so it is a "data-driven type".
It is possible to realize a melody analyzer of 1, and the configuration of the melody analyzer can be made simple. Furthermore, according to the present configuration, it is possible to analyze a melody from the viewpoint of style analysis, such as what music style (genre) a given melody belongs to.

Instead of the code assigning means, tonality assigning means, and note type pitch class generating means, note type pitch class set assigning means for assigning (inputting) a plurality of different pitch class sets for each of a plurality of note types may be used. In this case, there is no need to generate a pitch class set within the device, and the configuration of the melody analyzer is further simplified.

Furthermore, according to another aspect of the present invention, a melody assigning means for assigning data representing a melody, a code assigning means for assigning data representing a chord, a tonality assigning means for assigning data representing a tonality,
tone type pitch class set generation means for generating a plurality of different pitch class sets for each of a plurality of pitches based on the chord and the tonality; melody sound classification means for identifying the sound type of each note of the melody and forming a sequence of note types by determining the relationship with each of the class sets; A pitch evaluation means that evaluates pitches to form a pitch sequence, and a melody pattern rule that defines the characteristics of each melody pattern composed of at least one note sequence for each of a plurality of different music styles. /) melody pattern rule database means for storing the melody pattern rule /l/t that matches the arrangement of note types from the melody sound classification means and the arrangement of intervals from the pitch evaluation means; There is provided a melody analyzer characterized by comprising: a matching rule search means for searching from a database means; and an analysis result output means for outputting a music style related to each matching melody pattern rule from the matching rule search means. Ru.

In the case of this configuration, the above-mentioned style instruction means is not required, and the melody can be analyzed from the viewpoint of various styles, which is effective in determining the characteristics of the music style of the melody.

[Example 1 Hereinafter, the present invention will be described in detail with reference to the drawings.

Overall Configuration> FIG. 1 shows a block diagram of the hardware configuration of a melody analyzer according to an embodiment. A CPU 2 operates according to a program stored in a ROM 4 to analyze a melody, and controls the entire apparatus. In addition to programs, ROM4 contains various databases (chord tone database, tension note database, diatonic scale,
A melody pattern rule database) is stored.

During operation, the RAM 6 is used as a working memory for the CPU 2, and stores melodies, keys, chord progressions, melody analysis results, and other variable data. The input device 8 is used to input the melody, chord progression, key (tonality), etc., and to specify the style of the music, and includes a melody imparting means,
It constitutes a code assigning means, a tonality assigning means, and a style indicating means. Input device M8 may include a keyboard for real-time input of melodies and chord progressions. The monitor IO is a display device such as a CRT (or LCD),
It includes a sound source connected to a printer and an audio system, and is used to display and output analysis results.

Functions of the melody analyzer> The functions of the melody analyzer are shown in Figure 2.The melody memory Fl (located in the RAM 6 in Figure 1) stores the melody data input from the input device 8 (each note has a pitch class). (expressed in octaves and lengths). Key memory F2 (located in RAM 6) stores key (tonality) data representative of the particular pitch class entered by input device 8. A chord progression memory F3 (located in RAM 6) stores data representing the chord progressions (each chord being represented by a particular root pitch class and type) that have also been input to the input device 8.

The pitch evaluation section F4 (pitch evaluation means) is the melody memory F1.
receives the melody data from
The interval formed between notes) is evaluated in the form of scale degrees. The scale generation unit F5 receives key data from the key memory F2.

Generate a pitch class set of scales (eg diatonic scales) in this key.

The chord tone generation unit F7 receives the chord progression data from the chord progression memory F3, and refers to the chord tone database F6 that stores chord tone pitch class sets for each chord type assuming the root tone of the reference pitch class. A pitch class set of chord tones (chord constituent notes) for each chord is generated.

