JP2958794B2 - Melody analyzer - Google Patents

Description

Description: TECHNICAL FIELD [0001] The present invention relates to a music apparatus, and more particularly, to a melody analyzer that automatically analyzes a given melody.

[Background] The present applicant has proposed a melody analyzer that automatically analyzes a melody in Japanese Patent Application No. 1-126705. 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 sound of the melody based on the music knowledge incorporated in the device.

Unfortunately, this type of melody analyzer is basically a "program driven" configuration, in which musical knowledge for classifying each sound of the melody is expressed as part of the program. For this reason, it is not easy to express music knowledge (device), it takes time and cost for development, processing is complicated,
There was a problem that the configuration became large.

[Object of the Invention] Accordingly, an object of the present invention is to improve the melody analyzer as described above and to provide a melody analyzer capable of efficiently performing a melody analysis with a simpler configuration.

According to one aspect of the present invention, a melody providing means for providing data representing a melody, a code providing means for providing data representing a code, and a tonality providing means for providing data representing a tonality Means, a sound type pitch class set generating means for generating a plurality of different pitch class sets for each of a plurality of sound types based on the chord and the tonality, a pitch class of each sound of the melody and the plurality of A melody sound classifying means for identifying a sound type of each sound of the melody to form a sequence of sound types by determining a relationship with each of the different pitch class sets, and a melody sound classifying means formed between adjacent sounds of the melody. Pitch evaluation means for evaluating pitches to be formed to form a sequence of pitches, and for each of a plurality of different melody patterns. A melody pattern rule database means for storing a set of melody pattern rules defining at least one feature of a sequence of sounds constituting each melody pattern; a sequence of the tone types from the melody sound classification means; Analysis means for analyzing the melody by searching the melody pattern rule database means for a melody pattern rule which matches the arrangement of the pitches from the means.

According to this configuration, the music knowledge necessary for the analysis of the melody is constructed as a database by the melody pattern rule database means, so that a "data-driven" melody analyzer can be realized. The configuration can be simplified. Further, according to this configuration, it is possible to analyze the melody from the viewpoint of viewing as a combination of allowable melody patterns.

The sound type pitch class set generating means includes, for example,
Chord tone generation means for generating a chord tone pitch class set from a given chord, tension note generation means for generating a tension note pitch class set from the chord, and scale note generation for generating a scale note pitch class set from the tonality And an available note generating means for generating an available note pitch class set from the scale note pitch class set and the tension note pitch class set. In this case, the melody note classifying means classifies the melody note having the pitch class included in the chord tone pitch class set as a tone type “chord tone”, and converts the melody note having the pitch class included in the tension note pitch class set into a sound. A melody note with a pitch class included in the scale note pitch class set classified as the kind "tension note" and a melody note with a pitch class included in the available note pitch class set Is classified as a sound type “available note”, and a melody note having a pitch class not included in any pitch class set is classified as another sound type.

Each melody pattern rule stored in the melody pattern rule database means includes data representing at least one sequence of tone types and data representing a sequence of intervals between adjacent sounds in the sequence of at least one tone type. Can be represented by a data structure consisting of

Instead of the chord assignment means, the tonality assignment means, and the tone type pitch class set generation means, a tone type pitch class set assignment means for assigning (inputting) a plurality of different pitch class sets for each of a plurality of tone types may be used. Good. In the case of this configuration, it is not necessary to generate the pitch class set inside the device, and the configuration of the melody analyzer is further simplified.

Embodiment An embodiment of the present invention will be described below with reference to the drawings.

<Overall Configuration> FIG. 1 shows a block diagram of a hardware configuration of a melody analyzer according to the embodiment. The CPU 2 operates according to a program stored in the ROM 4 to analyze the melody and controls the entire apparatus. The ROM 4 stores various databases (code tone database, tension note database, diatonic scale, melody pattern rule database) in addition to the programs. The RAM 6 is used as a working memory of the CPU 2 during operation, and stores melody, key, chord progression, melody analysis result and other variable data. The input device 8 is used for inputting a melody, a chord progression, a key (tonality) and the like, and constitutes a melody adding unit, a code adding unit, and a tonality adding unit. For real-time input of melody, chord progression, the input device 8 may include a keyboard. Monitor 10 is a CRT
(Or LCD), including a sound source connected to a display device, a printer, and an audio system, used for displaying and outputting analysis results.

