JPH11161267A - Automatic music composing device, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To effectively support musical composition for various musics without preparing massive templates. SOLUTION: This device is provided with a template data base 21 having plural templates of passage units including characteristic information, a retrieving means 23 for extracting as a suitable template the template including the characteristic information consistent with a condition for musical composition input on the respective passages by a user, a melody preparing means 25 for preparing a melody data of an objective passage based on a reference condition as to the objective passage and the suitable template extracted by the means 23 as to the objective passage, and a melody storing means 26 for storing the melody data prepared the means 25 with every passage. The retrieving means 23 and the melody preparing means 25 prepare the melody data of the objective passage adding the condition for music composition in the foregoing passage, the selected template and a produced melody, based on the reference condition as to the objective passage.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic music apparatus for selecting a music template based on music composition conditions and composing music using the selected music template, and a computer or the like for reading out and executing the automatic music composition. The present invention relates to a recording medium on which software composing a composition device is recorded.

[0002]

2. Description of the Related Art Conventionally, various automatic music devices for performing automatic music have been proposed in order to reduce the burden on a composer. As an automatic composition device of this kind, the applicant has
An automatic melody generation device for automatically generating a melody based on a song template has been proposed (Japanese Patent Application No. 8-1322).
No. 08 (see the specification and drawings attached to the patent application). The automatic melody generation device searches a plurality of pre-stored music templates using the music composition conditions set by the user as keys, and extracts music templates that match the music composition conditions as candidates. Then, according to the user's instruction, one song template is selected from the candidates, the selected song template is modified as necessary, and a melody is generated using the selected (and modified) song template. Note that a “song template” is a melody template that stores a pitch pattern and a rhythm pattern for one song, and the feature information (the template of the template) is hierarchized in the order of the entire song, phrases, phrases, and measures from the top. Information representing musical characteristics).

[0003]

As described above, in the conventional automatic music apparatus, a music template to be searched stores a pitch pattern and a rhythm pattern of one music. The match / mismatch between the composition conditions over and the feature information of each template is determined. That is, if the feature information of a section of the template matches the desired composition condition but the feature information of another section does not match the composition condition, the song template is not extracted as a candidate. Therefore, in order to support the composition of various songs, it is necessary to prepare a huge song template database consisting of a large number of slightly different song templates. There was a point to be improved that it lacked the variety of songs that could be composed because it was not a target.

The present invention has been made in view of the above circumstances, and provides an automatic music composition apparatus and a recording medium that can effectively support the composition of various music pieces without preparing a huge number of templates. It is an object.

[0005]

In order to solve the above-mentioned problems, an automatic music composition device according to the present invention comprises: a template database having a plurality of templates including feature information for sections shorter than a music piece; A search unit that extracts a template including feature information that matches the feature information input as the composition condition for the target section; and, based on the template extracted by the search unit,
Melody creating means for generating melody data of the target section, wherein the search means refers to composition conditions in a preceding section located ahead of the target section as composition conditions in the target section, and It is characterized in that templates are extracted based on conditions.

According to a second aspect of the present invention, there is provided an automatic music composition apparatus comprising: a template database having a plurality of templates including feature information for a section shorter than a piece of music; and feature information input by a user as composition conditions for the target section. Search means for extracting a template including feature information that matches the melody, and melody generation means for generating melody data of the target section based on the template extracted by the search means, wherein the melody creation means A template extracted in a preceding section located in front of the target section is referred to as a template in the target section, and melody data in the target section is generated based on the referenced template.

Further, the automatic music composition device according to claim 3 is
A template database having a plurality of templates including feature information for a section shorter than the song, and a retrieval unit for extracting a template including feature information matching the feature information input by the user as composition conditions for the target section; Melody generating means for generating melody data of the target section based on the template extracted by the search means, wherein the melody creating means is generated in a previous section located in front of the target section. It is characterized by referring to the melody data and generating the melody data of the target section based on the referenced melody data. Furthermore, in addition to each of the above-described configurations, a previous section to be referred to for the target section,
Setting means for setting at least one of the reference modes at the time of reference may be provided.