The tension note generation unit F9 receives the chord progression data from the chord progression memory F3, and refers to the tension note database that stores pitch class sets of tension notes for each chord type assuming the root note of the standard pitch class. Generate a pitch class set of tension notes for each chord in a chord progression. The available note generation unit FIO receives the scale pitch class set from the scale generation unit F5 and the tension note pitch class set from the tension note generation unit F9, and converts the pitch class contents common to both pitch class sets into an available note pitch. Generate as a class set. Elements F5 to FIO constitute tone type pitch class generation means.

The tone type classification section Fil is a scale pitch class set from the scale generation section F5 and a food tone pitch class set from the chord tone generation section F7.

Tension note bitch class set from tension note generation unit F9, available note generation unit FIO
Based on the seven available note bitch class sets from melody memory Fl, each note of the melody is classified to identify the note type for each note.

The melody pattern rule database F12 stores a set of melody pattern rules that define each of a plurality of different (musically permissible) melody patterns for each of a plurality of music styles. Each melody pattern consists of at least one sequence of notes. Each melody pattern rule defines the characteristics of the sequence of notes in the melody pattern. In detail, a melody pattern consisting of one note is characterized (ruled) by the tone type of that note, and a melody pattern consisting of a sequence of two notes is characterized by the sequence of the tone type of each note and the two tones. It is characterized by the type of sound formed during the process. Similarly, 3
A melody pattern consisting of a sequence of three or more notes is characterized by a sequence of tones of each note and a sequence of types of pitches formed between adjacent tones in the sequence of notes. The style designation unit F14 (style designation means) variably designates one music style from among the plurality of music styles. A database selection section F15 (rule set selection means) selects a melody pattern rule set for the music style specified by the style specification section F14 from the melody pattern database F12.

The pattern inspection unit F13 (analysis means) receives the sequence of intervals related to the melody from the pitch evaluation unit F4 and the sequence of note types from the note type classification unit Fil, and selects melody pattern rules that match these sequences from the database selection unit F15. Search and analyze melodies from the selected melody pattern rule database.

In the case of the configuration shown in Figure 2, the musical knowledge necessary for melody analysis is not incorporated as part of the ZA-gram, but is stored as data in a database such as the melody pattern rule database F12, so the melody analysis knowledge is expressed. , changes can be made easily, and a ``data-driven'' melody analyzer can be realized.Furthermore, this configuration can provide an analysis result showing what kind of melody pattern a given melody is composed of.

Furthermore, this configuration can analyze a melody by taking into account the possibility of various musical styles that the melody is trying to analyze.
It is also effective for determining the style of a melody.

Details The details of the melody analyzer according to the embodiment will be described below.

Figure 3 shows the format of the pitch class identification data used in this melody analyzer.The pitch class of "C" is represented by numerical data "θ", the pitch class of C# is represented by numerical data "1", and so on. The semitone difference is the numerical difference “1”
', the pitch class of B is expressed as numerical data "11". The pitch class data of each note in the melody, the pitch class data of the root note of the chord, and the pitch class data of the key are expressed according to the format shown in FIG.

Figure 4 shows the format of the pitch class set data. Since there are a total of 12 pitch classes from C to B, each digit (each bit position!) of the 12-bit data is shown in the figure.
i) each pitch class, the least significant digit (bit 0)
is pitch class “c”, and the next digit is pitch class “C”.
#”, and the same goes for the highest digit as pitch class “B”
For example, pitch class set = (C ,D
, E, F, G, A, B) are expressed as 101010110101 (binary). Each data of the court tone pitch class set, tension note pitch class set, scale note pitch class set, and available note pitch class set follows the format shown in FIG.

Figure 5 shows the format of chord type identification data.0 According to the format shown in the figure, major type chords are represented by the numerical value "0", and minor chords are represented by the numerical value "0".
The seventh (7th) chord is represented by the number “2”.
”. The code type set may be extended to include other food types.

FIG. 6 shows the structure of melody data stored in the melody memory 42 (corresponding to melody memory F in FIG. 2).The melody data is a linked list of note data (-
dimensional array).

Each note data consists of pitch class, octave, length, and next note address data.