<Function of Melody Analyzer> The function of the melody analyzer is shown in FIG. The melody memory F1 (located in the RAM 6 in FIG. 1) stores melody data (each sound is represented by a pitch class, an octave, and a length) input from the input device 8. Key memory F2
(Placed in the RAM 6) stores key (tonality) data representing a specific pitch class input by the input device 8. The chord progression memory F3 (located in the RAM 6) stores data representing the chord progression (each chord is represented by a specific root pitch class and type) input from the input device 8.

The pitch evaluation unit F4 (pitch evaluation means) receives the melody data from the melody memory F1, and evaluates the pitch formed between adjacent sounds (notes) of the melody in the form of scale frequency. The scale generation unit F5 receives the key data from the key memory F2 and generates a pitch class set of a scale (for example, a diatonic scale) of the 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 a chord tone pitch class set for each chord type assuming the root of the reference pitch class, and performs chord progression. , A pitch class set of chord tones (chord constituent sounds) for individual chords is generated. The tension note generation unit F9 receives the chord progression data from the chord progression memory F3, and refers to a tension note database that stores a tension note pitch class set for each chord type assuming a root pitch of a reference pitch class, and performs chord progression. Generate a pitch class set of tension notes for each chord in. Available note generator F10 is scale generator F
It receives the scale pitch class set from 5 and the tension note pitch class set from the tension note generation unit F9, and generates pitch class contents common to both pitch class sets as an available note pitch class set. Elements F5 to F10 constitute a tone type pitch class generating means.

The tone type classification unit F11 includes a scale pitch class set from the scale generation unit F5, a chord tone pitch class set from the chord tone generation unit F7, and a tension note generation unit F9.
Based on the tension note pitch class set from the melody memory F1 based on the tension note pitch class set from the available note generation unit F10 and the note type for each sound Identify.

The melody pattern rule database F12 stores a set of melody pattern rules that define each of a plurality of (musically acceptable) different melody patterns. Each melody pattern consists of at least one sequence of sounds. Each melody pattern rule defines the characteristics of the arrangement of sounds in the melody pattern. Specifically, 1
A melody pattern consisting of two sounds is characterized (ruled) by the type of the sound. A melody pattern composed of a sequence of two sounds is characterized by a sequence of sound types of each sound and a type of a sound type formed between the two sounds. Similarly, a melody pattern composed of a sequence of three or more sounds is characterized by a sequence of sound types of each sound and a sequence of pitch types formed between adjacent sounds in the sequence of sounds.

The pattern inspection unit F13 (analysis means) receives the arrangement of the pitches related to the melody from the interval evaluation unit F4 and the arrangement of the tone types from the tone type classification unit F11, and stores a melody pattern rule and a matching melody pattern rule in the melody pattern rule database.
Search from F12 and analyze the melody.

In the case of the configuration of FIG. 2, the music knowledge required for the melody analysis is not incorporated as a part of the program but is stored as data in a database such as the melody pattern rule database F12, so that the expression and modification of the melody analysis knowledge are performed. And a “data-driven” melody analyzer can be realized. Further, the present configuration can provide an analysis result indicating what melody pattern the given melody is composed of.

<Details> Hereinafter, details of the melody analyzer according to the embodiment will be described.

FIG. 3 shows a format of pitch class identification data used in the melody analyzer. The pitch class of “C” is represented by numerical data “0”, the pitch class of C # is numerical data “1”, and similarly, the pitch class of B is numerical data “1” with a semitone difference of “1”. Expressed as 11 ".
The pitch class data of each note, the pitch class data of the root of a chord, and the pitch class data of a key in the melody are expressed in accordance with the format shown in FIG.

