JP2759110B2 - Automatic composer - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [0001] The present invention relates to a music device,
In particular, automatic songs that compose melodies
Belongs to the machine. [0002] 2. Description of the Related Art Weights to be considered regarding the quality of an automatic composer.
One of the key elements is that humans have been familiar with
Songs, in other words, songs that are not purely mechanical but rich in musicality
Whether the composer has the potential to generate
That is. For example, Japanese Patent Application No. 56-1256
No. 03 (Japanese Patent Publication No. 60-40027) has a series of pitches.
Individual pitch data from data (for example, 12-scale data)
Data is randomly sampled and the sampled data is limited
If the above conditions are satisfied, take it as a melody note.
If the conditions are not met, the melody note
If you do not make a selection and re-sample, repeat the condition check
A self-contained composer is disclosed. Therefore, this self
The melody generation process of the motion composer is basically trian
Error mode. Randomly sample pitch data
At that stage, a row of completely disordered pitches is created. This
In a row of chaotic pitches,
Infeasible (good melody due to astronomical contingencies)
May be possible). So what is this disorder
In order to bring about the order, a kind of conditional inspection
We do filtering (sorting). In this case,
Degree is an important factor. If sorting is too tight,
Melody will be one pattern and forgive
If so, the original disorder would be dominant. above
The automatic composer is called a melody
Rather than compose a melody that is elusive in style
It is mainly used for listening and training.
It is effective as a composer.
Musical scores and performances are generally difficult). In this sense, I'll give it to the beginning
Ability is not combined. The present invention is exactly this ability
Is considered. [0003] That is, the object of the present invention
The goal is to maintain the concept or essence of the song,
Can compose songs that can change in various ways according to the line
It is to provide an operation music machine. [0004] According to the present invention, there are provided plural
Melody over each music section
In the automatic music composition machine generated by
Chord progression hand that gives chord progression information to
The motif feature parameters that characterize the steps and motifs
Change the motif feature parameter generation means and tune
Change that generates a change parameter for transformation
Parameter generating means, the motif characteristic parameter,
Input the above-mentioned musical composition change parameters and music segment position information
Variables, and based on these input variables,
Feature the section melody to be generated in each music section
Section melody that generates section melody feature parameters
Feature parameter generating means and the section melody feature
Using the parameters and the chord progression as input variables,
Section melody in each music section based on these input variables
Section melody generating means for generating a musical note
An automatic composer is provided. [0005] According to this configuration, the motivation of music is mochi.
Motif feature parameters that characterize music content
(PA) is generated by the motif feature parameter generation means.
Tune change parameters to change
(PB) is generated by the fantasy change parameter generating means.
Motif feature parameter, motif change parameter and
The position information of the music section (eg, bar number) is used as an input variable.
, And based on these input variables,
Music content of the section melody to be generated between
Parameter (interval melody feature parameter)
Meter PC) as the section melody feature parameter generation means
Has been raised. Therefore, the section melody feature parameter
The time series for each section of the meter varies depending on the music progress.
Can be changed to Further, a section melody generating means is provided.
The section melody generating means is adapted to perform the section melody feature
The parameter and the chord progression from the chord progression giving means
Each music is used as an input variable based on these input variables.
Generate a section melody in a section. As a result,
A melody that can be changed in various ways is generated (comp
o)). Because the section melody feature parame
The time series property of the data is the time series of the section melody (melody
This is reflected in the flow of Also, section melody
The feature parameter generation means uses mochi as one of the input variables.
Since the feature parameter is used, the output variable
The inter-melody feature parameter is the motif feature parameter
to be influenced. Therefore, the melody of each section generated
The effect of the motif characteristic parameter is reflected on the motif
A melody with characteristic parameters as the essence of music
Is done. In addition to such a basic configuration, the
It would be more desirable to be able to reflect
New To this end, in one configuration example, the input from the user
Section specifying means for specifying a specific music section according to
Input from user for specified music section
According to the music content of the section melody in the music section
Set the section melody feature parameters that characterize
Inter-melody feature parameter setting means.
The section melody generating means performs the section melody feature parameter.
Section melody feature parameter set by meter setting means
From the section melody feature parameter generation means
Identification specified using instead of melody feature parameters
Feature to generate section melody in music section
And Preferably, the section melody feature parameter
Data consists of multiple types,
The meter setting means sets the type for the specified music section.
The section melody feature parameters are set for each class. Preferably, the section melody feature parameter
Data setting means (a) for the specified specific music section.
And the section melody feature parameter generation means
Display the data message of the section melody feature parameter
Display means for displaying, and (b) the displayed section melody feature
Parameter data message from user
Receiving means for receiving a parameter change input for each type;
Have. In a preferred embodiment, the above-mentioned section melody is used.
The de-feature parameter setting means further comprises:
Melody features from multiple feature parameter generators
Each parameter value to the corresponding subjective parameter
Data message and display the data message on the display
Message containing the value of this converted subjective parameter as
Conversion means for displaying the page, and the user
The modified value of the subjective parameter input as
Inverse conversion to the corresponding section melody feature parameter correction value
And inverse conversion means. In one configuration example, the section melody is
The feature parameter generation means uses multiple types of section melody features
It consists of means for generating parameters, each type of section
The melody feature parameter is obtained by the section melody generating means.
In the music content of the section melody that is the result of the action
It functions to control the corresponding musical feature. [0011] Preferably, the music content of the section melody is included.
Learn user preferences for individual musical characteristics
The learning means further comprises: (a) composer
Output the resulting melody over multiple music segments
Output means, and (b) a specific music section as a data item
Interval, the specific section melody feature in the specific music section
Type of section parameter and section melody feature
User receives a correction data record containing the parameter's correction value.
Correction data record created according to correction instruction input from
And (c) a file for the corrected data record.
File storage means for storing the file, and (d) composing again
When performing the above, the section melody feature parameter generation means
Each correction data in the correction file storage means rather than output
Records are prioritized and written on each modified data record
Also, it is possible to modify specific types of characteristic parameters for specific music sections.
The positive value is used by the section melody
And priority control means for generating [0012] BRIEF DESCRIPTION OF THE DRAWINGS FIG.
An embodiment will be described. <Overall Configuration> Overall circuit configuration of the automatic composer according to the present embodiment
The result is shown in FIG. In the figure, 1 is an input device, 2 is a code configuration
Sound memory, 3 is chord progression memory, 4 is motif memo
Re, 5 is parameter B memory, 6 is CPU, 7 is work
Memory, 8 is parameter C memory, 9 is learning data memo
Re, 10 is melody data memory, 11 is monitor, 1
2 is a CRT, 13 is a staff printer, 14 is a tone formation
15 is a sound system, 16 is an external storage device.
You. The motif memory 4 is provided by the input device 1
Stores information of input motif (input melody)
By the way. The motif information contains pitch and pitch (pitch) data.
Data columns. During automatic composition, CPU 6
Parameters that characterize the motif from the motif information
(Motif characteristic parameters) are extracted. The chord progression memory 3 has a chord name.
The chord progress information represented by a sequence of programs is stored. Co
The user can sequentially input code progress information from the input device 1.
You can also specify
In response to the setting (for example, the specification of the music format), the CPU 6
The chord progression may be automatically generated. Chord
Automatic line generation can be performed, for example, using basic code patterns (multiple
Concatenation, or permitted code
Is possible by the interconnection of
For example, a Markov chain model can be used. However
And the chord progression is specified directly by the user
Or automatically generated by the machine according to the present invention.
It is not important. The chord structure sound memory 2 stores various chords.
Constituent sounds (chord member pitch data) are stored.
In the case of this example, each address of the code progress memory 3 is used.
From the chord configuration sound memory 2
The storage area of the specific chord sound data is specified
It has become. The CPU 6 uses a code
At each change timing (for example, every bar),
The address of the memory 3 is advanced, and the contents
Calculate the address on chord composition sound memory 2 from the donated name
Then, each pitch data constituting the chord is read. On the other hand, the parameter B memory stores
Parameter B is stored for automatic operation.
At the time of music, the CPU 6 sets the parameter B
Feature parameters and music progress section variables (eg, bar number
) Is generated for each traveling section.
You. The generation of the parameter C will be described in detail later.
It has the property of controlling or characterizing the melody to be generated.
ing. The generated parameter C is a parameter C memory
Is stored in In the work memory 7, the CPU 6 stores an automatic song.
Intermediate data generated in the process
(Melody data during construction) is stored. Melody
The melody data that constitutes the completed song is stored in the data memory 10.
Data is stored. The completed song can be
Output to the monitor 11. For example, the tone formation times
You can listen to the music through Road 14 and Sound System 15
it can. In addition, a copy of the musical score is printed by the staff printer 13.
Obtainable. Through the monitor 11, the user partially
You may want to modify the song. In this embodiment
In such a case, through the CRT 12 and the input device 1,
User can request a fix and
Modifications are performed in an active manner. Corrected data
Are stored in the learning data memory 9 as knowledge. Later self
The CPU 6 uses this knowledge to play a melody
Generate The external storage device 16 stores the completed music
Copy-ups, learned knowledge, other transcripts,
Or, use it as a resource for an alternative automatic music program
Is done. <Automatic music composition function>
The overall function of the motion composer will be described with reference to FIG.
I do. In the figure, reference is made to a code starting with the initial letter I.
Blocks indicate information or information sources. example
For example, I1 is motif information, and in FIG.
This is information stored in the memory 4. I2 is a parameter
B, which is in the parameter B memory of FIG.
Information. I3 is chord progression information,
Is information given from the chord progression memory 3 of FIG.
You. I4 is the generated melody, which is the melody of FIG.
The data is stored in the rody data memory 10. On the other hand, initial F
The block indicated by the sign starting with
Is shown. As shown in the figure, the main function is
Motif evaluation function F1 for evaluating information
Motif parameters to extract motif parameters from
Extraction function F2, motif parameter extraction function F2
Based on motif feature parameters and chord progression information,
There is a melody generating function F3 for generating a ody. Further
Then, the generated melody I4 is monitored by the user.
Study to correct and learn the necessary parts through D1
Learning function F4 and parameter changing function F5 by learning
Contains. More specifically, the evaluation function F1 is used in this embodiment.
Now, extract the non-harmonic sound contained in the motif
Extraction function to extract various non-harmonic sounds
Can be. In this case, the preemptive sound extraction 21, the excitement extraction 22,
Embrace sound extraction 23, elapsed sound extraction 24, ornament sound extraction 25
And other non-harmonic sounds extraction 26
Is extracted. The non-harmony extracted by each of the extractions 21 to 26
The sounds include "pre-emption sound" and "pure sound".
However, the meanings of "preemptive sound" and "pure sound" here are based on harmony
It is not necessary to be completely the same as "pre-emption sound" and "biasing sound".
Another good thing is that the name of each non-harmonic sound in harmony
There are some differences between musicians, and the definition
Changes depending on the genre of music and the flow of the times.
Is not as ambiguous as natural language
And cannot meet the required definition on a computer.