Pitch class! The octave is expressed according to the format shown in the figure 3, with the octave starting from the middle C being the number "4", and the octave being assigned consecutive numbers in the order of the octave, and the length being the length of one bar being the number "16". In FIG. 6, Melody represents the base (start) address of the melody memory 42, and the next note address field in the last note data area of the melody contains a melody end identification mark indicating the end of the melody data list. “
0'' is stored.

FIG. 7 shows the structure of chord progression data stored in the chord progression memory 44 (corresponding to chord progression memory F3 in FIG. 2).The chord progression data is expressed as a linked list (-dimensional array) of chord data. Ru. Each chord data is the pitch class of the root note.

with information on the type, length and address of the next code,
The root note pitch class data follows the format shown in Figure 3, the type data follows the format shown in Figure 5, and the length data follows the same format as the melody note length data. , becomes the mark "0" indicating the end of the chord progression list. Cproc shown in FIG. 7 represents the base (start) address of the chord progression memory 44.

FIG. 8 shows the configuration of the key memory 46 (corresponding to key memory F2 in FIG. 2).
Data indicating the pitch class of the key (tonality) input from 8 is stored.

The key pitch class data follows the format shown in FIG. 3, and the address of the key memory 46 is given by Key.

Figure 6, Figure 7 and! The contents of each memory shown in Figure s8 are shown in musical notation in Figure 9.

FIG. 1θ shows the code tone database memory 62 (code tone database F6 in FIG. 2) located in the ROM4.
In the court tone database memory 62, pitch class set data of chord tones (chord constituent notes) for each chord type whose root tone is a reference pitch class C is stored. The base address of code tone database memory 62 is given in CTdB. The value of the code type identification data (see Figure 1''5) determines the relative address (address distance from the base address) of the code tone database memory 62.The code tone pitch class stored at each address of the code tone database memory 62. The set data follows the format shown in FIG.

FIG. 11 shows the configuration of the tension note database memory 64 (corresponding to the tension note database F8 in FIG. 2) located in the ROM 4. The tension note database memory 64 stores pitch classes of chord root notes. Tension note pitch class set data for each chord type C is stored. The base address of the tension note database memory 64 is TNdB.
is given by The value of the chord type identification data defines the relative address of the tension note database memory 64.

FIG. 12 shows the configuration of the diatonic scale memory 66 placed in the ROM 4. The address of the diatonic scale memory 66 is given by pi atonic,
This address has a diatonic major scale pitch class set with keynote pitch class C (
C, D, E, FG, A, B) are stored.

1$13 The figure shows a melody pattern rule database memory 6 for one music style stored in ROM4.
8 (corresponding to a part of the melody pattern database F12 in FIG. 82). The melody pattern rule database memory 68 is composed of a list (-dimensional array) of musically permissible melody pattern rules, and each melody pattern rule is expressed as a list of abstracted note (FNOTE) data. Each F
The NOTE data consists of note type identification data, pitch direction identification data and pitch distance identification data for representing pitch conditions for the next note, and the address value of the next note. The note type identification data follows the format shown in No. 14FgJ, and when the note type is “scale note”, it is “l”.
, 2'' for “Tension Note”, “3” for “Available Note”, “4” for “Avoid Note”, “5” for “Any Note”
The pitch direction identification data follows the format shown in Figure 15, and takes the value "0" when the pitch increases from the current note to the next note ("10"), and when the pitch decreases. (“-”) takes the numerical value “1”, when the pitch is the same, the numerical value “2” is taken, and when “any direction” is acceptable, the numerical value “3” is taken.The pitch distance identification data is the 16th According to the format shown in the figure, when the pitch distance between the current note and the next note is zero (same note), the number is "0", and when the pitch distance is a semitone, the number is "1", and from the current note to the next note. When the pitch is a progression (semitone or whole tone), the value is “2”, and when the pitch distance is a whole tone, the value is “3”.
The base of the melody pattern rule database memory 68 takes the value "4" when jumping from the current note to the next note (a pitch distance greater than total blindness), and takes the value "5" when an arbitrary pitch distance is acceptable. The address is P
It is given in RdB. Value “FH” in pitch direction area
'' represents the end of one pattern in the primary pattern rule address area. 6 ['0' represents the end of the melody pattern rule database.