According to another aspect of the present invention, there is provided a recording medium, comprising: a template database having a plurality of templates including feature information for sections shorter than a song; A first step of extracting a template including feature information that matches the feature information input as the composition condition of the first section, and a second step of generating melody data of the target section based on the extracted template. In the first step, a program for referring to a composition condition in a previous section located in front of the target section as a composition condition in the target section and extracting a template based on the referred composition condition is recorded. It is characterized by:

According to a sixth aspect of the present invention, there is provided the recording medium, wherein a user inputs, as a composition condition for a target section, a template section having a plurality of templates including feature information for a section shorter than a song. A first step of extracting a template including feature information that matches the criterion, and a second step of generating melody data of the target section based on the extracted template, wherein the second step Recording a program that refers to a template extracted in a previous section located in front of the target section as a template in the target section and generates melody data in the target section based on the referenced template. And

Further, according to the recording medium of the present invention, the user inputs, as a composition condition for a target section, a template database having a plurality of templates including feature information for a section shorter than a song. A first step of extracting a template including feature information that matches the criterion, and a second step of generating melody data of the target section based on the extracted template.
And in the second step, referring to the melody data generated in the previous section located in front of the target section, generating the melody data of the target section based on the referred melody data. It is characterized by recording a program to be executed.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. The hardware part of the present embodiment is realized by a general-purpose personal computer, and the software part is realized by the general-purpose personal computer as one of application software. First, the hardware part will be described.

FIG. 1 is a block diagram showing a hardware configuration of an automatic music composition apparatus according to the present embodiment. In this figure, 1 is CP
U (central processing unit), which controls other components based on a processing program described later. 2 is a RAM (random access memory) and 3 is a ROM (read only memory), which is accessed by the CPU 1 via the bus 6. The ROM 3 stores an initial program loader and the like. Reference numeral 4 denotes an external storage device such as a hard disk drive, a floppy disk drive, a CD-ROM (compact disk read only memory) drive, a magneto-optical disk (MO) drive, a DVD (digital multipurpose disk) device, and the like. Is provided with a storage (recording) medium (hard disk, floppy disk, CD-ROM, magneto-optical disk) storing an operating system, application software (control program) or various data. FIG. 1 shows an example of a hard disk drive. The software (including programs and data) is developed on the RAM 2 after being developed on the hard disk or directly. Note that R
The software may be written in the OM 3 and used. However, by adopting a method in which the software is installed on a hard disk or the like, a control program can be easily added or a version can be easily upgraded. Benefits are born. Reference numeral 5 denotes a display circuit such as a CRT or a liquid crystal display, which displays various information under the control of the CPU 1. Reference numeral 7 denotes a keyboard, 8 denotes a pointing device such as a mouse or a tablet (an example of a mouse is shown in FIG. 1), 9 denotes a sound source, and C denotes a sound source.
A tone signal is synthesized based on the performance information supplied from PU1. Reference numeral 10 denotes a sound system which generates a tone signal synthesized by the sound source 9. 11 is MIDI (Mu
sical Instrument Digital Interface) or an interface with external equipment.

2: Functional Configuration of Embodiment Next, FIG. 2 shows a functional configuration realized by executing hardware of the above configuration by executing application software. In this figure, reference numeral 21 denotes a template database storing a plurality of templates, which is realized in the external storage device 4 such as a hard disk. Each template stored in the template database 21 is shown in FIG.
As shown in FIG. 7, melody data for a section shorter than a song (a passage in this embodiment), that is, a rhythm pattern and a normalized pitch pattern are stored. In the example shown in FIG. 3, a rhythm pattern and a melody pattern are stored for each measure included in a phrase constituting a phrase. In addition, each template is associated with feature information (whole feature information) on the entire song such as genre (for example, American folk song) and tempo, and feature information for each passage (section feature information: passage feature information in FIG. 3). Yes,
It is possible to search using these information as keys. The phrase feature information for the rhythm pattern includes
Melodic for each phrase (flowing melody) /
Rhythmic (chopping melody) and coarse / dense for each measure (the first half of a measure is coarse and the second half is dense) are examples.
Also, as the phrase feature information about the melody pattern, the types of pitch patterns (horizontal, ascending, descending,
Linear, wavy, dynamic, harmony, etc.), and melodic / rhythmic for each phrase.

