JP3088919B2 - Tone judgment music device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [0001] BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a music device, and more particularly to a music device.
The present invention relates to a music device capable of determining a key. [0002] 2. Description of the Related Art Conventionally, the key of each section of music is determined.
No known music device is known. [0003] SUMMARY OF THE INVENTION
The purpose is to determine the key of each section of the music automatically
It is to provide a music device. [0004] According to the present invention, a chord progression is provided.
From the chord progression providing means
In the section of each chord in the chord progression given
The key,Along the timeline of chord progression, Up to the code
Key determination means for sequentially determining according to the history of chord progression;
A tonality judgment music device characterized by having
You. According to this configuration, each chord in chord progression
To determine the key of the section without using the key of the following section
Can be. Furthermore, according to the present invention, the chord progression
Code progress giving means for giving, and the above-mentioned code progress giving means
In each chord section in the chord progression given from the column
Chord progression along the time axis of chord progression
Key determination means for sequentially determining according to the history.
The determining means determines (a) the destination of the head code of the code progression;
A leading key determining means for determining a key of a head section, and (b) a code
Compare each code except the first code of progress with the first code
And compare the comparison result with the key of the section of each code for each code.
Calculated as the difference from the key of the first section (referred to as tonal distance)
Tonality distance calculation means and (c) calculation for the code of the current section
The tonality distance issued is determined for the code of the immediately preceding section.
Tonality distance, and according to the comparison result,
By changing the tonality distance calculated for the
Determining means for determining the tonality distance of the current section;
Based on the key of the determined first section
Conversion means for converting to a key and determining the key of the current section.
A tonality determination music device characterized by having
You. An embodiment described below is a key determination music apparatus according to the present invention.
Is applied to an automatic composer. [0005] Embodiments of the present invention will be described below. Book
Automatic composer divides sound into harmony and non-harmony
Approach. Basic data for composition
As chord progress information, motif (entered by user
Melody), melody rhythm or duration
The pulse scale used to control the train, the seed of the underlying scale
Kind is given. Each sound included in the motif is used in each section.
By using the chord information used, harmony and non-harmony
And identified. The part of the motif excluding the non-harmonic sounds
A motif arpeggio (dispersed chord). This alpe
From the geo, the characteristics of the arpeggio and the
Characteristic elements) are extracted. Also, harmony and non-harmony
For motifs that are distinguished from vocal sounds, non-harmonic
Music knowledge to classify sounds (this is the production
(Stored in the rule data memory)
The type of each non-harmonic tone included in the motif (special
Sign) can be extracted. In other words, which motif
Information indicating how such non-harmonic sounds are distributed
(Characteristics of non-harmonic sounds) is obtained. Further chord progression
Extracts the hierarchical structure and tonality structure of the song. A melody is generated by an arpeggio generation process.
The process comprises a non-harmonic sound adding step and a note length sequence generating step.
In the arpeggio generation process, it is extracted from the chord progression information.
The above hierarchical structure controls the generation of arpeggios
You. The hierarchical structure dictates the generation of new arpeggios
Arpeggio features (algorithms extracted from motifs)
Apegio features or modified versions of apegios)
Is generated, and the generated arpeggio is generated.
Expressed as a pitch sequence by applying the corresponding chord
An arpeggio is generated. The generated
A non-harmonic tone is added to the lupezio. With non-harmonic sound
The music knowledge described above is used again to add. Unharmony
In inference of sound addition, non-harmonic sounds that can be added are non-harmonic sounds
That it satisfies the characteristics of
Conditions. The scale note here is extracted from the chord progression.
According to the tonality structure issued, the main tones (keys,
This is the sound on the scale with the tonality turned. Classification of non-harmonic sounds
In addition to non-harmonic sounds, use common musical knowledge
Analysis and synthesis of melody by this device
"Reversibility" is given between. Complete reversibility and
Analyzed a certain melody and obtained a certain analysis result.
And on the contrary, a melody was generated based on the analysis results.
If the generated melody matches the original melody
Is Rukoto. To add a non-harmonic sound to the arpeggio
The melody pitch train is completed. Meanwhile, the melody sound
The long sequence is obtained as follows. That is, the basic
Rhythm (length sequence) becomes the target number of notes (for example,
Pulse scale so that the total number of non-harmonics)
Use optimal join or optimal split. Which notes are how much
Pulse scale selected or split
Depends on the weight of each pulse point. Therefore consistency
Rhythm control with a sense is possible. <Overall Configuration> FIG. 1 shows an automatic operation according to this embodiment.
1 shows an overall configuration diagram of a music machine. CPU 1 is the
This is a control device for realizing the operation music function. Input device
From 2, motif (melody), chord progression, use
The type of pulse scale used and the type of scale used are entered.
Is forced. Chord progression memory 4 stores chord progression information
The memory in the memory 4 analyzes the chord progression
To extract or generate arpeggios
Used by CPU1. Scale data memory 5
Memory for storing scale data representing various types of scales
Prior to composition, the user should select the scale to use in the song.
To select a specific scale from the scale set of this memory 5.
You can choose. Production rule data memo
Re6 stores musical knowledge for classifying non-harmonic sounds.
It is. This musical knowledge separates the non-harmonic sounds contained in the motif.
For example, when adding a non-harmonic sound to the arpeggio,
Used. The pulse scale memory 7 stores various pulse schedules.
Memory (set of pulse scales)
At the beginning of composition, the rhythm feature that the user gives to the song
From the pulse scale set
The desired pulse scale can be selected. Selected
Pulse scales generate melody rhythms
Used for The completed melody data memory 8 contains
Lody data is stored. External storage device is melody
Of the melody data stored in the data memory 8 and another sound
Used as a resource for easy knowledge and another composition program.
The work memory 10 stores various types of data used by the CPU 1 during operation.
Data, such as tonal structure, hierarchical structure, and various variables
It is memorized. Monitor 11 is CRT12, stave pudding
, The tone generation circuit, and the sound system 15
Composed, and composed and analyzed results through these devices
Can be displayed and output. <Composition General Flow>
Fig. 2 shows the general flow (composition general flow) of this device
Shown in In the initial setting of 4-1, the user sets the automatic
, Basic information for composition is input. Use
The information that the user notifies the automatic composer includes (1) BEA
T, (2) Type of pulse scale (3) Scale (4) All self
The motion motif input is shown. BEAT is a basic unit
The length of one measure expressed by the number of long notes (shortest notes)
And thus defines the time signature of the song. An example
For example, for a four-time tune, the basic unit length of a note is 1
If BEAT = 16 is set as 6th note, one bar
The length is four beats. The pulse scale selected in 4-1 is
Information that temporarily controls the rhythm of songs composed by an automatic composer
Information. The pulse scale has an interval of the basic unit length
At each pulse point, ease of combining notes or dividing notes
Is a weighted scale that represents
(See FIGS. 7 and 12), using this pulse scale
This controls the melody pitch sequence. Accordingly
Therefore, selecting the type of pulse scale is
Make sure that the rhythm feature of the song that the composer composes is selected.
And Type of scale selected in 4-1 (for example,
For example, diatonic scale)
This is the scale to use. In addition, initial setting 4-1
Is the composition performed automatically or based on motifs?
Is determined by the user. When fully automatic,
Important data are (1) chord progression, (2) product
(3) The pulse scale is the work memory 10
Read on (4-3), (1) Basic rhythm
(Duration pattern), (2) Arpeggio features (PCi:
(See FIG. 5), (3) Non-harmonic features (RSi: see FIG. 5)
Essence) is generated according to the user's instructions.
(4-4). On the other hand, composition mode using motifs
Then, as data reading 4-5, besides the above data,
The motif (input melody) is also read, and the essence and
(1) Rhythm (2) Arpeggio pattern (3) A
The characteristics of the lupegio pattern and (4) the characteristics of the non-harmonic
(4-5). In particular, non-harmonic sounds
Are obtained by inference based on production rules.