FIG. 4 shows the format of the pitch class set data. Since there are a total of 12 pitch classes from C to B, each pitch class is assigned to each digit (each bit position) of the 12-bit data, and the least significant digit (bit 0) as shown in the figure.
Is the pitch class “C”, and the next digit is the pitch class “C”.
# ", And so on, the highest digit is assigned as pitch class" B ", and the digit having the value" 1 "indicates the existence of a note having the pitch class assigned to that digit, thereby giving an arbitrary pitch. A class set can be represented, for example, pitch class set = (C, D, E,
F, G, A, and B) are represented by 101010110101 (binary). Each data of the chord tone pitch class set, the tension note pitch class set, the scale note pitch class set, and the available note pitch class set follows the format shown in FIG.

FIG. 5 shows the format of the code type identification data. According to the illustrated format, the major type code is represented by a numerical value “0”, the minor code is represented by a numerical value “1”, and the seventh (7th) code is represented by a numerical value “2”. The code type set may be extended to include other code types.

FIG. 6 shows the melody memory 42 (the melody memory shown in FIG. 2).
2 shows the structure of the melody data stored in F1). Melody data is represented by a linked list (one-dimensional array) of note data. Each note data includes pitch class, octave, length, and address data of the next note. The pitch class is represented according to the format shown in FIG. 3, the octave is represented according to a format in which the octave starting from the center C is set to a numerical value “4”, and consecutive numerical values are assigned in octave order, and the length is set to the numerical value “1 bar length”. It is expressed according to the format of 16 ". In FIG. 6, Melody represents a base (head) address of the melody memory 42. A melody end identification mark "0" indicating the end of the melody data list is stored in the next note address portion in the last note data area of the melody.

FIG. 7 shows the structure of chord progression data stored in chord progression memory 44 (corresponding to chord progression memory F3 in FIG. 2). The chord progress data is represented by a linked list (one-dimensional array) of chord data. Each chord data has information on the pitch class, type, length of the root note and the address of the next chord. The pitch class data of the root note follows the format of FIG. 3, the type data follows the format of FIG. 5, and the length data follows the same format as the melody note length data. When there is no next code, the next code address portion is a mark “0” indicating the end of the code progress list. Cproc shown in FIG. 7 represents the base (head) address of the chord progression memory 44.

FIG. 8 shows the configuration of the key memory 46 (corresponding to the key memory F2 in FIG. 2). The key memory 46 stores data indicating the pitch class of the key (tonality) input from the input device 8. Key pitch class data is 3rd
Follow the format shown in the figure. The address of the key memory 46 is Ke
given by y.

The contents of each memory shown in FIGS. 6, 7 and 8 are shown in musical scores in FIG.

FIG. 10 shows the structure of a code tone database memory 62 (corresponding to the code tone database F6 in FIG. 2) stored in the ROM 4. Code tone database memory 62
Stores pitch class set data of chord tones (chord constituent sounds) for each chord type having a root pitch as a reference pitch class C. The base address of the code tone database memory 62 is given by CTdB. The value of the code type identification data (see FIG. 5) determines the relative address (address distance from the base address) of the code tone database memory 62. The code tone pitch class set data stored at each address of the code tone database memory 62 follows the format shown in FIG.

FIG. 11 shows the tension note database memory 64 (the tension note database F8 in FIG. 2) stored in the ROM 4.
) Is shown. The tension note database 64 stores tension note pitch class set data for each chord type, where the pitch class of the chord root is C. The base address of the tension note database memory 64 is given in TNdB. The value of the code type identification data determines the relative address of the tension note database memory 64.

FIG. 12 shows the configuration of the diatonic scale memory 66 stored in the ROM 4. The address of the diatonic scale memory 66 is given by Diatonic, and the pitch class set of the diatonic major scale (C, D, E, F, G,
A, B) are stored.