The 25th Eiiff shows the structure of the entire melody pattern rule database (corresponding to the melody pattern rule database F12 of the 2nd fl!ff). As shown in the figure, the entire database stores melody pattern rule sets for each of the four music styles. . The nursery rhyme database 682 with the base address Doyo stores a set of melody pattern rules that are allowed in the music style "劃", and the enka database 684 with the base address Enka stores the set of melody pattern rules that are allowed in the music style "Enka". A hops database 686 that stores a set and has a base address Pops is a blues database 688 that stores a set of melody pattern rules allowed in the music style "Hops" and has a base address and glues.
stores a set of melody pattern rules that are allowed in the music style "Blues". In the configuration shown in Figure 25, the bottom right of the table of the enka database 684 is
A linker (pointer) of the nursery rhyme database 682 indicated by O: is shown. Therefore, the actual enka database 684 includes the contents of the nursery rhyme database 682 as a part thereof. Similarly, the actual hops database 686 includes the enka database 684 (and nursery rhyme database 682) as part of it, and the blues database 688 is created by the linker-Enka.
ps: contains the hops database 686 (and enka and nursery rhyme databases 684 and 682).

These four types of databases 682.684.686
．． 688 to be used in the melody analysis process is determined by the style instruction section 81.688 as illustrated in FIG. (
(a part of the input device 8) is selected according to a style instruction. When you press the “DOYO” switch, the song style “DOYO!”
! ” is selected, and the style change @5TYLE takes the value 0. Similarly, “ENKA”, “POPS”, “
In response to the press of each switch "BLUES", the music styles "Enka", "Hops", and "Blues" are selected, and the style variable 5TYLE is set to the values of 1'', 2'', and 3''.

During the melody analysis process, the header memory 69 having the base address Header: as shown in FIG. Read and use the database (of the specified style) pointed to by this header data.

Figure 17 shows a list of the main variables. The variable DTS is the second
This corresponds to the output of the scale generator F5 in the figure,
The variable DTS represents a diatonic scale (pitch class set of) having the pitch class stored in the keynote memory 46 as a key. It is obtained by rotating (rotating) to the left by the value of the data.

As an example, the tie 7 tonic scale pitch class set data for key C and the diatonic scale pitch class set data for key E are shown as follows: 101010110101 (key C) 1011
01011010 (Key E) The key E diatonic scale pitch class set (12 bits) data is the key C diatonic scale pitch class set (12 bits) data +-H data value "4" obtained by rotating to the left. variable TE
N corresponds to the output of the tension note generator F9 in FIG. 2, and is a 6-variable CRT (
2) represents the pitch class set of code tones for a given code.

Variable AVN (Available note generator F in Figure 2)
The available note pitch class set data AVN (corresponding to the output of IO) represents the pitch class set of the available notes.The available note pitch class set data AVN is obtained by performing a bitwise logical product of the variables TEN and DTS.

The variable N0TETYPE (corresponding to the output of the note type classification unit FIZ) represents the note type of the given melody note.
The variable INTERVAL represents the scale degree of the interval between a given (current) melody note and the next melody note (corresponding to the output of the interval evaluation unit F4).
TERVAL is obtained in a first-order manner.

INTERVAL=NOTE(Next)(OctNO
) X12+N0TE(Next)(PC:)iNO
TE (OctNO)
represents the next two-tone pitch class, N0TE(Oct
NO) represents the octave of the current note, N0TE(P
C) represents the pitch class of the current note.

Since the other variables are clear from FIG. 17, their explanation will be omitted here.

FIG. 18 shows the main flow of the operation of the embodiment, 18-
1, the CPU 2 waits for user input from the input device 8. If there is data input (melody, key, chord progression) (18-2), data input processing routine 18-3
The data obtained by λ is stored in the RAM 6. Input device W8
When there is a style instruction from the style instruction section 81 (18-4), the style setting process 18-5 is executed and the style variable 5TYLE is set to a value indicating the instructed style (1g-5). If an analysis instruction is given (98-6), each melody note stored in the melody memory 42 is classified to identify the note type.