Therefore, to be precise, the preemptive sound extraction 21 is a first type non-
Extracts harmony sounds (non-harmonic sounds that can be regarded as preemption sounds)
The function 22 has a function to extract a second kind of non-harmonic sound.
In the following, the same applies to the third and fourth types.
That is. As the last function in the evaluation function F1, HD
i (motif data) contains constants (non-
A function 27 for substituting a harmony identifier) is shown,
This function is built into each non-harmonic sound extraction function 21-26.
Can be taken. The above-described evaluation function F1 is based on the motif information I1.
Extract the parameters that characterize the motif from
Is a pre-processing function of
It can be considered as a part of the data extraction function F2. According to FIG. 2, a motif parameter extractor
No. F2 is the evaluation result of the motif evaluation function F1 (in this case,
Where the motif is what kind of non-harmonic sound
Motif data with non-harmonic identification information)
Extract the parameters that characterize the motif. In this figure
, The motif parameter extraction function F2 includes
Function 31 for extracting the number of harmony sounds contained in the motif
Function 32 for extracting the total number and the total number of non-harmonic sounds
Function 3 for extracting type parameters of harmony (dispersion chords)
3, and extract the parameters of the smoothness of the motif
Function 34 is included. Here, the motif evaluation function F1 and the mochi
The “extraction section” executed by the parameter extraction function F2
The position will be briefly described. The unit of this extraction section is
This is a predetermined section of the chief (input melody). For example,
Motif (input melody) changes in units of one measure
If you can think of having phrases that follow the chord progression
In most cases, this is true).
It can be a unit. In such a case, the motif is
Assuming that it has a length of N measures, the first, second,... Nth measure
For each section of, evaluation and extraction by functions F1 and F2
Is performed. For the sake of convenience, in the following description,
Unless you know, the length of the motif is one bar, and
The function F2 is the first measure of this song, and the function F2
Parameter (characteristic PA described later)
j) shall be extracted. On the other hand, the melody generating function F3 also
The "section" (the "progress section" of the song) that controls the generation of the lody
The concept exists. For example, if the chord progression information I3 is one bar
(For example, the measure i is C, i + 1
Measure is C, i + 2 is F, i + 3 is G)
Then, "length of one bar" is set as the "progression section" described above.
Can be used. In the melody generating function F3, the parameter
The C calculation function F31 outputs a packet in units of the “progression section”.
The parameter C is to be generated. Parameter C is above
Mochi given by the motif parameter extraction function F2
Properties depending on the characteristic parameters and the "progress section" (example
For example, the bar number). Mathematically
In other words, parameter C is PC, motif feature parameter
Is PA, and the number of the “progress section” of the song is i,
Meter C can be represented by PC = f (PA, i)
You. The relationship PC = f (PA) is expressed by the motif parameter P
The essence of the motif characterized by A is a song
Suggest that it is reflected throughout. PC =
The relationship of f (i) is determined by the parameter C
Is determined (assigned).
The parameter C generated by the parameter calculation function F31 is
As shown in FIG. 2, the remaining key of the melody generation function F3 is
Prime, ie, distributed chord generation function F32 and non-chord addition
A parameter that is provided to the function F33 and controls the generation of a melody.
Used as a meter. In other words, parameter C is
Parameters that control or characterize the generated melody
(Generation melody feature parameter). In FIG. 2, a parameter C calculation function F
31 is the above motif feature for calculating the parameter C
The parameter B (see FIG.
1) also uses the information stored in the parameter B memory 5).
It has become so. Parameter B memory 5 stores
In terms of the formation of data C, as "data compression"
It works. That is, the parameter C is formed.
Configuration (generation melody feature parameter generation means)
By the chief feature parameter PA, the parameter C
Data base {f (i)} that matches PA
Is possible, but then, for each i
Requires very large storage capacity to provide values
You. In short, in the case of this embodiment, the parameter B memory
5 is an element of the generation melody feature parameter generation means.
You. The distributed chord generation function F32 is a
This is where the disperse chord is generated from the loddy. That
Read function 41 of chord constituent sound, one of the constituent elements
Is a chord component sound memory according to the chord progression information I3.
2 (Fig. 1) to read each component sound of the corresponding chord (chord)
Is about to emerge. The next element, the conversion of chord components
The round function 42 reproduces the chord constituent sounds read out by the function 41 described above.
Parameters determined for each progress section (eg, bar section)
Code C in accordance with the data C
It is about to run times. This function 42 is mainly
Controls the range within each advancing section. In addition, distributed chord generation
In function F32, MEDi (harmonic sound) is determined from the immediately preceding sound.
Function 43 and the type (pattern) of the distributed chord
Functions 44 and the like. The distributed chord generation function F32
The generated chords protect the chord progression.
In addition, for each traveling section,
Has a pattern. The non-harmonic sound adding function F33 is provided with
Non-harmonic tones between the dispersed chords given from the generating function F32.
Is to be placed and added. In the illustrated example,
Addition function 51, elapsed sound addition function 52, swelling sound addition function
53, a decoration sound adding function 54; Here, too,
The names of “elapsed sound”, “swing sound” and “decorative sound” are
This is for convenience, and more precisely, the first in melody generation
It can be called the seed, the second kind, the third kind, and the fourth kind of non-harmonic sound.
Wear. One important point is that each function 51-54
Non-harmonic addition rules defined within these functions
In accordance with the parameter C given for each traveling section.
It is to use. For example, parameter C determines
In the case of prohibition, the biasing function 51
At this time, no bias is added. In addition, the number of impulses
It can be controlled by the meter C. That is, the functions 51 to 5
4 is a non-harmonic tone controlled by the parameter C.
Execute. But avoid randomness, random elements
Is freely and preferably introduced. Random (unusual
D) The introduction function is grouped in the additional functions 51 to 54 for each non-chord.
Or the parameter C calculation function F3
1 or a suitable
You can also perform a trade-off (assignment of random introduction function)
No. In the block F33 for adding a non-harmonic tone in FIG.
Also shows the functional block of the tone length modification 55. Length correction
Function 55 determines the length of each progression section (eg one bar)
To adjust the length of the melody,
(Sound value, market value) is about to be corrected. Melody above
The melody generated by the generation function F3 for each traveling section is a melody
The data is stored in the data memory 10 (FIG. 1). In addition,
The chief in this example is the melody at the beginning (measure 1) of the song
The melody generated by the melody generation function F3 is
It is a melody following the melody of the motif. Accordingly
For convenience of later processing, the melody data memory 10
Is the melody of the song in the form of motif data
The data is arranged. In FIG. 2, the automatic music function described above automatically
The composed melody is indicated by reference numeral I4. This completion
For the song, the monitoring D1 allows the user to
Listening can be performed. Monitoring D1
As a result, if the user is satisfied with the generated song (OK)
If so, the song data is not modified at all. Only
If there is an unsatisfactory point, the user enters the section
Force device 1 (Fig. 1)
Wear. In contrast, the automatic composer “learns” the section.
And how to change which parameters
The user can ask questions such as wanting to change
Contact In this case, the question
It is desirable that the format be easy to understand.
As described above, in this embodiment, subjective / objective parameters are used for this purpose.
Data conversion function). Inquiry from automatic composer
In response, the user specifies the desired parameters,
The automatic composer learns the contents of the instruction. Modification of Figure 2
The learning function F4 is where the above learning is performed.
Result (which section to change to which parameter)
Is stored in the learning data memory 9 (Fig. 1) as knowledge.
Is done. Automatic tunes for motifs input later
Machine activates parameter change function F5 by learning
Generates melody according to user's wish
I do. That is, in the section desired by the user,
From the result calculated by the data C calculation function F31.
The parameter provided by the parameter change function F5 is superior
The melody will be generated ahead of time. As a result
The generated melody reflects the user's preferences
And the learning outcomes will shine. In addition, this learning machine
Noh is applied when creating songs with completely different tendencies.
Better not. For this purpose, for example, from the user
Is given from the input device 1 (FIG. 1),
Activate the part of the learning function assigned to the specified song style
can do. Very apparently, the above part
Learning functions by field and field
It has the same aspects as the language learning function. Rice cake
Well, the substance is completely different. <Preliminary Matters> Hereinafter, detailed operations of the embodiment will be described.
Before we go into the explanation,
You. First, in the following description, unless otherwise specified, pitch
The assignment of data (pitch data) to be performed is shown in FIG.
Use things. In other words, the value increases by one semitone
Assigned consecutive integers to the pitch of the scale
Things. Rests are represented by the value zero. Other than this
Can be assigned some pitch data.
Is shown in FIG. 3a. In addition, note length (note value)
The minimum unit is a sixteenth note. That is, the sixteenth note
The pitch data is 1, which is twice as long as the eighth note.
Data is 2. In addition, the “extraction section” and “progression section”
The rank is one bar. In this connection, the chord progression information is also 1
There is one chord per measure. In addition, the motif (input melody
Unless otherwise specified, the length of d) shall be one bar. Also small
Remember that the length of the bars is the same regardless of the bar number.
Set. For convenience of explanation, each chord has four pitches.
It consists of Here, the sum of four independent voices
Sounds, for example (de, mi, so, shi) of C major seventh
Type and two voices have an octave difference 3
There is a type of chord (triad). Therefore,
In this example, the triad of (de, mi, so) is [Outside 1] And In connection with this, each of the chord structure sound memories 2
There are four chord areas, each address has the pitch of each constituent sound
Contains the value corresponding to. For example, In the case of [1], data of (1, 5, 8, 13)
ing. These assumptions are for convenience of explanation only.
It is merely an exemplary assumption. FIG. 4 shows an example of input data for explaining the operation.
You. The score shown in FIG. 4A is an example in the following operation description.
This is a motif. The rest of Figure 4 is self explanatory
The description is omitted here. FIG. 5 shows the formed menu.
An example of Lodi is shown. Of the scores shown above, the first
The measure is the input melody (motif),
2nd, 3rd and 4th bars are automatically adjusted by the automatic composer
The generated melody is shown. 2nd, 3rd, 4th small
Fmaj, G shown below the section₇, Cmaj are coded
Show the chords of the second bar, the third bar and the fourth bar of the line
(See also FIG. 4 (3)). Shown in the lower half of FIG.
After making the automatic composer perform the correction learning,
Melody generated by the automatic composer for the same motif
You. From the figure, you can see that the second measure has been corrected.
You. FIG. 6 is a flowchart showing the main steps used in the following flowchart.
FIG. 4 is a diagram showing a list of various variables. It's a self-explanatory statement
The description is omitted here. As above,
After the explanation, we will start the detailed operation explanation.
You. <Non-harmonic Sound Extraction> FIG.
Shows the flow. In the figure, i indicates the number of the motif data
Variable. In 7-2, the i-th sound (more precisely,
The i-th note in the bar)
Is calculated and set as a variable AFT. As well
, Calculate the number of notes in front of the measure, and assign it to the variable BEF
You have set. In 7-3, the i-th sound of interest is
The difference between the pitches of adjacent sounds before and after
About). For example, a1 is the i
One pitch after the pitch MDi + 2 that is two pitches after the th
A value obtained by subtracting the pitch MDi + 1 of a certain pitch (difference between both pitches) is set.