FIG. 13 shows the configuration of the melody pattern rule database memory 68 (corresponding to the melody pattern database F12 in FIG. 2) stored in the ROM 4. The melody pattern rule database memory 68 is composed of a list (one-dimensional array) of musically acceptable melody pattern rules. Each melody pattern rule is represented by a list of abstracted note (FNOTE) data. Each FNOTE
The data is composed of note type identification data of a note, pitch direction identification data and pitch distance identification data for indicating a pitch condition for the next note, and an address value of the next note.
The sound type identification data follows the format shown in Fig. 14, and is "1" when the sound type is "scale note", "2" when it is "tension note", "3" when it is "available note", and "avoid note". "4" for ""
In the case of an arbitrary "note", a value of "5" is taken. The pitch direction identification data follows the format shown in Fig. 15 when the pitch increases from the current note to the next note ("+").
Takes a numerical value “0” and when the pitch goes down (“-”)
Takes a numerical value "1", takes the numerical value "2" when the pitch is the same, and takes the numerical value "3" when "any direction" is acceptable. According to the format shown in FIG. 16, the pitch distance identification data takes a numerical value “0” when the pitch distance between the current note and the next note is zero (same tone), and takes a numerical value “1” when the pitch distance is a semitone, Takes a numerical value "2" when progressing sequentially from the current note to the next note (semitone or whole note), takes a numerical value "3" when the pitch distance is a whole note, and jumps from the current note to the next note (a pitch larger than the whole note) In the case of "distance", a numerical value "4" is taken, and when an arbitrary pitch distance is sufficient, a numerical value "5" is taken. The base address of the melody pattern rule database memory 68 is
Given in PRdB. The value “FH” in the pitch direction area indicates the end of one pattern. The value “0” in the next pattern rule address area indicates the end of the melody pattern rule database.

Figure 17 shows a list of the main variables. The variable DTS is equivalent to the output of the scale generation unit F5 in FIG. 2, and (a pitch class set of) a diatonic scale having the pitch class stored in the key note memory 46 as a key.
Represents Variable DTS is diatonic scale memory 66
Is obtained by inverting (rotating) the 12-bit pitch class set data, which is the content of the data, to the left by the value of the data in the key note memory 46. As an example, the diatonic scale pitch class set data of key C and the diatonic scale pitch class set data of key E are shown as 101010110101 (key C) and 101101011010 (key E). The diatonic scale pitch class set (12 bits) data of the key E is obtained by turning the diatonic scale pitch class set (12 bits) data of the key C to the left by the value “4” of the key E data. The variable TEN corresponds to the output of the tension note generation unit F9 in FIG. 2, and represents the pitch class set of the tension note for a given chord. Variable CH
T (corresponding to the output of the code tone generation unit F7 in FIG. 2)
Represents a pitch class set of chord tones for a given chord. The variable AVN (corresponding to the output of the available note generation unit F10 in FIG. 2) represents the pitch class set of the available note. The available note pitch class set data AVN is obtained by taking the logical product of the variables TEN and DTS for each bit.

The variable NOTETYPE (corresponding to the output of the note type classification unit F11) represents the note type of the given melody note. Variable INTERV
AL represents the scale frequency of the pitch between the given (current) melody note (corresponding to the output of the pitch evaluation unit F4) and the next melody note. INTERVAL is obtained as follows.

INTERVAL = NOTE (Next) (OctNO) x 12 + NOTE (Nex
t) (PC)-{NOTE (OctNO) x 12 + NOTE (PC)} where NOTE (Next) (OctNO) indicates the octave of the next note, and NOTE (Next) (PC) indicates the pitch class of the next note. , NOTE (OctNO) represents the octave of the current note, and NOTE (PC) represents the pitch class of the current note.

Other variables are apparent from FIG. 17 and will not be described here.