Can be Full automatic or motif use mode
Is also evaluated for chord progression 4-7, where (1) hierarchy
Structure (2) Tonality structure (3) Whether the scale is chord progression information
Extracted from This hierarchical structure is the song underlying the chord progression
It expresses the coherence and diversity of Also, the tonality structure
Tsukuri stipulates the key (fundamental tone) of the scale used in each section.
Is what you do. The process denoted by “scale” is a kind of
In the special chord section, regardless of the default scale
There is to use a specific scale. This process up to 4-7
The "analytical work" for composition has been completed. An example
For example, the arpeggio feature is the generation of arpeggio patterns
Information that is necessary for the arpeggio
This characterizes the non-harmonic sound to be added. Productivity
Is the non-harmonic sound added to the arpeggio appropriate?
Used to infer no. Also, the tonality structure
Constrain the candidate for Lody. Hierarchical structure
Can be used to decide whether or not to generate
You. The pulse scale is used for generating a rhythm.
In melody generation 4-8, (1) arpeggio pattern
(LLi: see FIG. 5), (2) arpeggio
Setting the range of the pattern (3) Creating an arpeggio for pitch expression
(4) Add non-harmonic sound, (5) Generate rhythm
It is. <Variable list and data format>
Fig. 3 shows a list of the main variables used in the data format.
Are shown in FIGS. The data format is just an example.
And any other suitable data format can be used.
You. <Initial setting> Composition general flow (Fig. 2)
FIG. 9 illustrates details of the initial setting 4-1 in FIG. this
Meaning of BEAT selected in the initial setting, pulse scale
Type (PULS), scale type (ISCALE)
The meaning of taste, fully automatic or motif input is explained with reference to FIG.
Therefore, the description is omitted here. The value of PULS is
A specific pulse scale stored in the scale memory 7 (FIG. 1)
Pointer, and the value of ISCALE is the scale data.
Pointer to specific scale data stored in data memory 5
Becomes <Data reading> The composition of FIG. 2
After initialization, 4-3 or 4-5
The data reading is executed at. Fully automatic
Does not use motifs as the basic data for composition
Does not read the motif data. Of each data
The reading will be described below. FIG. 10 shows a code progression memory 4 (FIG. 1).
An example of stored chord progression data is shown. Figure 11 shows the code
Load chord progression data stored in row memory 4
It is a flowchart. In the memory map shown in FIG.
Indicates that the code type (CDi) is placed at an even address,
The length of the code is placed at the next address (odd address).
Have been. For example, C with a value of 507 in hexadecimal notation
Di is G_7thAnd has a value of 10
i has a chord length of 16 which is a basic time length (for example, a sixteenth note)
It means that it is double. [0013] In FIG. 11, a composer is included in the register CDi.
Enter the chord that appears at the i-th of the song
Enter the length. CDNO contains the total number of codes.
The other points are clear from the description of FIG.
The description is omitted. FIG. 12 shows a pulse scale memory 7 (FIG. 1).
2 shows an example of pulse scale data stored in the memory. FIG.
Selected in the initial setting from the pulse scale memory 7
5 is a flowchart for loading a pulse scale of various types.
You. In the case of this example, the song
The pulse scale seed selected to select the system characteristics
(PULS) is a special feature of the pulse scale memory 7.
Points to a fixed address (for example, 0)
Contains the start address of the selected pulse scale data.
Is in. This start address has a pulse scale
Construct subscale (a scale with only 0 and 1 weights)
The number of sub-scales is stored in each subsequent address.
Is stored. For example, normal pulse
Kale has five subscales FFFF, 5555, 1
111, 0101, 0001 (hexadecimal notation),
The corresponding binary representation is shown in FIG. Normal Pal
In the case of scale, the first pulse point (the rightmost point in FIG. 7)
Position) has a maximum weight of 5, which means that
If you select the normal pulse scale, the generated resources
The first note in each section (measure) of music
It indicates that it is easy. FIG. 14 is a production rule data memo.
Production rule data stored in Re6 (Fig. 1)
Is shown. FIG. 15 shows the data stored in the memory 6.
It is a flowchart which reads data. Production
The entire rule classifies the non-harmonic sounds contained in the melody
It expresses the musical knowledge required to perform
Option rule data is data that defines the premise of the rule.
Function data indicating the lower limit data Li and the type of function
Data Xi, upper limit data Ui, and the conclusion part of the rule.
As data Yi and Ni. Function is melody to analyze
It is a numerical representation of the characteristics of the day, an example of which is
Shown at 31. Function value Fxi represented by data Xi
Is not less than Li and not more than Ui (Li ≦ Fxi ≦ U
i) is the premise of the production rules (proposition)
And the conclusion when this premise is satisfied is the data Yi
And the conclusion when this premise is not satisfied is the data N
Indicated by i. Then, the data Yi or Ni is correct.
If the value has the value of
Indicates the number of the production rule to be used and may be negative.
The type of non-harmonic sound is expressed by its absolute value.
It is. Forward-looking inference always starts with one rule.
The rule is called a route. Conclusion Yi with negative value
The forward inference ends when Ni is found. The production rule data shown in FIG.
Address assignment for each production rule
Are five consecutive numbers starting from the address of the lower limit data Li.
Is stored at the address. Divisible by 5 in detail
The address is Li and the remainder 1 divided by 5 is Xi
However, Ui is in the address of remainder 2, and Yi is in the address of remainder 3.
For i, the data of Ni is stored at the address of the remainder 4. In FIG. 15, RULENO contains a product.
The total number of action rules is set. Other points
It is clear from the above description and the description of the flow. FIG. 16 shows the mode stored in the motif memory 3.
An example of chief data (melody data) is shown. FIG.
Flowchart that reads motif data that is the basis of composition
It is. In the case of FIG. 16, note pitch data is stored in even addresses.
Data MDi is stored in the next odd address,
Data MRi is stored. In the MDNO of FIG.
The number of notes in the motif is set. Read data
I'll finish my explanation. <Generation of Essence>
In full automatic composition mode, after reading data,
Rhythm, arpeggio pattern features,
The feature of the harmony is generated (4-4 in FIG. 2). Figure 18
Is shown in the flowchart of the essence generation of FIG. These d
Licenses can be specified by the user or generated automatically.
You. For example, setting the reference rhythm pattern in 22-1
Set the rhythm pattern generation means, for example,
4/4 time signature, normal selected as pulse scale
When you are [Outside 1] Automatically generate a reference rhythm pattern.
Or, the user can enter a favorite rhythm pattern
Done. 22-2 Arpeggio pattern feature setting and
22-3.
Performed by user input. FIG. 19 shows random number generation and the like.
Flow that automatically sets the characteristics of the arpeggio pattern
The user inputs the non-harmonic features in FIG.
And FIG. Arpeggio pattern feature setting (FIG. 19)
PC in₁~ PC_FiveIs within a section of a certain length (eg, bar)
The number of harmony sounds that make up the arpeggio, the highest harmony sound,
The lowest harmony, the maximum difference between adjacent harmony, the next
Represents the minimum value of the difference between matching chords (see FIG. 5).
See). Each PC can be generated for each section.
Set the upper and lower limits of the value, and generate random numbers between them.
And obtained. Or a series of PCs for the progress of the song
Is prepared in the database and the desired PC series is selected.
You may make it. In the characteristic setting of the non-harmonic sound shown in FIG.
The keyword corresponding to each non-harmonic tone type a
It is displayed to prompt the user for input (24-2). R
Sequence of non-harmonic tone identifier a input by the user in the array of Si
(24-3, 24-6, 24-7). End of input
When the code EOI indicating the
And exit the flow (24-8). <Extraction of Essence> Utilizing a motif
In the composition mode, after reading the data,
Essence (rhythm, arpeggio pattern, arpeggio
Pattern features, non-harmonic features) are extracted (Fig.