The note type classification process 18-7 includes a time-based subroutine (not shown) for determining the chord in the chord progression memory 44 that exists at the time of each melody note in the melody memory 42. The temporally corresponding chords on the chord progression memory 44 are obtained as follows. The horizontal position of the melody note (accumulated length from the beginning of the melody to the immediately preceding note) is determined, and the chord length on the chord progression is accumulated from the beginning until the determined note position is exceeded. The chord that is exceeded is the chord CHORD that corresponds to the melody note. Furthermore, the tone type classification process 18-7 is performed based on the pitch class of the chord tone.
T, diatonic scale pitch class set DTS
, a tone type pitch class set generation subroutine for obtaining a tension note pitch class set TEN, and an available note pitch class set AVN.
CHT, DTS, TEN, AVN are rotation functions R
OTATE (a, b) (pitch class set data a
Rotate left by b) and CHT-ROTATE(” (CTdB+CHQRD(
Type)I.

CWORD(RootPC)) DTS=ROTATE(”rDiatoniclJEY
)TEN-ROTATE(t [TNdB◆GHOR[
1 (Type) ICWORD (RootPc)
) AVN-TE is given by bitwise logical product of N and DTS. Here, [al] represents the data stored at address a, and note type classification processing 18-7 shows the inclusion relationship between the four pitch classes (CHT, DTS, TEN, AVN and note pitch class N0TE CPC). In step 19-1, it is determined whether the pitch class N0TE (PC) of the melody note is included in the chord tone Virachi Class Sera) CHT. investigate. Specifically, the pitch class data of the melody note is converted from the format shown in Fig. 3 to the format shown in Fig. 4 (the 8th digit of N0TE (PC) of the 12-bit data becomes "1"), A logical product is performed for each bit between the note pitch class 12-bit data after format conversion and the code tone pitch class set 12-bit data CHT. If the result is not "0", the melody note's pitch class is included in the chord note's pitch class set; if the result is 0", the melody note's pitch class is not included in the chord note's pitch class set; the former In this case, set the chord tone identification value in the variable N0TETYPE to indicate that the note type is a "chord tone" (19-2).Similarly, if the pitch class of the melody note is 7V available. If it is included in the note bit class set AVN, the melody note is classified as the note type “Available” (19-2,
! 9-3), if the pitch class of a melody note is included in the diatonic scale pitch class set DTS, that melody note is classified as a note type "scale note" (Eng. 9-4.19-5), Pitch Class Set is Tension Note Pitch Class Set.

If it is included in TEN, the melody note is classified as a note type “tension note” (19-6, 19-7).
), the pitch class of the melody note is CHT, AVN,
DTS, TEN (71') If it is not included in the reno pitch class set, the melody note is classified as the note type "avoid note" (19-8). Let's analyze the note type identification flow (Figure 8819). The process is repeated for each note of the melody, and as a result, a sequence (-dimensional array) of note types corresponding to the sequence of notes of the melody to be analyzed is obtained.

Returning to FIG. 18, after the note type classification process 18-7, a pitch evaluation routine 18-8 is executed. This routine 18-
In step 8, the scale frequency of the tones between adjacent notes of the melody to be analyzed is evaluated, and the sequence of intervals related to the melody (
−-dimensional array).

Finally, the CPU 2 executes a pattern inspection routine 18-9 to check the melody pattern rule database memory 68 for the arrangement of note types and pitches obtained from the melody.
Search for matching pattern rules from .

Details of the pattern inspection routine 18-9 are shown in FIG. 20, and initialization is performed at 20-1. The details of initialization are explained in the 21st
The base address Mel of the melody memory 42 shown in the figure
Set ody to note pointer variable N0TE.2l
-1), base address DEC of pattern inspection pass/fail array
is set in the pass/fail pointer variable DECN (21-2)
, Next, the loop from 21-3 to 21-6 is executed, and each element DECN (DEC) of the pattern test pass/fail array is set to "0" indicating failure (no matching melody pattern) for the number of notes in the melody memory 42. Set. At the end (21-7), melody memory 4 is returned to variable N0TE.
2 base address Melady is set.