FIG. 18 shows a main flow of the operation of the embodiment. 18-1
The CPU 2 waits for a user input from the input device 8. If there is a data input (melody, key, chord progression) (18-2), the data input in the data input processing routine 18-3 is stored in the RAM 6. When an instruction to analyze a melody is given (98-4), each melody note stored in the melody memory 42 is classified to identify a sound type.
The sound type classification processing 18-5 is a time corresponding subroutine (not shown) for determining a chord on the chord progression memory 44 existing at the time of each melody note in the melody memory 42.
including. A chord on the chord progression memory 44 corresponding to a certain note on the melody memory 42 in time is 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 in the chord progression is accumulated from the beginning until the note position is exceeded. The code at the time of exceeding is CHORD corresponding to the melody note. Further, the tone type classification processing 18-5 includes a tone type pitch class set generation subroutine for obtaining a chord tone pitch class set CHT, a diatonic scale pitch class set DTS, a tension note pitch class set TEN, and an available note pitch class set AVN. including. Here, CHT, DTS, TEN, and AVN are obtained by using a rotation function ROTATE (a, b) (turn the pitch class set data a to the left by b) to obtain CHT = ROTATE (* [CTdB + CHORD (Type)], CHORD (RootPC)) DTS = ROTATE (* [Diatonic], KEY), TEN = ROTATE (* [TNdB + CHORD (Type)], CHORD (RootPC)) AVN = TEN and bit-wise logical AND of DTS. Here, * [a] represents data stored at address a. Further, the tone type classification processing 18-5 examines the inclusion relationship between the four pitch class sets CHT, DTS, TEN, and AVN and the pitch class NOTE (PC) of the note, and identifies the tone type of the note as shown in FIG. Including various subroutines. 19−
At 1, it is checked whether or not the pitch class NOTE (PC) of the melody note is included in the chord tone pitch class set CHT. More 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 (note (PC) -th digit of the 12-bit data is "1"), and The logical AND of each bit is calculated between the converted note pitch class 12-bit data and the code tone pitch class set 12-bit data CHT. If the result is not "0", the pitch class of the melody note is included in the pitch class set of the chord note, and if the result is "0", the pitch class of the melody note is not included in the pitch class set of the chord tone.
In the former case, the identification value of the chord tone is set in the variable NOTETYPE to indicate that the note type is "chord tone" (19-2). Similarly, the melody note pitch class is available note pitch class set AVN
, The melody note is classified as a sound type “Available” (19-2, 19-3). If the pitch class of the melody note is included in the diatonic scale pitch class set DTS, the melody note is classified as a sound. Classified as species "scale notes" (19-4, 19-
5) If the pitch class set of the melody note is included in the tension note pitch class set TEN, the melody note is classified as a tone type “tension note” (19-6, 19-7), and the pitch class of the melody note is If the melody note is not included in any of the pitch class sets of CHT, AVN, DTS, and TEN, the melody note is classified as a sound type "avoid note" (19-8). The tone type identification flow (FIG. 19) is repeatedly executed for each note of the melody to be analyzed, and as a result, an arrangement (one-dimensional array) of notes of the melody to be analyzed and corresponding tone types is obtained.

Referring back to FIG. 18, after the tone type classification processing 18-5, a pitch evaluation routine 18-6 is executed. In this routine 18-6, the scale frequency of the tone type between adjacent notes of the analysis target melody is evaluated, and a sequence (one-dimensional array) of intervals related to the melody is generated.

Finally, the CPU 2 executes the pattern inspection routine 18-7, and searches the melody pattern rule database memory 68 for a suitable pattern rule for the arrangement of tone types and the arrangement of pitches obtained from the melody.

FIG. 20 shows the details of the pattern inspection routine. 20-1
Is initialized. The details of the initialization are shown in FIG. The base address Melody of the melody memory 42 is set to a note pointer variable NOTE (21-1), and the base address DEC of the pattern inspection pass / fail array is set to a pass / fail pointer variable DECN (21-2). Next, the loop from 21-3 to 21-6 is executed, and "0" indicating rejection (no matching melody pattern) for each element DECN (DEC) of the pattern inspection pass / fail array by the number of notes in the melody memory 42 is obtained. set. Finally (21-7), the base address Melody of the melody memory 42 is set again in the variable NOTE.

Returning to FIG. 20, the rule pointer variable RULE is initialized to the base address PRdB of the melody pattern rule data memory 68 at 20-2. At 20-3, the address RULE (FNOTE) of the first sound of the melody pattern rule of interest is set in the pattern note pointer variable FNOYE, and at 20-4, NOTE is set in the note pointer variable NOTE0. 20-5
Then, the note type NOTETYPE and interval INTERVAL for the melody note pointed to by the melody note pointer NOTE0 are extracted from the tone type array and interval array, respectively, and the corresponding elements of the pattern note pointed to by the pattern note pointer FNOTE (note type, pitch By examining the logical relationship between the melody note and the melody note, it is determined whether or not the melody note satisfies the condition of the melody pattern rule. If the note type FNOTE (TYPE) of the pattern note is "arbitrary note type", the note type of the melody note is not.
This holds when (Type) is equal to the note type FNOTE (Type) of the pattern note. In addition, the pitch class set of available notes is defined as a common part between the diatonic scale pitch class set and the pitch class set of tension notes, so the sound type of the pattern note FNOT
The tone type condition may also be satisfied when E (TYPE) is “scale note” or “tension note” and the melody note tone type “NOTE (Type)” is “available note”. For pitch inspection, the pitch direction of the pattern note
From FNOTE (Dir) and the pitch distance FNOTE (Moti), a range of the allowable scale frequency of the pitch is obtained, and if the melody note interval INTERVAL falls within this range, the pitch condition is satisfied.