2, 4-6). Fig. 21 shows the rhythm evaluation of the motif.
Fig. 25 shows the extraction of the aeggio pattern.
FIG. 28 shows the feature extraction of non-harmonic sounds, and FIG. 29 shows the feature extraction of non-harmonic sounds.
For example. In these flows, each essence is a section
(Every measure). In the rhythm evaluation shown in FIG.
In Ps, the first note of the target bar is
Position information indicating whether the note is a note is entered, and Ps is represented by Ps.
The length of the note at the beginning of the indicated bar protruding into the previous bar
(Basic time expression) is entered, and Pe is the end of the target bar.
Position of the note (note before the first note of the next bar)
Enter information. Rr shown in 25-2 is the rhythm of the target bar
A 16-bit register for storing patterns.
If the node length is 16, the first bit of register rr
The position represents the first basic time of the measure, and similarly the Nth
Bit position represents the Nth basic time from the beginning of the measure
You. The processing from 25-3 to 25-9 is the musical note of the motif.
The position of the note from Ps to the note Pe is the motif length
Using the data MRi, the corresponding
This is the process of filling in the slot position. For example, as rr: 0001000100010001 Is obtained, this pattern rr is
Sound is generated at the first beat, second beat, third beat, and fourth beat of the bar
It represents that. Details of calculation of Ps, Pss, Pe, and Pee
Are shown in FIGS. Pee is the next sound of Pe
Note, the first note of the next bar interrupts the target bar
Indicates the length of the step. In the calculation flow of Ps and Pss in FIG.
And beat is the length of one measure (basic time length expression), ba
r is the number of the measure (the measure number specified by the user is
Express. The specified measure number is smaller than 1, but the song
If it is greater than the number of nodes (mno), it is an input error. Designation
If the number of bars is 1, set Ps = 1 and Pss = 0
(27-4, 27-5). The reason for Ps = 1 is first
Measure, the first note in the measure is not the first note in the song
Because it is nothing but the first note of all motif data
Yes, the reason for Pss = 0 is that there is no preceding bar
Because. A obtained in 27-2₁Is the target from the beginning of the song
This is the length of the bar to the front bar line, and this length a₁The
Length obtained by accumulating chief duration data MRi from the beginning
S (27-7, 27-8, 27-10, 27
-12). S = a₁Holds, the last accumulation in S
The next note after the i-th pitch data
Start with. Therefore, Ps = i + 1 and Pss = 0
(27-9). On the other hand, S> a₁Holds when
A note having the duration data added last to S, ie, i
The second note is the first note of the target bar. Accordingly
Let Ps = i. Also, Pss = MRi-S + a₁Toss
(27-9). The calculation flow of Pe and Pee shown in FIG.
Processing similar to that in FIG. 23 is performed. Where a₁Is the beginning of the song
The length from to the bar at the end of the target bar is included, etc.
The description of this point is omitted. The extraction of the arpeggio pattern shown in FIG.
In the low, the arpeggio pattern from the motif of the target bar
{LLi} is extracted. To summarize the processing, P
chord progression information for the measures indicated by s and Pe
Using the corresponding code in the motif data
Determine whether the sound is a harmony,
Search for the corresponding chord component sound from the chords.
To get the data in the form of LL.
Of motif data, the first to be evaluated
Find the note (Ps) and the last note (Pe) (29-
1). Next, component sound data is generated from the chord data.
(29-2, FIG. 26, FIG. 27). As shown in FIG.
The constituent sound data memory is a type of chord whose root is C.
By class, the chord constituent tones are
It is stored as data. Each bit position represents each note name
The lowest bit position is C (C). For example, c
c = 0091 (hexadecimal) is in the bit position of do, mi and so
There is “1”, which represents the constituent sound of C major. Now
If it is the code Gmaj of the target section, the CD is 000
7 (hexadecimal). From the chord configuration sound data memory,
Measure at address specified by upper 8 bits of CD
Cc (= 0091) is read out.
Then, as shown in FIG.
By turning to the left by the root value indicated by the 8 bits
The bit of “1” is bit position 7, which indicates “so”, “shi” and “re”,
Go to 11 and 2 and the Gmaj chord constituent sound is expressed
You. In this way, the code structure of the target section
Generated sound data is generated. Next, the scale counter i and the harmony
The sound counter k is initialized (29-3, 29-4). 2
The processing of 9-5 is to convert the pitch data MDi of the motif into a code structure.
This is the process of converting to the same data formation as the sound data cc.
You. For example, the sound of “S” has “1” in bit position 7.
Is converted to data mm. At 29-6, the pitch data mm is converted to the chord structure.
Checking whether or not it is a sound. This is the pitch data m
Logical product of m and chord data cc [Outside 2] Can be determined. In 29-7 to 29-13,
In bit “1” of chord configuration sound data cc,
What number matches the bit "1" of the pitch data mm
The eye is examined to see if it is an eye,
Probe number [Outside 3] To create arpeggio pattern data LLk.
You. At 29-15, the counter i is advanced to the next note, and
LL is determined until the signal reaches Pe (29-16).
The LLNO of 29-17 indicates the number of chords (al
Enter the length of the pegio pattern). The characteristics of the arpeggio pattern shown in FIG.
Extraction results of 25 arpeggio patterns {LLi}, LL
Derived from NO. In the feature extraction of the non-harmonic sound shown in FIG.
Patterns of non-harmonic types distributed in section motifs
I'm asking. Set the non-harmonic counter and note counter
(33-2, 33-3), the note of interest is non-sum
In the case of a voice sound (33-4), a mochi centered on the note
Function F representing the characteristic element of the
Perform forward inference with rules to find the type of non-harmonic sound
Substitute into RSj (33-5 to 33-8). This non-harmony
By performing sound classification processing until Pe is reached,
{RSj} contains the non-harmonic tone of the motif in the target section
The sort order is set. 33-11 PSNO
Contains the total number of non-harmonic sounds in the elephant section. Whether MDi shown in 33-4 is a non-harmonic sound
30 is shown in detail in FIG. This discriminating process
In the extraction of the arpeggio pattern (FIG. 25),
This is the same as the process for determining whether the note
Yes, and focus on the notes that make up the chords in the target section.
Can be determined based on whether or not the note name of the note to be included is included. In the calculation of the function F in 33-6,
Motivation needed to classify non-harmonics in inference
Calculate the conditions or factors of the melody. Fig. 31
FIG. 39 shows a specific example thereof. In this example, the function F
hand, F₁: Harmonic sound ahead of the note (non-harmonic sound) of interest
Is located (the position of the rear chord) F_Two: The harmony is located before the note of interest
Or (position of forward harmony) F_Three: Non-harmonic tones from forward to backward
number F_Four： Pitch difference between forward and backward harmony F_Five: Non-harmonic sound between forward and backward harmony
Height distribution F₆: The pitch of the melody from the front harmony to the rear harmony
Whether or not changes monotonically F₇： Pitch difference between back chord sound and previous sound (back sum
Pitch progression to voice) F₈： Pitch difference between the forward harmony and the succeeding one
Pitch progression from voice) Is calculated (FIG. 31). Besides this, whether it ’s a weak beat or a strong beat
Information and information for distinguishing the duration are added to the set of functions F.
You may. The flowchart for calculating each function F
The description is omitted because it is clear from the description of itself. Details of forward inference in 33-7 are shown in FIG.
Shown in First, the rule number pointer P is
Specify the rule that is the root in the application rules
Is set to "1" (44-1). After a while, Lou
Of the rule indicated by the number pointer P (Lp ≦ F
xp ≦ Up) is checked, and it is satisfied
The data Yp of the positive conclusion of the rule
Used as a pointer to the
The data Np of the rule's negative conclusion to the next rule
Used as However, when the data Yp and Np are negative
Value, the absolute conclusion has been reached, so its absolute value
(-Yp, -Np) as non-harmonic tone type identifier
Set in the logic register. According to the flow,
Condition Lp> Fxp or condition Fxp> Up of 44-5 is satisfied
, The premise Lp ≦ Fxp ≦ Up of the rule P is not satisfied.