Returning to No. 205J, in 20-2 the rule pointer variable R
Base address book of melody pattern rule data memory related to ULE designation style [Header: +5t
Initialize to ylego. Address RULE (F
NOTE) is set to pattern note pointer variable FNOVH, and note pointer variable N0TEO is set to N0TE at 20-4. 20-5, the note type N for the melody note pointed to by the melody note pointer N0TEO.
0TETYPE and pitch I NTERVAL are extracted from the note type array and pitch array, respectively.

Check the logical relationship between these and the corresponding elements (note type, pitch direction, pitch distance) of the pattern note pointed to by the pattern note pointer FNOTH to determine whether the melody note satisfies the pattern note conditions of the melody pattern rule. do. The note type condition is pattern note note 4FN.
If OTE (TYPE) is “any note type”, otherwise the note type of melody note is N0TE (T y p
e) is the pattern note type FNOTE (Type
) holds true when it is equal to Also, since the pitch class set of available notes is defined as a common part of the diatonic scale pitch class set and the pitch class set of tension notes, the tone type FNOTE (TYPE) of the pattern note is "scale note" or 6-ten. melody note type No. 0
The note type condition may also be established when TE (Type) is an “Available Note”.For pitch inspection, the pitch direction FNOTE (D
i r) and the pitch distance FNOTE (Mo t i), the range of allowable pitch scale frequencies is determined, and if the pitch I NTERVAL of the melody note is within this range, the pitch condition is satisfied.

If check 20-5 is not satisfied, the checked melody note does not satisfy the pattern note conditions of the melody pattern rule of interest, so the next melody pattern rule is selected (20-1120-3, 2O-4).
), the sequence of melody notes is checked again for the next melody pattern rule.

If check 20-5 is satisfied, the next element of the melody pattern rule and the next element of the melody are to be inspected when the next check 20-5 is executed.
Set E (NEXT) to FNOTE) L (2
0-6), N0TEO(NEXT)t-NOTEOf,
:Se/ ) to do (20-7).

At 20-8, book [N0TEO] = O (the melody note checked before is the end of the melody) and book [FNOTE
] sFH (pattern no tested before) [arrival 1f
(This is not the end of the current melody pattern rule), the melody pattern rule of interest is deemed not to match, and the next melody pattern rule is selected (20
-11), 20-9, when [FNOTE] = FH, the melody pattern rule that is being focused on matches the notes from the note indicated by NOTE to the note before NOTEO in the melody, so these melodies Execute the pass process 20-10 for the note. Figure 22 shows the details of the passing process. Variable N.

The base address M e 1 of the melody memory 42 is
Set o dy 22-1), and set the base address DEC of the pass/fail array to variable DECN 22-2)
, through loops 22-3 to 22-5, the address DECN of the element of the pass/fail array corresponding to the melody note pointer NOTE is obtained. Then, loops 22-6 to 22-7
is executed, and each element DECN (DEC) of the pass/fail array corresponding to the note pointed to by the melody note pointer NOTE to the note one before the note pointed to by the melody note pointer NOTEO is passed (existence of a matching pattern rule). Set “l”. If desired, the monitor 10 may be
(Furthermore, it is also possible to display the corresponding melody pattern rule along with the melodies/notes that passed).

Passing process 20-10 (i') Next, RULE (NEX
T) to select the next melody pattern rule (20-1 1), if the end of the melody pattern rule database 68 has not been reached (20-1
In step 2, [RULE] = O does not hold), repeat the process from step 24-3 to check whether the newly selected melody pattern rule matches the arrangement of the melody notes. When the end of the melody pattern rule database 68 is reached, select the next melody note with NOTE=NOTE (Next 20-13), and if the previous note is not the end of the melody (20-14, select the next melody note).
E] = 0 does not hold), perform the processing below 20-2,
A melody pattern rule matching the sequence of notes with the newly selected melody note as the first note is searched from the melody pattern rule arter memory 68. If check 20-14 is satisfied, the pattern inspection routine ends.