Is determined. However, in the process of 7-4 to 7-9, rest
(The value is zero, see FIG. 3.)
One MD (motif pitch) is zero (rest)
In the case of (1), a special value is put in the corresponding variable a. 7
-4 to 7-9 are extraction processes of each non-harmonic sound.
Need not be as shown. Which order in principle
May be processed in turn. This is the i-th sound in 7-4 to 7-9
(Or the sound before and after it) is tested for non-harmonic
If the extraction conditions in each extraction process are satisfied,
Is identified as a type of non-harmonic tone. As an aid to understanding
The right side of FIG. 7 shows the non-harmonic tones extracted in each extraction process.
Examples are indicated by arrows in the music score. That is, the sound indicated by the arrow
The mark is the corresponding type of non-harmonic sound. At step 7-10, the note number i in the motif bar is added.
The value is the total number of notes in the motif bar
Until the number exceeds No, the processing of 7-2 and below is repeated. But
Therefore, the processing of 7-1 to 7-11 is executed all at once.
Various non-harmonic sounds included in one bar motif
Is extracted. Each extraction process of 7-4 to 7-8
Are shown in FIGS. 8 to 12, respectively. example
FIG. 8 shows details of the prefetched sound extraction processing. If you pay attention
Is the last note in the bar (the i-th note here)
Yes (AFT = 0 is true) and the sound is the first in the next bar
If the pitch is the same as the sound (a2 = 0 is true), the sound is
This is a first type non-harmonic sound (preemptive sound). This is 8-1
This is the meaning of 8-3. (I) If AFT is zero,
And (ii) if a2 is zero (iii) MDi is of the first kind
Is a non-harmonic sound. According to this description, (i) and (i
i) is the premise of the rule (IF section, LHS), and (iii)
Is the action part of the rule (action part, conclusion part, RHS)
is there. In other words, the above rules ((i)-(ii
i)) defines the first non-harmonic tone in this example
It is. Therefore, changing or modifying the rules
With this, it is possible to update the definition of the first type non-harmonic sound.
Wear. At the implementation level, Figure 8
Flow (Rules (i) to (iii)) is a procedural program
Can be written in a programming language, or
It can also be described as a rule in a source system.
In the latter case, it is easier to update and improve the rules.
No. So, for example, in a logical programming language
Can also be expressed. From the viewpoint of the rules, another extraction processing 7
The same applies to -5 to 7-9. That is, 9-1 to -9 in FIG.
9-7 is the rule or definition of the second kind of non-harmonic sound
10-1 to 10-11 in FIG.
11-1 to 11-11 in FIG.
This is a rule of the kind of non-harmonic sound, and is 12-1 to 12 in FIG.
-16 is the fifth type of non-harmonic rule. In any case
First, the CPU 6 (FIG. 1) executes the processing in accordance with the flowcharts in the respective drawings.
If the extraction conditions are met, the related motif pitch data
Data as a non-harmonic sound. In addition, FIG.
Since the flowchart of FIG. 12 is a self-explanatory description,
The detailed description of is omitted. Instead, they help understanding
First, the motif of the first measure shown in FIG. [Outside 2] Briefly describe what extraction is done
Let me explain. In this example, the total number of notes in the motif is five.
Yes, the first sound is “do” (MD1 = 1). But
Therefore, at the time of this first sound, in 7-2 of FIG. 7,
AFT (the number of notes after the bar) is 4, BE
F (the number of preceding notes in a bar) is zero. Accordingly
AFT is not zero in the preemptive sound extraction process (FIG. 8).
It is not recognized as a preemptive sound. In the biasing process (FIG. 9),
Although the AFT and BEF judgments pass, the next a2 (9-4)
Then, a2 = MD2-MD1 = mead = 5-1 = 4
Do not pass through. In the humming sound processing (FIG. 10), A
The judgment by FT and BEF passed, but a2 was 4.
Do not pass through Processing of the elapsed sound in FIG. 11 and the decoration sound in FIG.
The same is true for the logic. After all, the first sound "do"
It doesn't even sound non-harmonic. increment i
A similar extraction test is performed for the eye sound "mi". here
Also, it is determined that "mi" is not a non-harmonic sound. i = 3
In other words, the operation for the third sound
Let's try. At this time, AFT = 2, BEF = 2, a1
= 5 [Outside 3] , A2 = 2 (= sofa), a3 = 1 (= family),
a4 = 4 (= mead). In the prefetching sound processing, since a2 is not zero, the
In skip (unsatisfied) and sine, skip a because a3 is 1.
A2 = 2, but a3 = 1
To skip to. This is established in the elapsed sound processing. Sand
Since a2 = 2 and a3 = 1, 11-1 → 11-
Go from 2 → 11-3 to 11-4 → 11-7, where A
FT = 2, but since a1 = 5, from 11-7
11-8, go from 11-8 to 11-9, where BEF
= 2 but a4 = 4, so 11-11 to 11-
10, 11-10 to 11-11, MDi (here
Then, the third pitch MD3) is a non-harmonic tone
It is recognized that. Not established in grace sound processing. After all, the third
The eye sound “fa” is determined to be a passing sound. The fourth and subsequent sounds
Is not described (the result is not satisfied). Extraction processing of each non-harmonic sound shown in FIGS. 9 to 12
The logical algorithms or rules are exemplary only. This
Vendors may accept other definitions and rules from the teachings of the present invention.
It can be easily created. For example, if desired,
In addition to the condition for the difference between adjacent pitches,
/ Conditions such as weak beats can be added. One index
From users who do not have enough musical knowledge
Of motif (melody) input
This is to give a voice extraction rule. Various mochi
It is also desirable that a rule with a high extraction success rate be applied to the program. <Extraction of motif parameters>
With the sound extraction function, the motif entered by the user
After the non-harmonic tone is evaluated, the motif parameter
Parameters that characterize the motif using the data extraction function
Data is extracted. The following describes motif parameter extraction.
And will be described in detail. Figure 13 shows motif parameter extraction
It is a flowchart of a process including (1). In the figure, 13-1 to 13-1
13-4 is a motif parameter extraction process.
You. 13-6 is a process for calculating or generating the parameter C.
This is discussed in a separate section. 13
-7 is a parameter changing process by learning.
However, it will be described in another section. Motif parameter extraction processing
In 13-1 to 13-4, the motif memory 4 (FIG. 1)
To extract certain motif pitch data for parameter extraction,
Processing is performed on the work memory 7. For example, motif
Discrimination of various non-harmonic tones instead of the pitch data variable MDi
Variable HDi (or register) with value is used
You. FIG. 14 shows the details of the process 13-1 in FIG.
Unique for non-harmonic sounds
Substituting the pitch data for the harmony
This is a flow to be kept. In FIG. 14, i is the mochi
This is the i-th variable in the bar measure. The flow of FIG.
-14-1 to 14-16 are explicit descriptions in themselves.
This description replaces the description in the specification. But
Therefore, a detailed description thereof will be omitted. Instead, the motif shown in the first bar of FIG.
The processing result for this will be described. <Non-harmonic sound extraction>
As mentioned earlier, this motif [Outside 4] In the numerical expression, as can be seen from FIG.
D2 = 5, MD3 = 6, MD4 = 8, MD5 = 13
), Only the third sound “F” is non-harmonic,
Is a passing sound, and the other sounds are not non-harmonic.
ing. Therefore, the result of executing the processing of FIG.
HD1 = 1, HD2 = 5, HD3 = −40
HD4 = 13). The processing in FIG.
Can actually be done at the stage of non-harmonic extraction
You. What I want you to keep in mind to help you understand later is
If HDi is positive, the value indicates the pitch of the harmony.
If HDi is zero, the note of the i-th motif is a rest
If there is (see FIG. 3b), if HDi is negative, the sound is non-
Harmonic, its value indicates the type of non-harmonic
That's what it means. FIG. 15 shows the details of the process 13-2 in FIG.
The number of each non-harmonic tone contained in the motif
It is a flow to extract as a meter. In the figure, PAj is
Parameters that characterize the chief (referred to as parameter A)
Or a register to store its parameters
Is shown. 15-1 to 15-4
This is the initialization of parameter A. Next 15-5 to 15-1
By processing up to 2, PA1 has a motif (1 bar)
The number of preemptive sounds included in the motif is included in the motif for PA2.
The number of percussions included in the motif is included in PA3.
The number of chords is the PA4
The number is PA5 and the number of ornaments contained in the motif is P
A6 stores the number of other non-harmonic sounds.
You. FIG. 16 shows the details of the process 13-3 in FIG.
The total number of non-harmonic sounds contained in the
It is a flow for calculating the total number. 16-1 to 16 shown
By executing the processing up to -8,
The number of notes included in the motif is included in the motif for PA8.
The total number of harmony tones is PA9
The total number of voice sounds will be stored. FIG. 17 shows the details of the process 13-4 in FIG.
, The type of dispersed chords (sum
Flow to extract the parameters of the vocal time series pattern)
It is. The purpose of the processing here is to follow the flow of the motif.
Each harmony is dispersed throughout the motif
Knows the pitch of the body in the collection of harmony
Is Rukoto. The secondary purpose is to make note number
Is to know what In the figure, the change indicated by ONPU
The number indicates the range (system range) used in this automatic composer.
It represents the range of the input motif.
It has been decided. Illustrated flows 17-1 to 17-1
At 6, the lowest pitch in the system range is used for
(Harmony) is included in the motif,
Increment the variable M of the harmony pitch number only when there is
And the value of LLi. Also, if HDi is zero
When the i-th motif data is a rest,
With the LLi set to zero, the i-th motif data is suspended.
I remember that it was a mark. Also, if HDi is negative
That is, skipping non-harmonic notes.
You. The processing of this flow is shown at the upper right of the flow in FIG.
An example of the logical result is given. In this example, one of the motifs
The third sound (indicated by HD1) is a harmonic sound,
It has the highest harmony among the chiefs. Second
Is a non-harmonic sound. The third sound HD3 is harmony
Sound (the second harmonic of the motif)
It is a harmony with a pitch lower than the harmony. The fourth of the motif
Note HD4 is also a harmony (the third appearance of the motif sum
Voice), but at a lower pitch. Motif
The fifth HD5 is a rest. The sixth of the motif (this place
The note HD6 at the end of the match is a harmony sound, and
The lowest pitch is the same as the fourth note. This
As can be seen from the figure, the flow of the harmony
It has a descending pattern as a body. Flow of FIG.