Therefore, the data Np of the negative conclusion part of the rule is a
(44-4, 44-6), otherwise the previous
Since the advisory department is established, a is the data of the positive conclusion of rule P.
Yp enters (44-2). This a is substituted into P (44-
7) If P is a positive value, the next rule number pointer
Returns to the inspection of the next rule, and when P is negative, −P
Is the classification result of the non-harmonic sound (44-8, 44-9). As an example of the classification of non-harmonic sounds, the function F
In the calculation of F₁= 1: One back-harmonic sound is the non-harmonic sound of interest
After F_Two= -1: The forward harmony is one of the non-harmonys of interest
In front F_Three= 1: The number of notes between the preceding and following harmony is one F_Four= 8: The pitch difference between the front and rear harmony is 8. F_Five= 2: The non-harmonic sound between the preceding and following harmony is the preceding and following harmony
Distributed in the middle of the pitch of the sound F₆= 1: The melody from the front harmony to the rear harmony is
The pitch is changing monotonically F₇= 1: Progression to the back harmony with a pitch difference of 1 F₈= 7: Proceeds with a pitch difference of 7 from the forward harmony Is obtained, and the production rule data
An inference is made using the data shown in FIG. First, when P = 1 (route),
Plaque 0 ≦ F_Two≤0 Is F_Two= −1, it is not established. Therefore,
Negative conclusion of route N₁= 3 is the next rule to check
Become an interchange. For P = 3, premise of rule 3 0 ≦ F₁≤0 Is F₁= 1, it is not established. Therefore, N_Three
= 5 is the pointer for the next rule. At P = 5,
Premise of Rule 5 0 ≦ F_Four≤0 Is F_Four= 8, so it does not hold. Therefore, N_Five=
6 becomes P. At P = 6, premise of rule 6 1 ≦ F₆≦ 1 Is F₆= 1 holds. Therefore, the rule
Positive conclusion data Y of 6₆= 7 is the next access le
Pointer P of the rule. At P = 7, rule 7
Premise 3 ≦ F₈≤∞ Is F₈= 7 holds. Where Y₇= -2 (negative)
It is. Therefore, conclusion = 2 (for the classified non-harmonic
Identifier). As can be seen from this example, any kind of
For non-harmonic sounds, a finite number of propositions hold (the premise does not hold)
This means that the proposition that makes the premise false is established.
Can be identified by the musical knowledge of "equal". This knowledge
Function F is calculated to express the sense
Rules have been created. That is, the function F is
Melody used in knowledge to classify vocal sounds
Check item information, production rule data
The rule sequence up to the classification result of each non-harmonic
They are linked by pointers. <Evaluation of Chord Progression>
One of the special features is to make the most of the chord progression for the song.
It is a sign. That is, in the composition general flow of FIG.
In the chord progress evaluation shown in 4-7,
Search for the hierarchical structure and tonality structure of the song based on the progress of the song
ing. The hierarchy is related to the consistency and diversity of the song,
For arpeggio generation control in melody generation described later
Used. On the other hand, the tonality structure
(Key) change, melody generation described later
Select the scale key to use for each section in
Used for In addition, special code
The process of planning to use a special scale for sections of
Running. Hereinafter, a hierarchical structure will be described with reference to FIGS.
Is described in detail. The flow shown in FIG.
Block with the length of
5 is a flowchart for calculating the similarity of the code progress. First, the song
The length SUM is the length CRi of each code in the code progress information.
It is obtained by accumulating (45-1), barno (measure
) Is converted to the basic pitch
And set the block length to 1 (45-2) and set the song length SUM to
Calculate the number m of blocks included in the song by dividing by the lock length l
(45-3). For the reference number of the block to compare
Is initialized to "0" (45-4). In 45-8 to 45-18, the i-th
Class of code between block and j-th block (j ≧ i)
The similarity Vij is calculated. As a function of similarity, (Equation 1) You are using Where l is the length of the block and V
s is the code of the i-th block and j-th
The code obtained when comparing with the code of the eye block
Indicates the number of matches of the key. The similarity function Vij is 0 to 1
It takes values up to 00, and when it is 100, two blocks
The chord progression is completely (100%) identical, and when it is 0, it is complete
All disagreements. Class of i-th block vs. j-th block
The calculation of the similarity is started from the position of j = i (45-6),
Each time the similarity is obtained, the next block is determined by j = j + 1.
(45-20, 45-7)
After going to the last block (45-19), i = i
The i-th block is shifted by +1 (45-2).
2, 45-5) and repeat until i becomes the last block.
Return. As a result, the j-th block for the i-th block
As the similarity Vij of the code of the eye block, [Outside 4] Is obtained. Note that Vij = Vji, that is, i-th
J-th block code similarity to the block, j
Code similarity of i-th block to i-th block
The degrees are equal. Also, Vii = 100. Calculation of code matching degree between blocks in FIG.
The result {Vij} is used for generating the hierarchical structure data shown in FIG.
Used. In FIG. 42, c represents the calculation of the hierarchical structure.
And a hierarchical structure of the j-th block in Hj.
The structure identifier is set. Hj is 0, 1, 2
Takes the integer value of ...... This is a,
a′b, b ′... (see HIEi in FIG. 6).
See). In the flow of FIG.
In addition, the block with 100% code match
Of the same value (even value) as the hierarchical structure identifier of the network
With child (46-10, 46-11), 70-100%
The block that matches in the range of
As a block with a line-qualified chord progression,
Hierarchical structure identification of the value obtained by adding 1 to the lock hierarchical structure identifier
A child is attached (46-12, 46-13). Also, 7
Blocks with a similarity of less than 0% are used as reference blocks.
The block is treated as a block having a hierarchical structure independent of the block.
Select the first block of the song as the first reference block
(46-2), 100% of this reference block
Or for blocks that match 70-100%
The evaluation values of Hj = 0 and Hj = 1 are attached, and the evaluation is completed.
Therefore, the flag flj of these blocks is set to “1”.
Set. This first evaluation loop (46-2 to 46)
Of the song blocks whose evaluation was not finalized in -15),
The youngest block is the reference block in the next evaluation loop.
(46-3, 46-4, 46-6)
The hierarchical structure identifier of the block is 2. The same applies to
It makes sense. As a result, every block of the song
Is assigned a hierarchical structure identifier Hj. The processing in FIG. 43 corresponds to the block obtained in FIG.
Convert unit hierarchy to bar-based hierarchical data
Processing. That is, HIEa in FIG.
The hierarchical structure identifier for the bar is set. Next, referring to FIG. 44 to FIG.
The structure extraction will be described. From tonal structure to chord progression
In this embodiment, the tonality of a general music piece is
The nature of the structure is taken into account. Its nature is (B) The tonality does not change frequently in the progress of the song,
It tends to maintain tonality. (B) The constituent sounds of a chord are on a musical scale of a specific tonality. (C) When transposition occurs, rather than turning to an unrelated key
It is easy to turn into relatives such as genus or lower genus. It is. The tonality structure to be extracted has the above properties.
Therefore, in this embodiment, a tonality distance is defined between the codes,
The code of the current section is at a predetermined distance from the previous section.
In some cases, the tonality of the current section is considered to be the same as that of the previous section.
You. FIG. 47 shows an example of the tonality distance between the chords.
You. As can be seen from FIG.