As described above, in this embodiment, the melody pattern rule database (682, 684, 68
6, 688), the knowledge necessary for melody analysis is made completely independent of the program as data, so it is easy to construct a melody analyzer and it is possible to analyze melodies in relation to styles.

[Modifications] This concludes the description of the embodiments, but various modifications and changes can be easily made within the scope of the present invention.

For example, in the embodiment described above, a music style is specified by the style instruction section 81, and during analysis processing, a database related to the specified style is selected from the melody pattern database memory for each style, but instead of this, There may be no style instruction input, and the melody may be analyzed for all style melody pattern rule databases during analysis processing and the results output. Specifically, the number of rule databases for each style may be set to 5RDBN. , repeat the pattern inspection routine in Figure 20 5RDBN times (5TLYE in 20-2 is
to 5RDBN-1), initialize 20-
1 (Figure 21), when initializing in 0 pass processing 21-20 (Figure 22), which secures the storage area of DFC and DFCN by 5RDBN and initializes the data in the area corresponding to the value of 5TYLE. Similarly, data in the area corresponding to 5TYLE is changed. Furthermore, each time the passing process is executed, or after pattern inspection for all styles is completed, the music style related to the matched melody pattern is displayed on the monitor along with the melody note that passed once.

Further, for example, in the embodiment, the melody pattern rule database 68 is secured in the ROM 4 as a permanent data source, but the melody pattern rule database may be configured to be editable on a read/write memory such as a RAM.

In addition, the pitch class set for each note type is inputted from the input device 8 or the like, and the F2. F3, F
The elements 5 to FIO may be unnecessary.

[Effects of the Invention] As described above in detail, according to the present invention, the musical knowledge necessary for analyzing melodies that have the possibility of various styles can be acquired using melody pattern rule-based means for each style. Since it has been created as a database, it can be easily structured and analyzed by associating melodies with styles.

[Brief explanation of drawings]

FIG. 1 is a system block diagram of a melody analyzer according to an embodiment of the present invention. FIG. 2 is a functional block diagram showing the functions of the melody analyzer of FIG. 1. FIG. 3 is an example of the format of pitch class identification data. Figure 4 is a diagram illustrating the format of pitch class set data, Figure 5 is a diagram illustrating the format of chord type identification data, Figure 6 is a diagram showing the configuration of melody memory, and Figure 7 is a diagram illustrating the chord progression. FIG. 8 is a diagram showing the structure of the key (tonality) memory; FIG. 9 is a diagram representing examples of input music data stored in the memories of FIGS. 6 to 8 in musical scores; Figure 1θ is a diagram showing the configuration of the court tone (constituent note) database memory, Figure 11 is a diagram showing the configuration of the tension note database memory, Figure 12 is a diagram showing the configuration of the diatonic scale memory, and Figure 13 is a diagram showing the configuration of the diatonic scale memory. The figure which shows the structure of the melody pattern rule database memory. Figure 14 shows the format of data for each note sm. Figure 15 shows the format of pitch direction identification data. Figure 16 shows the format of pitch distance identification data. Figure 17 shows a list of variables. Figure 18 is a main flowchart executed by the CPU in Figure 1, Figure 19 is a flowchart for note type classification, Figure 20 is a flowchart for melody pattern inspection, Figure 21 is a flowchart for initialization, and Figure 22 is a flowchart for initialization. 23 is a diagram showing the configuration of the music style instruction section; FIG. 24 is a diagram showing the configuration of the header memory for the melody pattern rule database memory by style; The figure is a diagram showing the configuration of a style-specific melody pattern rule database memory. Fl, F2...Melody memory F3.44...
...Chord progression memory F246...Key memory F6.62...Chord tone database F7
...Chord tone generation section F8, 64...Tension note database 66...Diatony 7 memory F5...
...Diatonic scale generation section F9...Tension note generation section FIO...Available note generation section Fll...Tone type classification section F4...Pitch Classification section Fl2.68... Pattern rule database FI3... Pattern inspection section F14.8... Style specification section F15...
... Data selection section 682 ... Nursery rhyme database 684 ... Enka database