Is as follows for this motif example. Ma
For HD5, zero indicating a rest as its LLi
Is assigned. For HD4, the model is
1 which is the lowest harmony among the chiefs
Assigned. The same applies to HD6. HD3
Is the second highest from the bottom of the motif as LLi
2 indicating that the sound is a harmony is assigned. HD
2 is a non-harmonic tone, and LLi is not assigned. HD
1 is the highest tone among the harmony sounds included in the motif,
Since it is the third harmonic pitch when counted from the bottom,
Is assigned 3. In other words, HD1 (first harmony
LL1 of sound) is 3, HD3 (second harmonic sound)
LL2 is 2, L of HD4 (the third harmonic that appears)
L3 has a rest after the first and third chords, and LL4 has
0, LL5 of the next harmony HD6 becomes 1, and Mochi
The flow or type (including rests here) of the harmony
It will be specified. That is, {LL} = (LL
1, LL2, LL3, LL4, LL5) = (3,2,
1, 0, 1) are obtained. The parameters of the type of this harmony (dispersion chord)
Is not shown in PA format, but it also features a motif.
This is a parameter to be assigned. This melody will be described in detail later.
Used without any changes in the generation
The result is a highly repetitive and consistent melody
Will be done. The motif of FIG. For 外, the LL of the first appearing harmony
1 = 1, LL2 of the second harmony sound, LL2 = 2, 3rd
LL3 = 3rd and 4th of harmony sounds appearing in eyes
Harmony [Outside 5] LL4 = 4. This is the ascending pattern. FIG. 18 shows details of the process 13-5 in FIG.
Fine, smoothness of the motif (degree of jump)
Seeking. In the illustrated examples 18-1 to 18-7,
To find the difference in LL between adjacent chords
The smoothness value is calculated and set to PA10.
You. Particularly in this example, the difference between the LLs of adjacent chords is
The maximum value is the value of PA10. From FIG. 14 described above
The algorithm shown in FIG. 18 is merely an example. This
Other algos, as disclosed by the present invention, may be
You can easily create a rhythm. In short, functionally
Equivalent or similar ones can be used.
In other words, it is sufficient if the required function is realized. Further
The motif feature parameters other than those described above (for example,
May be extracted.
This concludes the description of the motif feature parameters. [0050]Melody occurrence The melody generation function F3 (FIG. 2)
Menu based on the feature parameters and chord progression information.
This is a function to automatically generate a lody. In this embodiment
The melody generation function is roughly divided into
For calculating parameter C that controls or characterizes
Melody according to the function and given parameter C
And a function to generate di. The latter function is
Dispersion of code in progress with reference to code progress information
A function to generate chords, before and after or
It consists of the function of adding a non-harmonic sound in between. <Generation of Parameter C> Generation of Parameter C
The raw function is a part of the melody generating function F3.
However, before going into the detailed description,
A quick review of the basic properties of distant songs
Like. In general, songs are considered to be uniform and diverse
You. For example, a highly repetitive melody is a well-
Lodi. For example, the range of the melody changes over time
Therefore, when it changes, diversity appears. Time and
In relation, there is unity and diversity. Of course, easy
Depending on the song, the degree of unity and diversity varies. Ah
Some songs emphasize unity and some songs take time
Sometimes the melody keeps flowing. But completely
Rows of random note values and pitches are still, and probably now
Even after that, it is still considered to be a song in a musical sense.
Unlikely. In some respects, music is a table of emotions
Present, completely disorderly and without any regularity
A sequence of sounds is meaningless. In this invention, the chord progression
Melody, which creates one order
Has been introduced. Further, in the present invention, a user input
Parameters that characterize the motif from the empowered motif
Data and generate a melody based on the parameters.
Alive. This creates a motif in the generated music
It is reflected and the essence is kept. The uniformity and variety in the music are described above.
Is closely related to time, for example, the phrase played before
Something as it is or changing
Flowing out again with similarity in the sense of
I often experience it. Therefore, a relatively long sound
It changes completely over a comfortable time (eg one song long)
Controlling melody generation by rules that do not
It is generally disadvantageous and does not produce good results.
The automatic composer takes this point into consideration. sand
In other words, with this automatic composer, the song
To control the melody for each section
Parameter C and assign it to each section
Melody of each section by parameter C
Has occurred. In particular, in the present embodiment, the length of this section
By the way, I have selected one measure. If the section number is i
Parameter C (PC) is a function of the section number. Good
In other words, the parameter C has a value depending on the section number.
(First property of parameter C. In fact, chord progression
Information has similar properties). Further characterize the motif
Parameter C also depends on the parameter
(The second property of C). In this case, the motif feature parameter
Data is independent of the section number, that is,
Over time, it can be incorporated in the parameter C. FIG. 19 shows an example of a parameter map.
It is shown. <Motif parameter A>
Are the motif parameters for FIGS. 15 to 18 described above.
Motif feature parameters obtained by meter extraction processing
(However, the chord type parameter LLi is
Column not shown). <Melody generation parameter C>
The column shows an example of the parameter C. Parameter A
A dotted line with an arrow connecting the group and the parameter group is one parameter.
Data A is reflected in the parameter C indicated by the arrow.
I taste. The connection relationship indicated by the dotted line with arrows is an example.
is there. Some of parameter C depends on parameter A.
Some do not exist. One of the reasons for this is that
Indicates that the motif has a short length of about one bar.
Many parameters above or unreliable parameters
Is not extracted. FIG.
B> is also shown, where the amplitude parameter PB
a, the period parameter PBf and the DC parameter PBd
Have been Such a parameter B is a parameter
The means for generating C is an arithmetic type (this embodiment is an arithmetic type).
) Can be used as an information source. In this case, the parameter C is a section number (example
For example, besides i) PBa, PBf, PBd, P
A (parameter A) at least one of which is a variable
Thus, it is defined as a function that can be calculated. That is,
If the parameter C is represented by PC, PC PC = f (i, PA, PBa, PBf, PBd) In the form However, specifically, one PC
PC = f (i, PA), a certain PC is PC = f (i, PB
d) and so on, independent of certain parameters
Will be. FIG. 20 shows a characteristic example of the parameter C.
ing. The horizontal axis is the direction of music progression
I). (A) is a sign type 1. This type of
Parameter C is an expression containing cosine or sine.
Can occur in arithmetic. For example, PC = cos (i × 2π / PBf) × PBa (I is a bar number). (B) of FIG.
Is also sine type 2, but non-negative PC =. This Thailand
For example, PC = | cos (i × 2π / PBf) × PBa | Is given in the form Same thing, F = cos (i × 2π / PBf) × PBa , And F>If 0, PC = F, if F <0, PC =-
The calculation rule may be F. (C) of FIG.
It is an example of a parameter of the lock type. This type is, for example, F = i MOD PBf (the remainder after dividing i by PBf) And F is zero (if divisible by the period PBf)
If PC = constant, if F is not zero, PC is another constant value
(For example, PA). (D) of FIG.
Is a DC type. This type of operation requires no explanation.
FIG. 20E shows an example of a segmented type. Here, the period
In the work type, only a certain section (for example, the section of rust)
Using arithmetic. This type is, for example,
N-type operation in addition to₁<I <N_TwoIf holds
For example, calculate the function for the section of i, and use the result as PC
Obtained by: Note that N₁<I <N_TwoOther sections
Of course, it is not limited to the peak type. FIG.
(F) is an example of a periodic ramp type. This type is like
If F = Ki (K is a constant), (N + 1) × PBf> F> N × PBf Find an integer N that satisfies (N = INT (F / PBf)), F−N × PBf = Ki−N × PBf And calculate the result as the value of parameter C.
Can be The characteristics (a) to (f) are generated by calculation.
This is only an example of the parameter C that can be easily formed. Again
The components and elements that constitute the data C may be used. Sand
Each type exemplified in (a) to (f) is appropriately combined.
The result can be used as the value of the parameter C. example
For example, a combination of sign types with different periods, a sign type
Combination of DC and DC type, combination of sine type and periodic peak type
There is no end to counting, including matching. By the way, the second
Fig. 2 shows PC2 (parameter of the number of chords of the generated melody).
Data) and PC8 (smooth parameters of generated melody)
FIG. 22B shows an example of calculation for
Various values of (PBf2, PBa2, PBd2, PA8)
Shows how PC2 changes according to measure number i.
ing. One of the advantages of the arithmetic type is that it saves storage space.
You. The operation type is the method adopted in this embodiment,
In the embodiment, the CPU 6 of FIG.
Data and parameter B from parameter B memory 5.
Each parameter C is generated in measures. this
Are the contents of the process 13-6 in FIG. In contrast,
The characteristic example of the parameter C shown in FIG.
Is not suitable for
Parameter C (i) (parameter for controlling each bar of the music)
Of the parameter C that stores the
To "select" an appropriate parameter C (i) from the
Is realized. In any case, the parameter C is the first
This is apparent from the example of FIG.
It is. That is, the parameter C is a progress section of the music.
(Each measure in the embodiment that executes the flow)
With the given value. The parameter C is a motif characteristic parameter.
It has a second property that reflects the meter. For example, figure
The DC type shown in (d) of FIG.
By using as a DC component of
Be incorporated into parameter C in a measure-independent manner
And Examples shown in FIGS. 20A to 20F are described above.
In the arithmetic type, all parameters C are
Has regularity. Such regularity is desirable.
And not a few. From another perspective, the parameters mentioned above
In the example of occurrence of data C, the component parameters (measure number i, parameter
Meter A, parameter B, etc.)
The value of the corresponding parameter C, which is the calculation result of
It is determined. However, looser regularity is desired
In many cases, and for the user,
Motif always produces the exact same melody
Rather than every time you tell me to create a melody
They also want to generate a changed melody
available. The present embodiment takes such points into consideration.
You. That is, it is a function of randomization. This random
Although the conversion function is not specified in FIG.
The arithmetic function F31 or a specific melody is generated.
Distributed chord generation function F32, non-chord generation function F33
Can be built inside. Or random machine
Function between these functions F31, F32, F33
May be shared. For convenience of explanation, FIG.
Function F31 responds to parameter C generated internally
Randomization shall be introduced. In this case,
The parameter C before input is an intermediate parameter,
The parameter after introduction is the final parameter C, that is,
This is a parameter that directly controls the generation of the melody.
FIG. 21A shows the characteristics of the parameter C before random introduction.
In FIG. 20, the periodic peak type shown in FIG.
Belong to. Other characteristic examples are not shown in FIG.
This is merely for convenience, such as saving space. FIG.
1 (b) shows a random characteristic example 1. This feature
First, the meaning of gender is dependent on the direction of the music.
Second, random width (variable)
Third, a random value whose moving width does not depend on the intermediate PC value
Is combined with the original intermediate PC value (eg,
). For example, RND (N)
The pseudo-random functions are -N, -N + 1, ..., 0, +1,
... + N (2N + 1) integer values with equal probability
(It is common for such pseudo-random numbers to occur
Is known). The result of this random number operation
RND (N) is represented by the value of the intermediate parameter C (PC).
Will be added). PC + RND (N) The result is expressed by the final value of parameter C (γ
To). That is, γ = PC + RND (N) The parameter γ after random introduction is the value of the original PC.
As a heart, the values discretely scattered in the width of ± N before and after
You. Parameter characteristics after introduction by such randomization
A sex example is illustrated in FIG. FIG. 21 shows another random characteristic example.
((D) in the figure). This characteristic example (d) is the characteristic described above.