The tonality distance between C (eg Am and C) is zero and
Therefore, it has the same tonality (C). Also, 5 degrees below or
The tonality distance to the chord that is completely 5 degrees above is 2 or -2
are doing. Now Key C Datonic Scale (Doremi
Fasoraside), the tonality distance of ± 2 from code C
Six of the codes C, Am, G, Em, F, and Dm that are located apart from each other
The chords of which the chord constituent sounds are all
On the tonic scale. As described later, this embodiment
Then, the tonality will be
I try to keep it. In FIG. 44, 48-1 to 48-5
In FIG. 47, the processing in
Divide the tonality distance data according to the definition of the tonality distance
I'm guessing. That is, 48-1
Key key for the first chord₁Set "0" as
Then, in 48-2 to 48-5, the following code CP
i key tones KEYi with the first code CD₁Tonality KEY₁When
Is calculated by calculating the distance. The tonality of 48-3 is
This is shown in FIG. In 50-1 [Outside 5] Is the extraction of the root data of the i-th chord CDi (see FIG. 4).
And the result is a₁And a_TwoIs assigned to While s
t is the first code CD₁Contains the root note data. 50-
As shown in FIG.₁Root sound data is 50-3 to 50-
Each time the circuit goes around the loop of No. 6, it is turned up five times._Twoof
The root data is turned down 5 degrees (the ring shown in FIG. 47).
Counterclockwise or clockwise).
50-3.₁= St holds for CDi
When the root sound data i is turned up 5 times
At -4_Two= St holds for CDi
When the data is turned down 5 times i. Therefore
For the former, put ix (-2) as the tonality distance in x.
(50-7) For the latter, ix2 is inserted into x. 5
0-9 to 50-17 are the first code CD₁And focus on
Code CDi is major or minor
Depending on whether or not x is being converted
You. For example, CD₁Is Am and CDi is Gmaj
And x = + 4 due to 50-7 (root sounds A and G
And for comparison). This is, according to FIG. 47, x = −2.
Have to become In this case, in FIG.
Go from 50-10 to 50-11, 50-13, and x = x
-6 gives x = -2. Also CD₁Is Cma
If CDi is Bmin in j, x = 50 by 50-8
It is -10. This corresponds to the definition of the tonality distance in FIG.
Therefore, x = -4 must be satisfied. in this case,
In FIG. 46, 50-15, 50-15, 50-1
7 and x = -4 is obtained by x = x + 6. No.
The calculation result x in FIG. 46 is transferred to KEYi. The processing example of 48-1 to 48-5 is shown in FIG.
This is shown in (1). Chord progression C, C, F, G₇, Bb,
F, G₇, C, as the tonal distance {KEY}, KE
Y₁= 0, KEY_Two= 0, KEY_Three= + 2, KEY_Four= −
2, KEY_Five= + 4, KEY₆= + 2, KEY₇= -2,
KEY₈= 0 is obtained. The tonality distance ΔKE thus obtained
Y｝ in the subsequent processing from 48-6 to 48-14
It is converted so as to have the tonality property described above. Sand
The tonality data immediately before is placed in the key and the current code
The tonality data is within ± 2 from this immediately preceding tonality data skey
At the distance of
Converted to gender data to maintain tonality and to distances exceeding ± 2
The tonality data of the current code
The value ± 2 is used as final tonal data. The processing side from 48-6 to 48-14 is shown in FIG.
(2). Chord progression C, C, F, G₇, Bb,
F, G₇, C, 0, 0,
0, 0, 2, 2, 0, 0 are obtained. In this way,
The tonality structure that has the desired properties for music
Data format. From 48-15 to 48-25 in FIG.
Represents the tonal structure data of distance expression
It is being converted into a sound name expression. This note name expression
Here, the number of “0” for C, the number of “1” for C #,.
A value is assigned. For example, change the first chord of the song to Cma
j, the i-th code is Fmaj, and this Fmaj
Let's assume that the tonality is “2” in the distance expression. to this
The corresponding note name expression is “5”. For this conversion,
The process goes from 48-15 to 48-16, 48-17,
Where a₁= KEY₁−KEYi × 7/2 is executed, and KEi
Y₁= 0 and KEYi = 2, so a₁= -7, 4
By the processing of 8-18 and 48-19, a₁= 5 and this
This becomes KEYi (48-20). The first of the song
If the chord is major, [Outside 6] Gives the root of the scale of the first chord section,
In the case of a minor system, according to the relationship of Am = C, [Outside 7] To get the root note of the scale of the first chord section.
(48-16, 48-21 to 48-23). In FIG. 45 (3), 48-15 to 48-2
5 shows a processing example. Chord progression is C, C, F, G₇, B
b, F, G₇, C, use in each code section
The keys of the scale (tonic) are C, C, C, C, F, F,
C, C. As will be described later, the raw
Each sound of the resulting melody was extracted by the above processing
Selected from a scale with tonality data
You. Next, the scale evaluation will be described with reference to FIG.
Will be explained. The purpose of this process is to use special codes
Is used as a scale for melody generation.
And use a special scale. FIG. 48
ISCALE is selected in the initial setting 4-1 (Fig. 2).
This is the selected scale. Code CDi is Deminish
When the code is “dim”, the scale used in that section
Combination Deminish as Le SCACEi
Set the scale and code CDi is the augment code
When "aug", the whole tone scale
And set the code CDi to 7th code “7t”
h ”for dominant 7th scale
Is set. Also, the key of the section of these chords
Instead of the tonality data obtained in the previous tongue structure extraction process.
Uses the root of the chord. Therefore, these
The first scale in sections other than exceptional code sections
The scale of the type selected in the period setting is used
Is determined by the tonality data obtained in the tongue structure extraction process described above.
Can be <Melody Generation> The apparatus of this embodiment is
Provides the basic data of the composition
After analyzing and evaluating the data internally,
Move on. Melodies shown in 4-8 of the composition general flow (Fig. 2)
FIG. 49 illustrates a schematic flow of the generation of a di. In FIG.
HIEi is the code extracted in the previous chord progression evaluation.
Hierarchical structure data for each section, 53-2 to 53-4
As shown, the hierarchical data HIEi is arpeggio data.
It is used for controlling the generation of the pattern (LL). Of this control
Details will be described later. Furthermore, the hierarchical structure data is 53
-5 and 53-6, the arpeggio pattern (L
It can also be used to control the range of L). The arpeggio pattern is the basis of the melody sequence.
A fundamental skeleton is formed, and the range of the arpeggio pattern is melody
Basically regulate the range of D. In this embodiment, the control
Uses the hierarchical structure obtained from the chord progression,
This is one of the features of the embodiment. Arpeggi
Hierarchical structure that extracts control factors of pattern from chord progression
It is not necessary to limit only to the structure information, for example, the hierarchical structure
And random numbers, and weighted sum of both data
Control the lupeggio pattern and let the user
You may enable it to be specified. In short, the composition of the user
Transform the above intentions into the arpeggio generation
Can be The arpeggio pattern (LL) is a code data
By using CDi, the pitch expression format,
Is converted to a data format (arpeggios) (53-
7). A non-harmonic sound is produced by this arpeggio.
It is added according to the action rules (53-8). this
The production rules used for adding non-harmonic sounds are mochi.
Same as those used to classify non-harmonic sounds
And therefore the melody analysis and synthesis
Reversibility holds. By adding a non-harmonic tone to the arpeggio
The melody pitch train is completed. Melody pitch line
After completion, generation of melody length sequence (rhythm pattern)
(53-9). Here, it consists of a predetermined number of notes
Basic rhythm pattern (essence generation 4-4 or essence
The length sequence determined in the sense extraction 4-6) is the initial setting.
Deformed by the pulse scale selected in Rule 4-1
Is converted to a sequence of note lengths equal to the pitch sequence of the melody.
Is done. Generation of arpeggio, addition of non-harmonic sound, duration
In the case of the flow of FIG. 49, the data is generated for each code section.
Is executed. For this reason, in 53-10,
The created melody data is moved to the continuous area. FIG. 50 shows a raw arpeggio pattern (sound pattern).