Claims (4)

[Claims] (1) A melody assigning means for assigning data representing a melody, a code assigning means for assigning data representing a chord, a tonality assigning means for assigning data representing a tonality, and a plurality of melody assigning means for assigning data representing a melody; by determining the relationship between the pitch class of each note of the melody and each of the plurality of different pitch class sets; , a melody sound classification means for identifying the note type of each note of the melody and forming a sequence of note types; and an interval for forming a sequence of intervals by evaluating intervals formed between adjacent notes of the melody. an evaluation means; a melody pattern rule database means for storing a set of melody pattern rules defining characteristics of each melody pattern composed of at least one note sequence for each of a plurality of different music styles; Style specifying means for variably specifying one style from among different music styles; and rule set selecting means for selecting a set of melody pattern rules related to the style specified by the style specifying means from the melody pattern rule database means. and a melody pattern rule that matches the arrangement of note types from the melody sound classification means and the arrangement of intervals from the interval evaluation means from the set of melody pattern rules selected by the rule set selection means. A melody analysis machine comprising: analysis means for analyzing the melody based on a specified style by searching; (2) melody assigning means for assigning data representing a melody; note type pitch class set assigning means for assigning a plurality of different pitch class sets for each of a plurality of note types; pitch classes for each note of the melody and the above. melody sound classification means for identifying the sound type of each note of the melody and forming a sequence of note types by determining the relationship between each note of a plurality of different pitch class sets; and a melody pattern that defines characteristics of each melody pattern composed of at least one sequence of notes for each of a plurality of different music styles. melody pattern rule database means for storing a set of rules; style specifying means for variably specifying one style from among the plurality of different music styles; and melody pattern rules related to the style specified by the style specifying means. a rule set selection means for selecting a set of melody pattern rules from the melody pattern rule database means; A melody analyzer comprising: analysis means for analyzing the melody based on a designated style by searching from a set of melody pattern rules selected by the rule set selection means. (3) a melody assigning means for assigning data representing a melody; a code assigning means for assigning data representing a chord; a tonality assigning means for assigning data representing a tonality; by determining the relationship between the pitch class of each note of the melody and each of the plurality of different pitch class sets; , a melody sound classification means for identifying the note type of each note of the melody and forming a sequence of note types; and an interval for forming a sequence of intervals by evaluating intervals formed between adjacent notes of the melody. an evaluation means; a melody pattern rule database means for storing a set of melody pattern rules defining characteristics of each melody pattern composed of at least one sequence of notes for each of a plurality of different music styles; a matching rule search means for searching the melody pattern rule database means for a melody pattern rule that matches the note type arrangement from the classification means and the pitch arrangement from the pitch evaluation means; A melody analyzer comprising: analysis result output means for outputting a music style related to each compatible melody pattern rule. (4) melody assigning means for assigning data representing a melody; note type pitch class set assigning means for assigning a plurality of different pitch class sets for each of a plurality of note types; pitch classes for each note of the melody and the above. melody sound classification means for identifying the sound type of each note of the melody and forming a sequence of note types by determining the relationship between each note of a plurality of different pitch class sets; and a melody pattern that defines characteristics of each melody pattern composed of at least one sequence of notes for each of a plurality of different music styles. melody pattern rule database means for storing a set of rules; and melody pattern rule database means for storing melody pattern rules that match the arrangement of note types from the melody sound classification means and the arrangement of intervals from the pitch evaluation means. A melody analyzer comprising: a matching rule searching means for searching from the matching rule searching means; and an analysis result outputting means for outputting a music style related to each matching melody pattern rule from the matching rule searching means.