The difference from gender example (b) is that the random width (variation
Width) depends on the intermediate PC value. Illustrated example
Then, the fluctuation width becomes larger as the intermediate PC value becomes larger.
Is chosen as This is, for example, the value of RND (N)
Is multiplied by an increasing function U (PC) of PC, and the result is multiplied by PC
Can be realized by adding the value of For example γ = PC + RND (N) × U (PC) It is. Such a lander is attached to the PC shown in FIG.
(E) of FIG. (E) and
As can be seen from the comparison with (c), in the case of (e), the introduction
No fluctuation occurs when the previous PC is zero. example
For example, this kind of parameter is
Can be used as a determinant. For example,
Prohibits the addition of certain non-harmonic tones when the value of is zero
Can be used as a parameter. Or control the additional position
It can be used as a parameter. One of FIG. 21 (d)
Is based on the value of the PC to be referenced.
That is, the degree is not uniform. like this
Parameter of melody control with random value with distorted probability distribution
May be used as data. 21 (b) and (d) of FIG.
What is common is that any random introduction function
The fluctuation range is introduced based on the PC value of
It is a point. In more general terms, the run is based on the original PC value.
The value after dam introduction is controlled. The other is
The fluctuation width does not change with the direction of the music.
You. This often gives favorable results. I
However, if desired, the range of variation depends on the measure number.
You may choose to change the randomization. Seed of parameter C
Depending on the type, in general, the greater the randomization,
Greater change in the melody generated by it
Is given for each generation. However, based on the original PC value
In the case of the random introduction function
To use as a fine adjustment of the generated melody
Can be. Another random introduction function that was not mentioned above
One advantage is that the non-operation type as illustrated in FIG.
That the characteristic of parameter C can be obtained stochastically
It is. <Specific generation of melody>
Detailed description of the specific occurrence of the melody in the state
You. FIG. 23 is an overall flow of melody generation. That
The main part is in the melody generation in step 23-9, where
Melody is created sequentially in bar units. Rest of the flow
Is a process for transfer between memories and the like. sand
In other words, the processing from 23-1 to 23-5 is based on the motif method.
The motif data in Mori 4 (Figure 1) is melody data
This is a process of transferring data to the memory 10. No shown in 23-5
The number of notes contained in the motif. 23-6 and 23-1
The MNo shown in Fig. 4 is the same as when creating a melody continuously.
This is the total number of melody notes that have been created. Melody
Is performed in measures, so the number of measures in the motif is calculated.
(For example, it can be calculated from the duration data of the motif data.
(23-7). Here, the motif is less than 2 bars
It indicates that it is OK to do so. For motifs of two or more measures
The handling will be described later.
Or one bar. Calculated number of motif measures
Is added to (23-8), and a melody of one measure is generated.
(23-9), the data is stored in the melody data memory
10 (23-10 to 23-13). 23-1
No. 3 is of course one bar generated in the process of 23-9
The number of notes in the melody. CNo shown in 23-15 is
This is the total number of codes used in code progression.
Since the bar number has reached the total number of chords,
At this point, the melody is completed. By the way, the main part
Rody generation processing 23-9 includes processing for generating a dispersed chord,
Processing related to adding non-harmonic sounds and processing related to pitch correction
Contains. Hereafter, the generation of dispersed chords and the addition of non-chords
Description will be given in the order of addition and correction of the sound duration. [0061]Generation of dispersed chords FIG. 24 is an example of the flow of the distributed chord generation process. the first
The processing is the reading of the chord constituent sounds (24-1).
Details are shown in FIG. Reading the chord constituent sounds in FIG.
Process from 25-1 to 25-4
Is the code number from the chord progression memory 3 (see FIG. 26).
When reading bar data (code name) sequentially
It is about time. 25-2 is a chord progression method indicated by the value of i.
The contents of the address of memory 3, that is, the i-th code name
In the register CNi. 25-3
Is stored at the address next to the last codename.
End code (code)
At this point, the reading of the code name is completed. FIG.
In the example of the chord progression, Cmaj, F
maj, G₇And ends with the next Cmaj. FIG.
The chord progression memory 6 is adapted to this example. did
Therefore, CN1 = 1, CN2 = 7, CN3 = 8, CN4
= 1. The processing from 25-5 to 25-12 in FIG.
The process is based on each chord name
Referring to FIG. 2 (see (2) of FIG. 26 and FIG. 4), each code
When reading the pitch data of the chord components of the name
It is about time. In this example, each chord consists of four constituent sounds
It is assumed that the four
Each pitch data is in the lower order in the continuous address. 2
5-7 j = (CNi-1) × 4 + 1 is the reading of each code.
Calculation of the delivery start address, from 25-8 to 25-1
0 reads four pitch data from the start address.
Is being set in the register KDij.
The chord progression is Cmaj, Fmaj, G shown in FIG.₇, C
When taking the maj
KD for Cmaj in the first measure₁₁, KD₁₂,
KD₁₃, KD₁₄Is KD₁₁= 1 (is C), KD₁₂= 5
(M), KD₁₃= 8 (so), KD₁₄= 13 [Outside 6] And in Fmaj of the second measure, (KD_{twenty one}, K
D_{twenty two}, KD_{twenty three}KD_{twenty four}) = (1, 6, 10, 13) = [Outside 7] , G in the third bar₇Then, (KD₃₁, KD₃₂, KD₃₃,
KD₃₄) = (3,6,8,12) = (Le, Fa, So,
The fourth bar is the same as the first bar, and (KD)
₄₁, KD₄₂, KD₄₃, KD₄₄) = (1, 5, 8, 13)
Becomes Note that the data structure of the chord structure sound memory 2 of this example is
The structure is based on C tone (tonality distribution
Consideration). In the following description, KDi₁, KDi_Two, KD
i_Three, KDi_FourInstead of simply KD1, KD2, KD3,
KD4. KD1 is the lowest harmony
Sound, KD2 is the next lower harmony, KD3 is the next lower harmony
The sound, KD4, is used as the register of the highest harmony
You. At this stage, KD1, KD2, KD3, KD4
Has the pitch of the chord component (in the ith measure)
In the basic form, that is, according to the chord configuration sound memory 2
It is in. 24-2 in FIG.
FIG. 27 shows an example of a detailed flow of the turn of the sound.
You. The turning function of the chord constituent sound is the range of the melody to be generated
Is changed over time (in this example, in measures),
To control the excitement of the song.
You can troll. In the flow of FIG. 27, for example, [Outside 8] Then in one turn [Outside 9] In two turns [Outside 10] …… and if you're a refassi, one turn [Outside 11] In two turns [Outside 12] I am trying to be. (8) Both ends are in octave relationship as in (Cmaj)
Code and Refassi (G₇Both ends are ok as in)
Turn logic is different from code that does not
You. That is, the arrangement of the pitches of the chord before turning is represented by KD1
(Old), KD2 (old), KD3 (old), KD4 (old)
(Lowest order)
After the turn, KD1 (new), KD2
(New), KD3 (new), KD4 (new) KD1 (new) = KD2 (old) ...... (fa) KD2 (new) = KD3 (old) ... (SO) KD3 (new) = KD4 (old) ... (S) (M1) KD4 (new) = Octave higher than KD1 (old) ...
… [Outside 13] And for chords that are octave related, KD1 (new) = KD2 (old) ... (mi) KD2 (new) = KD3 (old) ... (SO) KD3 (new) = KD4 (old) ... (do) Up to this point, (M2) KD4 (new) = Octave higher than KD2 (old) ...
… [Outside 14] Need to be Whether or not it is an octave relation is described in 27-3.
And (M1) and (M2) are distinguished by 27-5 and 27-4.
And shift from 27-6 to 27-10, 27-1
1 is the finish of the distinction between (M1) and (M2). PC7 in FIG.
Is a parameter indicating how many times the turning is performed. Mochi
This is one of the parameters C, as you can see from this example.
As described above, the parameter C is a parameter controlling the generation of the melody.
Data. Here, based on the example of FIG.
Let's see what the code turns. Parade of code turning
The calculation of the meter PC7 is as follows for the i-th measure. PC7i = (+ cos ((i + 2) × 2π / 4)) × 1
+1 (See FIG. 4). When i = 2, PC7 = 2 Becomes The second bar code is Fmaj, and its basic
Shape (before turning) [Outside 15] KD1 = 1, KD2 = 6, KD3 = 10, K
D4 = 13. Because PC7 = 2 turns 2 times,
Fruit [Outside 16] KD1 = 10, KD2 = 13, KD3 = 1
8, KD4 = 22. Now, as shown in FIG.
After the turn of the chord constituent sound, the process proceeds to 24-3. Here PC
Reference numeral 9 denotes a parameter for the number of dispersed chords to be maintained. PC ≧ 1
When standing, the flow moves to the maintenance side flow 24-4.
No. PC1 of 24-4 is the correction parameter of the variance chord type
And when PC1 ≧ 1, the dispersion sum is 24-5.
A sound correction is performed. A detailed flow example of the correction of the dispersed chord is shown in FIG.
29. 29-1 to 29-5
The meaning is shown on the right. That is, in this flow example
Motif dispersed chord type {LLi} = (LL1, LL
2, ...) to the opposite relationship, the so-called retrograde
(As mentioned above, LLi appears in motif measures
= The number of the chord in the motif from the bottom
Indicates the pitch of the sound. However, rest LL = is set to zero
doing. See FIG. 17). 24-6 to 24-
9 is the number PC9 that maintains the dispersed chord, and K
The value of D is stored in the intermediate melody register MEDi (current measure
Register that stores the i-th melody pitch data in
It is moving to. That is, the variance sum here
Is the sound type {LLi} a parameter type according to the motif or 24?
This is the type that modified it in -5.
Constituent of chord (chord) progressing in bar ｛KD
From i｝ = (KD1, KD2, KD3, KD4)
I select one and write it to MEDi
is there. Here, {KDi} is already in 24-2,
The range is shifted by the number of the code turning parameter PC7.
It is a collection of harmony sounds. In FIG. 24, reference numerals 24-10 denote the types of dispersed chords.
Is the first step in the flow that branches when
Here, the previous sound (here, the last melody of the previous bar)
Deterministic parameter PC15 from (1) is true (value 1)
If so, proceed to 24-12, where the last of this measure
Decide the first harmony MED =, note in bar at 24-13
The number i is the second and the deterministic parameter PC15 is false.
If i = 1 in 24-11, a lander of 24-14 or less
Proceed to the appropriate distributed chord generation flow. Shown in above 24-12
The detailed flow of the process for determining MED1 from the previous sound is shown in FIG.
8 is illustrated. The logic of this flow is the current measure
Last melody note of the previous measure (previous note)
To the note at the beginning of this measure (next
Sound). Previous sound is MEDj-1, next sound
Is MEDj. Flow 28-1 to 28-10
The description is clear and does not require further explanation. From 24-14 to 24-21 in FIG.
Is run within the range of the smoothness parameter PC8.
The dam is generating a dispersed chord. γ
₁Is 0, 1, 2, 3, 4 according to RND (4).