It is a detailed flowchart of generation, save, and load (Figure
49, 53-2, 53-3, 53-4). In this example
Is an arpeggio pattern based on the hierarchical structure data HIEi
The control of LL is performed as follows. First, pay attention
Is the section that is structurally different from the past section?
The hierarchical structure data of the passage of interest
It is determined by comparing with the layer structure data. Different composition
New arpeggios only for passages recognized as
A geopattern LL is generated. This generation is based on the
Performed based on the feature parameter PC of the lupe-geo pattern
It is. For passages recognized as having the same structure,
No new arpeggio pattern LL is generated. Instead,
Focus on arpeggio patterns generated in the past
Generated in the section with the same structure as the section
Use an arpeggio pattern. For example, consider a song composed of four passages.
And the structure of these four passages is a, b, c,
Let's assume it is a. The structure of the first passage is in the past
Features of the arpeggio pattern
An arpeggio pattern is generated according to the data. Likewise
And the second and third passages have new structures b and c
Arpeggio pattern or generated independently
It is. However, the last passage has the same structure as the first passage of the song.
It is. Therefore, the arpeggio pattern of the last passage
Arpeggio pattern generated for the first passage
Use the An arpeggio for a phrase with a new structure
Creating a new pattern means this new structure.
It means that another motif arises from the style passage.
Now, the arpeggio created for the first bar of the passage
Repeat use of the turn for subsequent measures of the passage
If you do, you will notice a motif that is one bar long.
Generally, the length of a motif is one to several measures,
The length of the motif often changes. In the example of FIG.
Taking this into account, it is assumed that a new structure passage has been detected.
The length of the motif in the passage is 1 or
Two measures can be used. Two bar motif selected
When selected, the first and second measures of the passage are independent.
An arpeggio pattern is generated at the beginning, followed by an odd number
Measures use the arpeggio pattern of the first measure,
The even-numbered bar that follows is the arpeggio pattern of the second bar.
Use Reference to Hierarchical Structure Data in Past Section
, Repeat arpeggio pattern in the past section
For this purpose, a sound (LL) data buffer is prepared. sound
FIG. 51 shows an example of the type data buffer. Referring to FIG. 50, it is easy to write 54-1.
Bar count within a bar (section for each barno shown in FIG. 41)
Is set to "1". 54-2 shows the hierarchical structure of the current bar.
Structured data HIEi is converted to the hierarchical structure data HIE of the immediately preceding bar.
Compare with i-1 to see if it is the beginning of a passage.
Click. For example, | HIEi-HIEi-1 |
The time when it is established is the start of the passage. The start of the passage is determined
Resets the bar counter in the passage to "1"
(54-3). Then search the sound data buffer
The current passage should generate a new arpeggio pattern.
It is checked whether it is a node (54-4). Sound data buffer
Is performed as follows. First, the sound pattern data
The data at address 0 of the buffer (how many sets of
Data that indicates whether it has been generated)
Address 1 to N is an address.
And then read the data (header information of each sound type) sequentially.
The hierarchical structure data indicated by the upper 8 bits is assigned to the level of the current bar.
Compare with the structure data HIEi. In sound type data buffer
The hierarchical structure data similar to the hierarchical structure data HIEi of the current bar
Data (for example, the same value as HIEi or (HIEi-
When there is no data with the value of 1), the current bar is arpeggiated.
This is the first measure of the passage for which a new pattern is to be generated.
You. If a header with similar hierarchical data is found
When the following arpeggio pattern is
Load as pattern. For new passages,
The number of bars in the motif is determined (54-
5). This determination can be realized by random number generation, for example. 2 bars
If it becomes a motif, set flag fl to “1”
(54-7, 54-8) and the next bar (the second
A new arpeggio pattern is generated for
To be And generate an arpeggio pattern
(See FIG. 52), and saves the data in the sound pattern data buffer (5
4-9). In detail, HIEi x 0100 + measure count
The header is created based on the value of data × 0010 + the number of measures in the motif.
Of the sound pattern at address 0 of the sound data buffer.
The number N of pairs is incremented, and the header N is added to the address N.
Write the address, and from that address, header, LLN
O (length of generated arpeggio pattern), LL₁, L
L_Two... LL_LLNO(Data of generated arpeggio pattern
Write). After that, other melody generation processing
(54-10) and increment the bar counter
(54-11). The phrase is not changed in 54-2.
Is found, the flag fl is checked (54).
-13). When fl = 1, the current bar is a two-bar mochi
Arpeggio because it is the second measure of the passage that generates the
Generate the pattern again and load it into the sound data buffer.
(54-15), the flag fl is reset to "0"
(54-16). When fl = 0, the hierarchical structure of the current bar
Search for buffer corresponding to data HIEi from buffer
And the length information of the motif indicated by the header is one measure
At that time, load the following sound pattern data into the header,
If the length of the motif is two measures, the remainder of the measure counter is 2
Compare the remainder with the measure number of the header, and if they match, the header
And the arpeggio pattern of the current bar
And load. The processing is executed at 54-9 and 54-15 in FIG.
Fig. 52 shows the details of generating arpeggio patterns (sound patterns).
Show. The ckno of 56-1 indicates the number of chord constituent sounds.
You. The number of chord constituent sounds is 16 bits of chord constituent sound data.
By counting the number of "1"
(See FIG. 26). In the example of FIG.₁~ PC_FiveRaw
Arpeggio pattern control parameters
You. r₁Takes a random value in the range of 1 to ckno and codes
It means the constituent sound number (56-4). r_TwoIs PC_Three(Al
The lowest note of the pegio pattern) ~ PC_Two(Arpeggio putter
Octave chord random number value
(56-5). a =
r₁+ R_TwoCalculate LL candidates generated by × 0100
(56-7), when this candidate a satisfies the condition of PC
Candidate a is adopted as LL (56-8, 56-
12, 56-14). However, depending on the value of PC,
LL to execute (for example, first LL₁) Is decided,
Subsequent LL_TwoNo longer satisfies PC requirements
there is a possibility. For example, PC_Two= 501 (of the arpeggio
The highest note is the first chord component of the fifth octave), PC_Three
= 401 (The lowest note of the arpeggio is the first of the fourth octave
Chord composition sound), PC_Four= 3 (maximum difference between adjacent LLs)
The value is for 3 chords), PC_Five= 3 (adjacent LL
Is LL when the minimum value of₁When
Then LL₁= 403 (the first LL is the fourth octave
Assuming that the first chord component sound is obtained, the PC_Four, PC_Fiveof
To meet the conditions, LL_Two= 503 or 303 (composition
When the number of sounds is 3), this is the PC_Two, PC_ThreeOn the condition
Do not fit. To prevent this, the loop counter LOO
Prepare a PC and set the loop counter LOOPC to a certain value (example
For example, when it becomes 100) or more, the candidate a is forcibly set to LL.
(56-9, 56-10, 56-1
1). FIG. 53 shows details of the check 56-8 in FIG.
It is. In order for LLi candidate a to satisfy the condition of PC, (B) a ≦ PC_Two(Below the highest note) (B) a ≧ PC_Three(Must be at least the lowest note) (C) | a-LLi-1 | ≤PC_Four(Difference from previous LL
Is the maximum value PC_FourBelow) (D) | a-LLi-1 | ≥PC_Five(Difference from previous LL
Is the minimum value PC_FiveAbove) Must be established. In the flow of FIG.
Flag OK is set to "0" when the condition of
(Olda in the figure represents the immediately preceding LL. 56
-13). FIG. 54 shows the details of 53-7 in FIG.