If the check 20-5 is not satisfied, the checked melody note does not satisfy the condition of the pattern note of the melody pattern rule of interest, so the next melody pattern rule is selected (20-11, 20-3, 20-20). -4), the arrangement of the melody notes is checked again for the next melody pattern rule.

If the 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, so that FNOTE (NEX
T) is set to FNOTE (20-6), and NOTE0 (NEXT) is NOT
Set to E0 (20-7). * [NOTE0] = 0 at 20-8
(The last melody note checked is the end of the melody)
And if * [FNOTE] ≠ FH (the pattern note inspected immediately before is not the end of the melody pattern rule of interest), the melody pattern rule of interest is not matched and the next melody pattern rule is selected. (20-11). If * [FNOTE] = FH in 20-9, the melody pattern rule of interest matches the notes from the note indicated by NOTE to the note immediately before NOTE0 in the melody, so pass processing for these melody notes Perform 20-10. Details of the pass process are shown in FIG.
Base address Melody of melody memory 42 in variable NOTE1
Is set (22-1), the base address DEC of the pass / fail array is set in the variable DECN (22-2), and the loops 22-3 to 22-2 are set.
By -5, the address DECN of the element of the pass / fail array corresponding to the melody note pointer NOTE is obtained. Then loop 22
-6 to 22-7 are executed and the melody note pointer NO
Each element DECN (DEC) of the pass / fail array corresponding to the note pointed by the TE to the note immediately preceding the note pointed by the melody note pointer NOTE0 is set to "1" indicating success (the existence of a matching pattern rule). If desired,
In the pass processing routine 20-10, the passed melody note may be displayed through the monitor 10 (further, the corresponding melody pattern rule can be displayed together with the passed melody note).

After the pass processing 20-10, RULE is set in RULE (NEXT) and the next melody pattern rule is selected (20-1).
1) If the end of the melody pattern rule database 68 has not been reached (* [RULE] = 0 is not established in 20-12),
It is checked whether the newly selected melody pattern rule matches the arrangement of the melody notes by repeating the processing of 24-3 and below. Melody pattern rule database 68
If the end of the melody is reached, the next melody note is selected by NOTE = NOTE (Next) (20-13), and if the previous note is not the end of the melody (* [NOTE] = 0 is not established in 20-14) ) And 20-2 and below, the melody pattern rule data memory 68 searches for a melody pattern rule that matches the newly selected melody note as the first note. When the check 20-14 is satisfied, the pattern inspection routine ends.

As described above, in this embodiment, the music knowledge necessary for the melody analysis is completely independent of the program as data by the melody pattern rule database 68, so that various melody analyzers can be easily constructed, and the melody is allowed. The melody can be analyzed from the viewpoint of viewing as a combination of the melody patterns to be performed.

The description of the embodiment is finished above, but various modifications and changes can be easily made within the scope of the present invention.

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

Also, a pitch class set for each sound type is input from the input device 8 or the like so that F2, F3, F5 to F5 to F5 shown in FIG.
The element of F10 may be unnecessary.

[Effects of the Invention] As described above in detail, according to the present invention, the music knowledge necessary for melody analysis is stored in the melody pattern rule database means as a database, so that a data-driven melody analyzer is provided. And the advantage that the configuration of the melody analyzer can be simplified. Furthermore, according to the melody analyzer of the present invention, it is possible to analyze a given melody from the viewpoint of viewing it as a combination of melody patterns.