It is a random number that takes any number of them. P shown in 24-21
C2 is assigned to the melody bar currently being generated
The number of harmony tones (one of the parameters C)
When this is reached, the distributed chord generation processing within that measure is completed.
You. The route from 24-9 to 24-21 is distributed
After generating the dispersed chord by the number PC9 that the chord maintains
The rest is to generate random chords randomly.
You. Here, for the sake of understanding, the second measure in FIG.
It will be easy to see what will happen in the distributed chord generation process.
Just let me state. Note that the turning of the chord components is not
The result (a collection of available harmonies)
Is) [Outside 17] KD1 = 10, KD2 = 13, KD3 = 1
8, KD4 = 24. The number of generated harmonies PC2,
For the value of parameter PC9 for the number of retained chords
PC2 = 6, PC9 =
4, PC1 = 1. In this case, the maintenance example from 24-3
Flows to the flows 24-4 to 24-9 of FIG.
LL1 = 4, LL2 = 3, LL3 =
2, LL4 = 1 is obtained. And 24-6 to 24-
From the beginning of the bar, [Outside 18] (MED1 = KDLL1)
= KD4 = 24 = [Outside 19] , MED2 = 18, MED3 = 13, MED4 = 1
0). The remaining two dispersed chords are from 24-14 to 24-2
Created by one random occurrence, for example, [Outside 20] Becomes So far, the melody of the second bar is [Outside 21] It is. This concludes the description of the generation of dispersed chords. [0069]Add non-harmonic sound After the generation of the dispersed chord is completed, a non-harmonic tone is added. Less than
Hereinafter, the addition of the non-harmonic sound will be described in detail. In FIG.
The following is an example of the flow of adding a bias sound (a first type non-harmonic sound). 30-
1 is a random number γ indicating the presence or absence of a bias sound₁To f (RND
(1)), is being calculated by PC3. This
Here, PC3 is the weight of the percussion assigned to the current melody bar.
And a random number γ in a format controlled by PC3.₁Get
ing. For this random introduction, <parameter C
See the second half of> and FIG. In 30-2,
Random number γ at the position where the sound is attached_TwoThe location of the bias
It is obtained in a format controlled by the parameter PC4. 30-3
To determine whether or not to add a bias sound.
30-9, an array of No intermediate melody data {ME
Of Di｝, γ_TwoNext melody data after
To γ_TwoThe second is secured as a biasing position.
At 30-7, the parameter PC10 of the difference pitch of the bias is set to γ._Three
, And a biasing sound is added at 30-9. Sand
Γ_TwoMEDγ_TwoMEDγ on the right of_{2 + 1}To
γ_ThreeIs added to the value of the pitch addition position.
γ_TwoMEDγ_TwoIs written in. 30-10 is 1
Melody note
The number of notes in a bar is being increased by one. FIG. 31 shows an example of a flow for adding the elapsed sound. This
The meaning of the flow is that adjacent melody notes
Whether to add a passing sound between MEDi and MEDi + 1
Is added if the specified condition is satisfied, and the condition is not satisfied
Is not added. Here, given conditions
Are AND conditions of the following (i), (ii), and (iii). (i) that adjacent melody notes have different pitches,
(ii) The pitch difference between adjacent melody notes is a major third (a =
4) Not greater than (iii) Weight of elapsed sound PC6
Controlled random number γ₁Is the additional value (γ₁= 1)
Showing. If the addition of a passing sound is allowed,
Between the matching melody notes, between the two melody notes
Melody notes of different pitches are written to MED as elapsed sounds
No. The pitch used as the transitional sound is a so-called available note
(Available Notes). Figure
31-10 of 31 is merely exemplary. The above explanation and figure
With the lapse sound from the description of the flow 31-1 to 31-13 of 31
Since the additional operation is clear, further explanation is omitted.
You. FIG. 32 shows an example of a flow for adding an oxalate sound.
You. This flow is also performed under a predetermined condition (weight of PC1).
When the condition that is restricted by
Is added. Same height of adjacent melody notes
Is also one of the conditions. Added
The pitch of the sound is the pitch of the previous melody note.
It is the height shifted by the difference pitch of the hum. Theories above
From the clear description of the flow of
The behavior of the humming sound addition is clear, and no further details are given.
No detailed explanation is required. FIG. 33 exemplifies the flow of adding a grace tone.
Things. What this means is that two adjacent
The loud note pitch is equal and the parametric weight
Random number parameter γ controlled by meter PB₁Is zero
When the melody notes are not
Add before MEDi + 1 and MEDi + 3 (MED
i and MEDi + 2). What
Here, γ shown in 33-10, 33-12, 33-13₁Is
Γ generated in 33-4₁Use a different value than
Can be For this purpose, for example, step 33-
Between 9 and 33-10, γ₁= PCX
) Pattern may be performed. The above explanation and figure
33 from 33-1 to 33-16
The work is clear. Incidentally, the second operation sequence described as the operation sequence
The melody of the measure is the non-harmonic sound added here (Fig.
Let's see what happens in 33). Non-harmonic
Before the addition, the melody of the second bar in FIG.
The sound sequence {MEDi} [Outside 22] Ie, MED1 = 24, MED2 = 18, MED3
= 13, MED4 = 10, MED5 = 13, MED6 =
It was 18. First of all, the bias sound addition (Fig. 30).
In measures, γ₁= 0 (ejection addition prohibited value).
I did not do it. Next, in the case of the elapsed sound addition (FIG. 31),
It looks like this: If the melody note number i in a bar is 1
Come a = | MED1-MED2 | = 4 Since the weight PC6 of the elapsed sound is the second measure, c
os (4 × 2π / 4) × 2 + PA4 (where PA4 =
According to 1), PC6 = 3. Where the random number γ₁
Is calculated, but for the value 3 of PC6, γ₁= 2
And For the next shift processing from 30-4 to 30-9
Than,   {MED} = (MED1, MED2, MED3, MED4, MED5, MED               6, MED7)             = (22, 18, 18, 13, 10, 13, 18) [Outside 23] Becomes And the register MEDi +₁
Is MED2 here, MED2 = MED3- (MED3-MED1) / 2 = 20 [Outside 24] Becomes At this point, the melody sequence {MED} [Outside 25] It is. Increment i to i + 1 until No
In the following, γ₁Or the value of a
No elapsed sound was added. Likewise
And then decide whether or not to add below
And add them. Assume that the result was not added. Non-harmonic sound addition shown in FIGS.
The flow (rule) is only an example. For example, non-sum
To the value of parameter C of voice weight (or similar)
Therefore, the rules for adding each non-harmonic sound in a different manner
Can be changed. For example, when the weight is small enough
(For example, at 1), the corresponding non-
Limit the number of chords to one at most. When the weight is 2, for example
Is not allowed to add more than one non-harmonic
However, continuous addition is prohibited. When the weight is large enough
(For example, when the number is 3 or more)
You. To achieve this, for example,
Use a counter to count the number of executed additions for each non-harmonic
(For example, increment after 31-10), 1
The counter indicates that the corresponding non-harmonic tones have been added.
If you do, PC = 1, in that bar
Completes the process of adding the non-harmonic sound
Use a flag to indicate that a non-harmonic tone has been added at the note position
And if this flag indicates true and PC = 2 is true
After setting the flag to false, adding a non-harmonic tone (for example, 31-7)
~ 31-11) is skipped and continuous addition is prohibited.
Flag true and PC>If 3 is also true, set the flag to false
Free non-harmonic sounds without adding non-harmonic sounds
Allow addition. Further, in the flow of adding the tone in FIG.
Although the number of added percussion sounds is at most one,
Instead of this, it is possible to allow continuous biasing.
Good (depending on the weight of the bias). Other various rules
Changes are possible. The explanation of the addition of non-harmonic sounds has been given above
Finish. [0076]Length correction For the sequence of note lengths {Rhy},
When the generation of the chord is completed, the total duration is
So that the length of the knot, for example, 16 (4 beats)
Assign the note length appropriately. For example, if the number of dispersed chords is 4
Then, 16/4 = 4 (length of one beat) is set to each pitch Rhy1,
Rhy2, Rhy3, and Rhy4 are allocated. Fraction
When the number of chords is not divisible by 16 like 5
Is, for example, Rhy1 = 4, Rhy2 = 2, Rhy3 =
2, unnatural, such as Rhy4 = 4, Rhy5 = 4
Allocate in a form that is not. In one allocation logic, 2
Use split logic. For example [Outside 26] When adding one note at the time of the rhythm, [Outside 27] Divided into two as shown by, and one of them is further divided into two
And [Outside 28]One of them [Outside 29] To As a result, for example, [Outside 30] Rhythm, that is, a sequence {Rhy} of the length of each sound of the melody
Is obtained. To add one more note, split it into two
Already, when [Outside 29] is not corrected, [Outside 31] Where [Outside 29]. For example, [Outside 32] Is obtained. At the stage of adding a non-harmonic sound,
The number of notes in the melody increases by one or two
You. Therefore, for example, when a non-harmonic sound is added
Performing allocation updates with allocation logic at
(Not shown). Enter the note length correction flow of FIG.
At this stage, note length assignment as described above has been completed.
(Completion of tone length adjustment as preprocessing). Hereinafter, FIG.
A description will be given of an example of tone length correction. In this example,
The first is to maintain the consistency of the duration pattern, the second is
Match the length of the note length pattern to the length of the bar
It is to let. Is the pitch pattern consistency 34-7?
34-9. From 34-10
The portion indicated by 34-19 represents the total length of the pitch pattern at the current small
We are adjusting to the length of the knot. More specifically, SIG is the number of melody sounds.
The ratio between No. and the number of notes in the motif No. 1 is much different from 1
Is a variable whose value is switched when
Half the number of notes in a melody bar
To shorten the length of each pitch pattern of the motif
Or lengthen the resulting pitch pattern
The first half of the measure is a pitch pattern. For example,
The melody number of the current bar is included in the motif (for one bar)
When the number of sounds is less than half (No1 / No ≧ 2), SI
Set G to 2. In this case, for example,
If it is 4, 2, 2, ..., then the melody length sequence is 8, 4, 4
...... This is temporarily a half of the motif
This is the tempo pattern. On the other hand, the melody of the current bar
The number of notes is less than twice the number of notes included in the motif (No1 /
When No ≦ 0.5), SIG is set to 0.5. This place
In this case, for example, the length sequence of the motif is 4, 2, 2,...
And the melody tone length sequence 2, 1, 2,.... This is mochi
This is a pattern that temporarily doubles the tempo.
Such integer ratio pattern conversion is generally consistent with the rhythm.
It does not destroy the sex. If the number of melody sounds is close to the number of motif sounds
Ki (0.5 <No1 / No <2), pitch length of motif
Turn the turn to the middle of the melody pitch pattern
To be sure, up to half of the total number of notes in the melody bar)
Turn. By the above-mentioned rewriting of the tone length, the melody sound
The length of the long pattern deviates from the measure length (16 in this case)
It can happen. Therefore, the melody after rewriting the pitch
Calculates the total length of the note duration pattern, which is the bar length
Check if it matches 16, and if not,
Adjust the note length from the last note. For example, melo
If the total length of the note length pattern is longer than the length of the bar,
First look at the duration of the last note in the bar, and if it is 3 or more
If it is, shorten it by one and compare it again with the bar length.