The purpose is indicated by (octave number + chord number)
The format of the arpeggio pattern LL
Using data cc (octave number + note name number)
Melody pitch data format and convert it to MEDi
It is to store. The processing of 58-5 and 58-6 is L
Li chord composition tone number [Outside 8] Is greater than the number of chord components (CKNO) of the chord in the current section
At the time, the code number of the LLi chord
Change to the highest chord component number
Processing. In the figure, c is a counter of chord constituent sounds, [Outside 9] Is the octave number of LLi, and j is the note name counter
You. FIGS. 55 and 56 are the same as those in FIG.
7 shows details of the addition of a non-harmonic sound. The purpose of this process is
Add the desired non-harmonic sound to Pesio to create a melody pitch train
It is to be completed. To add a non-harmonic tone,
Of non-harmonic tones {RSi}, tones obtained by chord progression evaluation
Gender structure {KEYi}, expresses knowledge to classify non-harmonic sounds
Production rules are used. Discord added
The voice must meet the following conditions: (B) Sound within the specified range (B) A schedule with KEYi as the main tone obtained in the chord progression evaluation
Sound (C) Sound outside the chord configuration (D) Conclusions and planned non-
Match with harmony identifier RSi In FIG. 55, 59-4 to 59-18
The outer loop is the number of planned non-harmonic identifiers RSi
It is a loop that repeats
The loop repeats the number of arpeggio notes. 59-8-5
In 9-14, as a candidate for a non-harmonic tone,
Each pitch data k within the range up to the limit up is sequentially examined.
(See FIG. 57). If the pitch data k is a scale tone and
(59-8, 59-9), the sound
Calculate the number F (59-10) and get the production rule
(59-11), and the conclusion is
Check whether the selected non-harmonic tone identifier RSi matches
(59-11). When they match, the pitch data k
Satisfies all of the above conditions for non-harmonic sounds
You. Therefore, the non-core that counts the number of added non-harmony
The tone tone counter nctct is incremented and checked.
The pitch data k of the found non-harmonic tone is input to VMnctct.
And set the non-harmonic sound addition position j to POSTnctct
And the associated flag fl_jIs set to “1” (59
-19 to 59-22). In this example, at most one
Non-harmonic sound is added, and fl_j= 0
Is the harmony MEDj and MED_{j + 1}A non-harmonic sound is still in between
Indicates that it has not been added. Conclusion at 59-12 = The condition of RSi is
If not established, the pitch data k of interest is a non-harmonic tone
Does not satisfy the condition as
Is incremented (59-13), and the process is repeated. 59-
14, when k> UP is satisfied, the harmony MED
j and MED_{j + 1}That no non-harmonic sounds were added during
Means Therefore, j is incremented (59
−15), whether non-harmonic sounds can be added between the next harmonic sounds
Proceed to inspection. Details of the setting of candidate sounds in 59-6 are shown in FIG.
At 58 is shown. In this example, the front and rear chords MEDi,
MED_{i + 1}Lower than 5 semitones higher than the higher sound
The search range is up to 5 semitones below (6
2-5 to 62-7). However, when i = 0,
Before the first harmony of the section of interest,
Is to be added, 5 semitones above the first harmony
The pitch range is set to up to 5 semitones below (62-1, 2), i =
In the case of Vmedno, that is, the last sum of the section of interest
When trying to add a non-harmonic sound after the voice,
Pitch range from 5 semitones to 5 semitones below
(62-3, 4). Whether the pitch data k in 59-8 is a scale tone
The details of the check are shown in FIG. SCA in the figure
LEi represents the type of the scale used in the section i.
Address pointer of scale data memory 5 as shown at 0
It has become. 12-bit length scale at this address
Data * SCALEi was obtained by the above-described chord progression evaluation
Turn by KEYi (65-2). For example, SCAL
If Ei is “0” (diatonic scale)
Data represents Doremifasoraside with C as the tonic.
When KEYi is "5" (F), five data turns
By this, it is converted to scale data a with F as the tonality
You. 65-3 is pitch data k (indicated by MD in the figure)
Is converted to the same data format as the scale data.
And the logical product of the result b and the scale data a is "0".
It is concluded that the pitch data k is not a scale sound, and the logical product is
If it is not "0", it is concluded that the scale is (64-5-6).
4-7). The details of the calculation of F in 59-10 are shown in FIG.
It is shown in FIG. In this example, there is one non-harmony between the previous and next
Is added, so some functions (in the figure,
F₁~ F_ThreeIs set to a predetermined value. 59
Details of the forward inference of −11 are shown in FIG. 40 described above. When the processing of FIG. 55 is completed, nctct
Stores the total number of added non-harmonic tones, and the array {V
The ith of｝ is added to the ith in the processing of FIG.
Pitch data of the selected non-harmonic tones are stored in the array ｛POS
The i-th of Ti｝ is added to the i-th in the processing of FIG.
The position information of the selected non-harmonic sound is stored. These data are obtained by the processing of FIG.
Is converted to the melody pitch sequence {VMEDi}
You. The sequence {VMEDi} is the arpeggio {MED
i｝ is initially set. 60-2 to 60-9 are added
Array {POSTi}, {VMi} in order of position
This is the process to be performed. In 60-10 to 60-19,
Non-harmonic pitch at the position indicated by the position data POSTi
Data VMi is inserted. In this example, one non-harmonic sound is added between harmonic sounds.
Only one non-harmonic sound can be added.
The processing may be changed so that it can be performed. So far, Melody
The row of pitches is completed. The rest of the processing is
Generation. FIG. 62 shows a flow of generating a melody note length sequence.
Is shown (details of 53-9). First, essence generation or
Pays attention to the number of notes of the reference rhythm pattern obtained by extraction.
Note number Vmedno (melody pitch)
To calculate the difference a between the two.
(66-1). The number of generated notes is
When the number is smaller than the number of notes (when a> 0), the pulse schedule
The best combination of notes by the number of differences
(66-2 to 66-6). The number of generated notes is
More than the number of notes in the rhythm pattern (when a> 0)
Performs the optimal division of the note repeatedly by the difference (6
6-7 to 66-11). In this example, the rhythm pattern data
16-bit data is used as the data format.
A position with a value of “1” is assigned to each timing.
This indicates that sound is generated at the
(66-12). FIG. 63 shows the details of the combining process. In the figure, P
SCALEj is the jth component of the pulse scale used
Where RR is the rhythm pattern to be processed
is there. Pulse scale when RR bit is “1”
By setting the point with the smallest weight to “0”, notes are connected.
Combine. For example, if the reference rhythm pattern is [Outside 10] , The normal no pulse scale (see FIG. 11)
)), The result of combining one note is [1] This is obtained as follows. First,
RR initially 0001 0001 0101 0001 It is. On the other hand, the normal pulse scale is 1213 1214 1213 1215 It is. Normal pulse in bit "1" of RR
The point with the smallest weight on the scale is number 7 from the right end.
Eye position. The bit at this position becomes “0”. I
Thus, the resulting RR is 0001 0001 0001 0001 And this is ## EQU1 ## FIG. 64 shows the details of the division process. The split is
When the RR bit is "0", the pulse scale weight is
This is executed by setting the maximum point to “1”. example
Rhythm pattern [Outside 11] On the other hand, if you divide the note with the normal pulse scale
And the result is [Outside 12] Becomes Details of Conversion to MER Data Format 66-12
The details are shown in FIG. In the figure, c₁Is a note counter
And c_TwoIs a counter for measuring the duration of each note. This
In the above example, the first “1” appears from RR in MERo
Spans the boundary (bar line) of the section
Notes can also be processed (measures against syncopation). FIG. 66 shows the details of data movement 53-10.
You. First, MERo (the blank part at the beginning of the current generation section)
Is added to the duration data MELRmeldno of the last note that has been generated.
I can. Here meldno shows the number of notes already generated.
I forgot. The pitch sequence VMED generated this time₁~ VMEDvmedno
Is moved to the MELD, and the duration string MER generated this time is moved.₁~ M
Move ERvmedno to MELR (70-2 to 70-
6). After updating meldno, the process ends (70-7). <Summary of Composer> It is clear from the above description.