[Brief description of the 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 functions of the melody analyzer of FIG. 1, and FIG. 3 exemplifies a format of pitch class identification data. FIG. 4 is a diagram illustrating the format of the pitch class set data, FIG. 5 is a diagram illustrating the format of the code type identification data, FIG. 6 is a diagram illustrating the configuration of the melody memory, and FIG. FIG. 8 shows a configuration of a key (tonality) memory. FIG. 9 shows a musical score of an example of input music data stored in the memories of FIGS. 6 to 8. FIG. 10 is a diagram showing the structure of a chord tone (constituent sound) database memory, FIG. 11 is a diagram showing the structure of a tension note database memory, and FIG. FIG. 13 is a diagram showing a configuration of a melody pattern rule database memory, FIG. 14 is a diagram showing a format of tone type identification data, FIG. 15 is a diagram showing a format of pitch direction identification data, FIG. 16 is a diagram showing a format of pitch distance identification data, FIG. 17 is a diagram showing a list of variables, FIG. 18 is a main flowchart executed by the CPU of FIG. 1, and FIG. 20 is a flowchart of a melody pattern inspection, FIG. 21 is a flowchart of an initialization, and FIG. 22 is a flowchart of a pass process of a melody pattern rule. F1, F2: Melody memory F3, 44: Chord progression memory F2, 46: Key memory F6, 62: Chord tone database F7: Chord tone generator F8, 64: Tension note database 66: Diatonic Memory F5: Diatonic scale generation unit F9: Tension note generation unit F10: Available note generation unit F11: Sound type classification unit F4: Pitch classification unit F12, 68 ... Pattern rule database F13: Pattern Inspection unit

Claims (4)

(57) [Claims] 1. A melody providing means for providing data representing a melody, a code providing means for providing data representing a code, a tonality providing means for providing data representing a tonality, and based on the code and the tonality. Sound type pitch class set generating means for generating a plurality of different pitch class sets for each of a plurality of tone types, and determining a relationship between a pitch class of each sound of the melody and each of the plurality of different pitch class sets. By this, a melody sound classification means for identifying a sound type of each sound of the melody and forming a sequence of sound types, and forming a sequence of pitches by evaluating a pitch formed between adjacent sounds of the melody For each of a plurality of different melody patterns. Melody pattern rule database means for storing a set of melody pattern rules defining at least one sound arrangement feature; arrangement of the tone types from the melody sound classification means; and arrangement of the pitches from the pitch evaluation means. Analyzing means for analyzing the melody by searching the melody pattern rule database means for a melody pattern rule suitable for the melody pattern rule. 2. The melody analyzer according to claim 1, wherein
The tone type pitch class set generating means includes: a chord tone generating means for generating a chord tone pitch class set from the chord; a tension note generating means for generating a tension note pitch class set from the chord; and a scale note pitch from the tonality. Scale note generating means for generating a class set; and available note generating means for generating an available note pitch class set from the scale note pitch class set and the tension note pitch class set. Melody analyzer. 3. The melody analyzer according to claim 1, wherein
Each melody pattern rule stored in the melody pattern rule database means includes data representing at least one arrangement of sound types to be compared with the arrangement of sound types from the sound type classification means, and data from the sound type evaluation means. A melody analyzer comprising: data representing a pitch sequence between adjacent sounds in the at least one tone type sequence to be compared with the tone type sequence. 4. A melody assigning means for assigning data representing a melody; a tone type pitch class set assigning means for assigning a plurality of different pitch class sets for each of a plurality of tone types; and a pitch class of each sound of the melody. Melody sound classification means for identifying a sound type of each sound of the melody to form a sequence of sound types by determining a relationship between the melody and each of the plurality of different pitch class sets; and adjacent sounds of the melody. Pitch evaluation means for evaluating a pitch formed between the melody patterns to form a row of pitches; and a melody pattern defining at least one feature of a row of sounds constituting each melody pattern for each of a plurality of different melody patterns. A melody pattern rule database means for storing a set of rules; Analysis means for analyzing the melody by searching the melody pattern rule database means for a melody pattern rule that matches the arrangement of the tone types and the arrangement of the pitches from the pitch evaluation means. Melody analyzer.