If the duration of the last note is 2 or less, one more hand
Look at the previous note and do the same. Conversely, the duration pattern
If the total length of the bar is shorter than the length of the bar,
And if it is less than 5, extend the note length by 2.
And then again compare to the measure length. The length of the last note
If it is 6 or more, the same applies to the length of the previous note
I do. As a result, the total length of the pitch pattern and the bar
Are matched, and the sound length correction process is completed. Theories above
From the clear description of the flow in FIG.
The description is omitted here. By the way, the melody raw of the second bar in FIG.
In the case where the pitch is
Rody's pitch pattern {Rhy} is 2, 2, 2, 2,
2, 2, 4, that is [Outside 33] It has become. In the processing from 34-1 to 49-9,
Is the same as the motif, {Rhyi} is 4,
2, 2, 4, 2, 2, 4, that is, [Outside 34] become. In the latter half of the adjustment from 34-10,
Become. The SUN is 20 and is longer than the bar length of 16. Last
Note length Rhy5 is 4 and greater than 3, so the last note length
Becomes 2. Here, SUN is 18 and is still longer than a bar.
So far {Rhyi} is 4, 2, 2, 4, 2, 2, 2
It is. If you search for a note with a duration of 3 or more, the fourth note
Is 4, so it becomes 2 and SUN = 16 holds,
Correction is complete. At this time, {Rhyi} is 4, 2,
2, 2, 2, 2, 2, that is [Outside 35] Becomes The pitch of the melody already obtained
When the turns {Rhyi} are shown side by side, [Outside 36] This is exactly the completion of the melody shown in the second bar of FIG.
It is. The melody of bars 3 and 4 in Fig. 5
The progress of the process is omitted, but the result is as shown in the figure.
Let's say That is, G in the third bar₇Then [Outside 37] In Cmaj of the 4th bar (here the last bar) [Outside 38] It is. With the above, the song is completed. <Correction Learning> As described above,
The user responds to the automatic music machine through the monitor
Modifications can be requested. Below, about correction learning
This will be described in detail. In this embodiment, the correction is performed for each bar.
It is. FIG. 35 shows a flow of the correction learning. First, first
Then, in the same manner as described above,
Extract the motif feature parameters (35-
1). Here, every measure of the melody to be generated is
To generate and store parameter C, etc.
Good (the melody will be the same as before)
It is not preferred in terms of quantity. Which bars you want to modify
The user inputs (35-2). In contrast,
The operation composer activates the internal parameter C calculation function and repairs it.
Calculate the parameter C (PCx) of the measure for which there was a correct request
(35-3). For example, the turning parameter PC7i = co
s ((2 + 2) × 2π / 4) × 1 + 1 = 2
Next, all the PCs are calculated. Next, subjective data is obtained from objective data (parameter C).
Parameter conversion into parameters for data (35
-4). The purpose of this parameter conversion is for the user.
Information that is easy to understand and understand.
You. For example, the turning parameter PC7i mainly has
PC7i = 2
"The degree of excitement is 90% in this bar
The person who informs me subjectively
Correction work is easier for the user. This
The logic of this parameter conversion (function operation etc.)
Considering the correlation between data (PC) and subjective data
You can decide. For example, on a development machine,
Estimated correlation (for example, analysis of songs actually composed,
Results obtained by subjective evaluation method)
And the results obtained from the
It can be incorporated into the logic of meter conversion. Parameters
The result (message) of the conversion is to CRT12 (Fig. 1) etc.
It is displayed (35-5). On the other hand, the user can select the type of parameter
The EDB is input (35-6). For example, a paradise
Meter, and the parameters of the excitement and turning
EDB is a parameter if the data
The value indicates the turning parameter as the type of the data C. Subjectivity
Parameter type {S} and objective parameter {O} type
If there is no 1: 1 correspondence, the specified subjective parameter
Objective parameters O (j) corresponding to the type of data S (i),
O (k),... May be obtained by conversion (not shown).
In this flow, a 1: 1 correspondence is assumed). Then you
The user's correction value EDC 'is input (35-7). example
If you want to change the excitement from 90% to 50%
In this case, the user sets the type of the correction parameter to 35-6.
Enter 50% of the correction value (EDC ') 35-7
That is. On the other hand, the automatic composer has a subjective parameter
Inversely converts the value of (EDC ') into the value of the objective parameter
(35-8). In the example above, the excitement should be 50%
Inverse conversion is performed for the request
The value of 7 (EDC) is set to 1. Then, from 35-8 to 3
As shown in 5-9, the contents of correction are stored in the learning data memory 9.
Write in. Here, P is a point of the learning memory 9.
It is. Pointer P is incremented and corrected one after another
The data is being written. The measure you want to modify
Or that there are no more parameter types or values
The user inputs the correction completion, and the correction learning is thus performed.
Is completed (35-12). The above correction learning generates user's preference.
Works to reflect on the melody. This is the 36th
From the operation of parameter change by learning illustrated in the figure,
Will be apparent. That is, parameter change by learning
The parameter C calculation function F31 (FIG. 2)
More usually, the correction learning function than the generated parameter C
Prioritize parameter C learned from F4
You. At 36-3 in FIG. 36, the normal parameters
Calculation of C, already described in <Parameter C>
It is solid. In the sections from 36-4 to 36-11, the learning method
Parameters used to search for melody 9 and generate melody
C is replaced with the learned one. That is, the current
A corrected measure that matches the measure is found (i = * P
a), a parameter that matches the type of parameter of interest
Data type is found (j = * (Pa + 1))
And the correction data EDC indicates that parameter in that measure.
Data PCj (PCj = EDC). In the example of FIG.
In other words, the turning parameter PC7 at the second bar
Instead of 2 (indicating that the number of turns is two)
1 indicating that the number of times of rotation is one is assigned. In FIG. 36, i is a measure number, and j is a parameter.
Parameter C in the meaning used as a subscript PCj of PC
Is a variable to indicate the type (parameter name) of
Pointer of learning memory 9, * Pa is learning pointed by pointer
Memory data. Here, Pa is a multiple of 3
* Pa indicates the number of the corrected measure, and Pa is a multiple of 3
* In the case of Lass 1, * Pa indicates the type of correction parameter
* When Pa is a multiple of 3 plus 2, * Pa is a correction parameter.
It represents the value of the meter (the flow of FIG. 35).
On the right). 36-13 to j> number of parameters
When all the parameters for the ith measure
About C, all the learning content is included
Parameter C is completed. 36-14 is a single measure
The melody of the order is generated, and 23-9 to 23-1 in FIG.
4 is almost supported. In particular, I've shown it here
Is the corrected parameter C up to 36-13
To make it clear that a melody is generated
It is. In short, once stored in the learning memory 9
The read learning data is read each time a user requests composition.
And the place where it has been learned is directly opposed to the parameter C.
Is reflected (see also FIG. 13),
Thus, the generation of the melody is controlled. Accordingly
And the composed melody reflects each user's preference
This is where learning outcomes come into play.
You. For reference, this correction learning is applied to the lower music score in FIG.
This shows the melody after modification. In this example, the turning
Parameter PC changed from two turns before correction to one turn
Has been corrected, and the result appears in the melody of this measure.
ing [Outside 39]. Of these, the harmony of Fmaj [Outside 40] and la and fa [Outside 41] is the basic form Is turned only once). <Features of Embodiment> From the above detailed description
The features of the automatic composer according to the present embodiment are obvious. That
Some are listed below. (B) Motif feature parameters obtained by evaluating motifs
Melody is generated based on chord progression information
The essence and concept of the song that the motif has
It is reflected throughout the body and
Controlled and varied melody
Can be (B) The melody to be generated is controlled in the melody generation means.
C (generation melody feature parameter,
Means for generating the generated melody control parameter)
Parameters for each measure in units of the progression section (measure).
Data C is determined. Parameters assigned to each measure
The melody of each bar is generated according to the value of meter C.
Raw control. Therefore, in the melody flow
Can express unity and diversity. (C) The means for generating the parameter C is the compressed data
(Third parameter B) and motif feature parameters
The parameter C is generated by calculation. did
Therefore, the value of various parameters C is relatively small.
You can get it in storage space. (D) Do you treat non-harmonic sounds and harmony sounds differently?
So the song flow becomes very musical. (E) The motif has come to be input by the user.
You can make a song that reflects the motif
The user composes the song with a sense of participation in composition
At the same time. (F) In addition, no specialized music knowledge is required,
As long as you think of the motif, then the automatic composer
He composes them together. (G) A correction learning function is built in,
This function will learn. The automatic composer has this learning function
The subsequent composition is performed with priority given to the acquired learning data. I
Therefore, the composed song reflects the taste of the user
By now, users are becoming more and more interested
Become. (H) Also, depending on the parameter settings, a distributed chord generator
Function. For example, whether or not non-harmonic sounds can be added
Parameters (eg, PC4, PC6, PC11, PC1
When 3) indicates a prohibited value, any non-harmonic machine
Noh is also prohibited from doing so, and as a result
(In FIG. 30 to FIG. 33, PC4 etc.
When γ₁= 0). <Modification> The present invention is limited to the above embodiment.
However, various modifications, changes, and improvements are possible. For example, on
In this embodiment, the length of each measure is set to the same length for all measures.
However, it may be a measure of variable length. This is for example
Provide a bar counter and assign it to the count value (bar number)
Can be realized by using the specified bar length.
You. In the above embodiment, the bar of the motif is the beginning of the song.
It was supposed to be given as a head, but what number
The measure may be the input measure of the motif. Change for this
The shape is easy. In the above embodiment, the input mochi
The length of the bar was assumed to be one bar, but
You may. For this, for example, in the case of two measures
Is the first motif feature parameter from the first bar motif.
A meter (eg, a sound type parameter LLi)
Second motif feature parameter from the second bar motif
And one of the two types of parameter B and the first mochi
The first parameter C is calculated from the characteristic parameter
The second motif feature parameter of the other parameter B
The second parameter C is calculated from the data. And the first
Parameter C is used to generate a melody for odd-numbered measures, for example.
Used to control, the second parameter C is an even number
Melody generation of measures (measures with zero modulo 2)
(Measure value of bar number counter is 2
Then, which parameter C should be generated?
Can be determined later). However, depending on the music, for example,
Takes the form A, B, A, where A is a bar with 8 measures and B is 7 measures
A phrase in the clause, the last A being an eight-bar phrase, etc.
For example, the value of each passage (8, 7, 8) = (first
The number of bars in the passage, the number of bars in the second passage, and the number of bars in the third passage
And the value of the measure number counter are compared.
It became the ninth bar from the beginning of the song,
When the start of a passage is detected, the bar number counter is incremented.