Thus, this device as an automatic composer has various features.
Some of them are shown below. (A) Using music knowledge in melody analysis and synthesis
A production system that performs inference
You. (B) a process of classifying non-harmonic sounds included in the melody;
The same process as adding non-harmonic sounds to arpeggios
Based on production rule data
Therefore, both processes are reversible. (C) Analyze the chord progression given as the composition material
By extracting the hierarchical structure and tonality structure of the song,
Is planned. (D) The extracted tonal structure is the key of the scale used in each section.
Stipulate. This results in a natural sound with a rich tonality
Songs are guaranteed. (E) The extracted hierarchical structure is used for controlling arpeggio generation
Is done. This adds variety and consistency to the resulting song
Can be included. The above-described embodiment is merely an example.
Various modifications, changes, and improvements are possible. For example, on
In the described embodiment, the arpeggio pattern extracted from the motif
The feature PC is used to generate the arpeggio pattern LL.
Is used as control data.
May be changed as the music progresses.
This is, for example, using the progress position of the song or the hierarchical structure as variables
It can be realized by means of calculating functions. Similarly, regarding the characteristic {RSi} of the non-harmonic sound
However, it can be changed as the music progresses. example
One of the characteristics of non-harmonic tones extracted from motifs
Is replaced with another non-harmonic identifier. This
This is randomized from a set of non-harmonic identifiers.
This can be realized by selecting one. Further, with respect to the rhythm,
Rhythm control by combining and dividing notes with scale
The dominant miniri included in the motif
Extracts the rhythm pattern and generates the melody's duration
You may make it incorporate in a row. [0090] As described in detail above, the present invention
According to the chord progression, the tonal structure inherent in the song is extracted
can do.In particular, each chord section of chord progression
Chord progression history along the time axis of chord progression
To determine the key of a certain section
There is no need to refer to the key of the following section above.Also, extraction
By using the generated key information to generate music,
Natural music that does not deviate can be generated.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is an overall configuration diagram of an automatic music composer according to an embodiment of the present invention. FIG. 2 is a flowchart showing an overall operation in a composer mode. FIG. 3 is a diagram showing a list of main variables used in processing. FIG. 4 is a diagram showing a format of data used. FIG. 5 is a diagram showing a format of data used. FIG. 6 is a diagram showing a format of data used. FIG. 7 is a diagram showing a format of data used. FIG. 8 is a diagram showing a format of data used. FIG. 9 is a flowchart of initialization. FIG. 10 is a diagram showing an example of chord progression data stored in a chord progression memory. FIG. 11 is a flowchart of reading chord progression data. FIG. 12 is a diagram illustrating pulse scale data stored in a pulse scale memory; FIG. 13 is a flowchart of reading pulse scale data. FIG. 14 is a view exemplifying production rule data stored in a production rule data memory. FIG. 15 is a flowchart of reading production rule data. FIG. 16 is a diagram illustrating melody data (motif data) stored in a motif memory; FIG. 17 is a flowchart of reading melody data. FIG. 18 is a flowchart of essence generation. FIG. 19 is a flowchart for setting characteristics of an arpeggio pattern. FIG. 20 is a flowchart for setting characteristics of non-harmonic sounds. FIG. 21 is a flowchart for evaluating the rhythm of a motif for each section. FIG. 22 is a flowchart for calculating Ps, Pe, Pss, and Pee. FIG. 23 is a flowchart for calculating Ps and Pss. FIG. 24 is a flowchart for calculating Pe and Pee. FIG. 25 is a flowchart for extracting an arpeggio pattern of a motif. FIG. 26 is a diagram showing an example of constituent sound data of each chord. FIG. 27 is a flowchart for generating constituent sound data from chord data. FIG. 28 is a flowchart for extracting features of an arpeggio pattern. FIG. 29 is a flowchart for extracting characteristics of non-harmonic sounds. FIG. 30 is a flowchart for classifying melody sounds into harmony sounds and non-harmony sounds based on codes. FIG. 31 is a flowchart for calculating a function F representing a melody situation to be analyzed; Flow chart of the calculation of Figure 32 functions F _1. Flow chart of the calculation of Figure 33 functions F _2. Figure 34 is a flowchart of the calculation of the function F _3. Figure 35 is a flowchart of the calculation of the function F _4. Figure 36 is a flowchart of the calculation of the function F _5. FIG. 37 is a flowchart of calculating a function F ₆ . FIG. 38 is a flowchart for calculating functions F ₇ and F ₈ . FIG. 39 is a flowchart for temporarily storing a calculated function F; FIG. 40 is a flowchart for inferring types of non-harmonic sounds. FIG. 41 is a flowchart for calculating the degree of coincidence of chord progression between blocks. FIG. 42 is a flowchart for generating hierarchical structure data from the calculated degree of coincidence. FIG. 43 is a flowchart of converting hierarchical structure data of a block into hierarchical structure data for each code section. FIG. 44 is a flowchart for extracting a tonal structure from a chord progression. FIG. 45 is a view exemplifying a tonal structure extraction process. FIG. 46: First code CD ₁ and i-th code CDi
9 is a flowchart for calculating a tonality distance between the two. FIG. 47 is a diagram showing a definition of a tonality distance between chords. FIG. 48 is a flowchart showing scale (scale) processing. FIG. 49 is a flowchart of melody generation. FIG. 50 is a flowchart showing generation, saving, and loading of a sound pattern (arpeggio pattern); FIG. 51 is a diagram illustrating a sound data buffer; FIG. 52 is a flowchart of arpeggio pattern generation. FIG. 53 is a flowchart of a check in arpeggio pattern generation. FIG. 54 is a flowchart for converting an arpeggio pattern into a melody data format. FIG. 55 is a flowchart for adding a non-harmonic sound to an arpeggio. FIG. 56 is a flowchart for adding a non-harmonic sound to an arpeggio. FIG. 57 is a view showing the order of non-harmonic sound addition processing. FIG. 58 is a flowchart for setting a range of pitches that are candidates for non-harmonic sounds. FIG. 59 is a flowchart for calculating a function F; FIG. 60 is a diagram showing an example of scale data stored in a scale data memory. FIG. 61 is a flowchart of identification of scale sounds. FIG. 62 is a flowchart for generating melody tone length data. FIG. 63 is a flowchart of the optimal combination of notes. FIG. 64 is a flowchart of optimal note division. FIG. 65 is a flowchart for converting a generated rhythm pattern into a MER data format. FIG. 66 is a flowchart of moving generated melody data to a continuous area. [Description of Signs] 1 CPU 4 Chord progression memory CDi i-th code KEYi Key in i-th code section

Claims (1)

(57) [Claims] A chord progression providing means for providing a chord progression, the tone in the interval of each chord in the chord progression that is applied from the chord progression providing means, along the time axis of the chord progression sequentially according provenance of chord progression to the code A tonality determination music device, comprising: a tonality determination means for determining. 2. Chord progression imparting means for imparting a chord progression; and key judging means for sequentially judging the key in each chord section in the chord progression imparted from the chord progression imparting means according to the history of the chord progression along the time axis of the chord progression. The key determining means includes: (a) a leading key determining means for determining a key of a leading section to which a leading code of a chord progression belongs; and (b) each code other than the leading code of the chord progression is referred to as a leading code. Tonality distance calculating means for comparing and comparing the comparison result for each code as a difference (referred to as tonality distance) between the key of the section of each code and the key of the first section; (c) calculating for the code of the current section The tonality distance thus determined is compared with the tonality distance determined for the code in the immediately preceding section,
Determining means for determining the tonality distance of the current section by changing the tonality distance calculated for the code of the current section according to the comparison result; and (d) determining the tonality distance of the determined current section. Conversion means for converting to a key based on the key of the leading section to determine the key of the current section, and a tonality determination music device.