You only need to reset it. In short, the odd-numbered small element in each passage
Generate the first parameter C at the bar, and at the even-numbered bar
What is necessary is just to generate the second parameter C. Also, in the above embodiment, the
The number of codes is assumed to be 1 (1 chord / measure).
The number of chords per measure can be more than 2
You. For example, the first two beats are the first chord (eg C)
Assume measures where the last two beats are the second chord (eg F)
Let's try this. In the first configuration example, the parameter C generation function
Generates the parameter C in bar units. On the other hand, the dispersed chord
The generation function follows the first code in the section of the first code.
To generate a dispersed chord, and in the section of the second chord, the second chord is generated.
Generate a chord according to the chord. However,
Tab C is used in the bar section. The remaining non-harmonic sound addition function is small
A non-harmonic sound is added for each clause section. In the second configuration example,
The non-harmonic sound addition function is also non-harmonic in two beat units (chord units).
Is added. Other configuration examples are also conceivable. Also desired
If so, change the note length correction function and
The total length of each chord to the chord length (or
A function for correction may be included in (3). Further, even if a selection function such as music style is added,
Good. For example, the entire data of the parameter B is classified.
That is, the data structure of the parameter B memory is classified.
Structure. And by the selection input from the input device,
Determine the data to be read from the parameter B memory
Parameter C is generated by the selected parameter B
Let As described above, parameter C is replaced with parameter B.
Depends on the value of parameter B.
The value of C also changes, resulting in a melody
Changes. Also, the chord composition sounds related to chord progression
The data structure of the memory 2 is also divided based on the song style and the like.
It can be a categorized structure. Selected code
From the set of constituent notes,
Each chord component sound according to the chord name string
Is read. As a result, the distributed chord generating function F32 is generated.
Generated distributed chords are based on a different set of chord components
It is different from the case. As a result, the characteristics of the melody also change.
Become In other words, the setting of the selected chord component
Is a set of chords in the selected field
You. Another alternative approach is to use the parameter C operation
Function F31 is different for each field while maintaining common functions
Built-in functions (arithmetic functions), fields from input devices
Activate the selected function by designated input
You. Distributed chord generation function F32 and non-chord sound addition function F33
A similar approach is possible. The motif evaluation function F1 has a chord decision function.
Functions such as functions may be incorporated. For example, two
Motif for one bar can be input according to the
First, the non-harmonic sound extraction function extracts non-harmonic sounds in measures.
To find out which chord the rest of the note is.
If there is no chord to play, there are two chords in one bar
The non-harmonic sound extraction function extracts non-harmonic sounds in units of two beats.
Chord determination function that determines chords by extracting them
Incorporate. In addition, regarding the melody generation, etc.
In the form, bars are generated one by one sequentially, but by parallel processing
May be generated in parallel. In this case, the connection between measures
Where you need (the melody note from the previous note to the beginning of the next measure)
Need the information of the previous bar
The measure melody generation function starts after the previous measure melody is generated.
Be moved. In addition, various changes can be made. [0097] According to the first aspect of the present invention, the progress of music is achieved.
Generate melody that can change music content variously as you go
And music through multiple music sections
A characteristic that does not lose the song essence (characteristic of the motif)
(Consistency) can be added to the melody flow
(Effect of the first invention). Claim 2The described inventionAccording toNo.
1 In addition to the effects of the invention,For specific music sections,
Using the section feature parameters set through the user input
By generating melody by
You can modify the melody to suit your taste. Claim 3,
According to 4, the music content of the generated melody is included for each characteristic element.
Can be modified to suit user preferences. Claim5To
According to various section melody feature parameters,
You can control various characteristic elements of the music content of D for each element,
Melody generation control can be performed finely. Claim
According to the invention described in Item 6, in addition to the effects of the first invention,said
In the music content of the generated melody, the user
Unwanted part, music feature element by part, feature element by
Can be specified in detail and can be modified.

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 functional block diagram of the automatic music composer. FIG. 3 is a diagram showing an example of pitch data. FIG. 4 is a diagram showing an example of input data for explaining an operation. FIG. 5 is a diagram showing an example of a formed melody. FIG. 6 is a diagram showing a list of main variables. FIG. 7 is a diagram illustrating a part of a famous song together with a flowchart of non-harmonic sound extraction to explain the non-harmonic sound. FIG. 8 is a flowchart of prefetched sound extraction. FIG. 9 is a flowchart of a bias sound extraction. FIG. 10 is a flowchart for extracting an oxalate sound. FIG. 11 is a flowchart of extracting elapsed sound. FIG. 12 is a flowchart of decorative sound extraction. FIG. 13 is a flowchart of parameter extraction and calculation. FIG. 14 is a flowchart for substituting a constant corresponding to each non-Japanese voice into HDi. FIG. 15 shows the number of each non-harmonic tone as a motif feature parameter P
9 is a flowchart for extracting as A1 to PA6. FIG. 16 is a flowchart for extracting the number of harmonic sounds, the number of non-harmonic sounds, and the total number as motif feature parameters. FIG. 17 is a diagram showing a flowchart for extracting a harmony type parameter as a motif feature parameter, and a graph showing an operation example thereof. FIG. 18 is a flowchart for extracting smoothness parameters as motif feature parameters. FIG. 19 is a diagram showing an example of a map between parameters. FIG. 20 is a graph showing a characteristic example of a parameter C (melody control parameter). FIG. 21 is a graph showing an example of randomization for explaining a random introduction function. FIG. 22A is a diagram illustrating an example of an operation for a specific parameter C, and FIG. 22B is a diagram specifically illustrating an action on a specific parameter by a component parameter and a bar number; FIG. 23 is a flowchart of melody generation. FIG. 24 is a flowchart of generating a distributed chord. FIG. 25 is a flowchart of reading a chord constituent sound. FIG. 26 is a diagram showing contents of a chord configuration sound memory and a chord progression memory. FIG. 27 is a flowchart of turning a chord component sound. FIG. 28 is a flowchart for determining MEDi from a previous sound. FIG. 29 is a diagram showing a flowchart of a distributed chord type correction and an operation example thereof in a graph. FIG. 30 is a flowchart of adding a bias sound. FIG. 31 is a flowchart of adding a passing sound. FIG. 32 is a flowchart of adding an oxalate sound. FIG. 33 is a flowchart of adding a decorative sound. FIG. 34 is a flowchart of tone length correction. FIG. 35 is a diagram showing contents of a learning data memory together with a flowchart of correction learning. FIG. 36 is a flowchart of parameter change by learning. [Description of Signs] 1 Input device 2 Chord configuration sound memory 3 Chord progression memory 4 Motif memory 5 Parameter B memory 6 CPU 8 Parameter C memory 10 Melody data memory F1 Evaluation (non-harmonic sound extraction) function F2 Motif parameter extraction function F3 Melody Generating function F31 Parameter C calculating function F32 Distributed chord generating function F33 Non-chord sound adding function

Claims (1)

(57) [Claims] In an automatic composer that generates a melody over a plurality of music sections using each music section as a generation unit, a chord progression providing means for providing chord progression information for the plurality of music sections, and a motif feature parameter characterizing the motif are generated. A motif feature parameter generating means, a motif change parameter generating means for generating a motif change parameter for changing the motif, and the motif feature parameter, the motif change parameter, and the position information of the music section as input variables. An interval melody feature parameter generating means for generating an interval melody feature parameter characterizing an interval melody to be generated in each music interval based on the input variable; and using the interval melody feature parameter and the chord progression as input variables, these Automatic composer which is characterized by having a section melody generating means for generating a section melody of each musical interval on the basis of the force variable. 2. Create melody that spans multiple music sections
In the automatic music composition machine generated as a synthesis unit , information on chord progression for the plurality of music sections is added.
And a motif feature parameter that characterizes the motif.
Motif feature parameter generating means for generating a motif change parameter for changing a motif
Meaning change parameter generating means, the motif characteristic parameter, the thought change parameter
And the position information of the music section as input variables
Based on et input variables, Oite raw to each music section
Interval melody features that characterize the melody to be created
Section melody feature parameter generator that generates parameters
Step, the section melody feature parameter and the chord progression
As input variables, each sound based on these input variables
Interval melody raw to generate interval melody in easy section
Wherein the automatic composer comprising a forming unit, and a section specifying means for specifying a particular music section in accordance with an input from a user, for the specified particular music section, a section melody in the music section in accordance with an input from the user Section melody feature parameter setting means for setting a section melody feature parameter to be added, wherein the section melody generation means converts the section melody feature parameter set by the section melody feature parameter setting means to the section melody feature parameter generation means An automatic composer, which generates a section melody in a specified music section using a section melody feature parameter instead of a section melody. 3. 3. The automatic composer according to claim 2, wherein the section melody feature parameter is composed of a plurality of types, and the section melody feature parameter setting means sets the section melody feature parameter for each type of the specified specific music section. An automatic composer. 4. 4. The automatic composer according to claim 3, wherein the section melody feature parameter setting means (a) displays a data message of the section melody feature parameter generated by the section melody feature parameter generation means for a specified specific music section. (B) a receiving means for receiving, from the user, a parameter change input for each type in response to the displayed section melody feature parameter data message; 5. 2. The automatic composer according to claim 1, wherein said section melody feature parameter generating means comprises means for generating a plurality of types of section melody feature parameters, and each type of section melody feature parameter is used as an operation of said section melody generating means. An automatic composer characterized by functioning to control a corresponding musical feature element in the music content of the resulting section melody. 6. Create melody that spans multiple music sections
In the automatic music composition machine generated as a synthesis unit , information on chord progression for the plurality of music sections is added.
And a motif feature parameter that characterizes the motif.
A motif characteristic parameter generating means, the Kyokuso change of that
Change parameter to generate the change parameter
Meter generating means, the motif characteristic parameter, and the tune change parameter
And the position information of the music section as input variables
Based on these input variables,
Interval melody features that characterize the melody to be created
Section melody feature parameter generator that generates parameters
Step, the section melody feature parameter and the chord progression
As input variables, each sound based on these input variables
Interval melody raw to generate interval melody in easy section
An automatic composer comprising a forming means, and the interval melody feature parameter generating means includes a plurality of kinds of Gu
Means for generating inter-melody feature parameters.
The various types of section melody feature parameters
Section melody music that is the result of the operation of the ody generation means
Control the corresponding musical feature in the content
And learning means for learning the user's preference for individual musical feature elements in the music content of the section melody, the learning means comprising: (a) a plurality of music sections, which are the result of composition; Output means for outputting a melody; (b) a modified data record including a specific music section as a data item, a type of a specific section melody feature parameter in the specific music section, and a correction value of the section melody feature parameter of the type; Correction data record creation means for creating in accordance with a correction instruction input from a user; (c) correction file storage means for storing a file of the correction data record; and (d) section melody feature parameter generation means when composing again. Priority of each correction data record in the correction file storage means over the output of Priority control means for generating a melody by causing the section melody generation means to use a correction value of a specific type of characteristic parameter of a specific music section written in each correction data record. Composer.