Numeral 22 designates composition condition input means for inputting overall characteristic information and section (section) characteristic information of the music to be composed as the overall composition condition and the section composition condition. The keyboard 7, the pointing device 8, and the display circuit 5 are provided. , C
It is composed of PU1, ROM3 and the like. The overall composition conditions include genre, tempo, music composition, overall tone, universal music rules, etc. The phrase composition conditions include melodic (flowing melody) / rhythmic (mellowing melody) for each phrase. Separately, the start sound / end sound / amplitude of the pitch pattern, the density of each bar, chords, adoption / non-adoption of music rules, and the like are mentioned.
There is a rule in which the beginning and end of a phrase are chord sounds, a rule in which a sound between adjacent chord sounds separated by one sound and separated by one sound is changed to an elapsed sound, and a rule in which a strong beat is a chord sound. In this embodiment, the composition condition input means 2
Numeral 2 includes a composition condition storage means (implemented by the RAM 2) for storing the input composition conditions. In the present embodiment, the processing subsequent to the completion of the input processing of all composition conditions (template search processing) Etc.) to start. Note that the composition condition input means 22 determines that the overall composition conditions such as genre and tempo are one for all passages.
It is configured so that the input of the degree is sufficient.

Reference numeral 23 denotes search means for searching the template database 21 based on the overall composition condition and the passage composition condition input by the composition condition input means 22 and the reference condition input by the reference condition input means 24 described later. One or a plurality of template candidates including feature information satisfying the condition are extracted from the template database 21. When a plurality of candidates are extracted, the user selects one of them. This search means 23
The present embodiment includes a search result storage unit (implemented by the RAM 2) for storing search results. In the present embodiment, the subsequent melody creation process is started after the search process for all candidates in the processing target phrase is completed. I have.

Reference numeral 24 denotes a reference condition indicating whether or not to refer to the composition condition of the preceding passage in which the composition condition has already been inputted, and if so, a reference condition indicating how to refer, or a forward condition in which the template has already been searched. Reference conditions indicating whether to refer to the template of the phrase, how to refer to it, or whether to refer to the phrase located ahead of the target phrase in the melody creation process, and if so, which Is a reference condition input means for inputting a reference condition indicating whether or not to make a reference, and comprises a keyboard 7, a pointing device 8, a display circuit 5, a CPU 1, a ROM 3, and the like. In the present embodiment, the reference condition input means 24 includes a reference condition storage means (realized by the RAM 2) for storing the input reference conditions. The search process and the melody creation process are started after the input process of the reference condition is completed.

Reference numeral 25 denotes a search result obtained by the search means 23, a reference condition inputted by the reference condition input means 24, melody data of a previous phrase stored in a melody data storage means 26 described later, and a search result of the search means 23. Melody creating means for sequentially creating melody data for each passage, and comprises a CPU 1, a RAM 2, a ROM 3, and the like. Here, FIG. 4 shows a type of pitch pattern in the melody data created by the melody creating means 25, and FIG. 5 shows an example of a rhythm pattern in the melody data. As shown in FIG. 4, in the present embodiment, each pitch pattern is “horizontal” according to the overall direction of pitch change,
"Ascending", "descending", and "linear", "wavy", "dynamic", "harmonic" according to the process of pitch change, and the combination of both as a type of pitch pattern I have. However, types that cannot exist, such as “horizontal” and “dynamic” pitch patterns and “horizontal” and “harmonic” pitch patterns, are excluded.

The melody storage means 26 stores melody data (pitch pattern and rhythm pattern) for each passage created by the melody creation means 25, and is constituted by the RAM 2 or the external storage device 4.

In the present embodiment, as the reference condition input by the reference condition input means 24, information for specifying the passage to be referred to, information indicating the type of reference to each reference destination, and the target passage reflecting the reference result are included. (In the present embodiment, a phrase). Here, the type of reference refers to a division such as whether to create melody data similar to the melody data of the section of the reference destination or to create melody data in contrast to the melody data of the section of the reference destination, and any melody element of the reference destination. In this embodiment, each section can be set for each reference passage.

More specifically, the type of reference includes setting at least a part of the composition conditions set in the reference section as the composition conditions of the main section, and setting the composition conditions set in the reference section. Set a composition condition in contrast to at least a part of the condition as a composition condition for this section, use a rhythm or pitch template selected in a reference section as a rhythm or pitch template for this section, see At least a part of the melody data itself generated in the previous section is directly used as at least a part of the melody in the present section, and at least a part of the melody data itself generated in the reference section is processed (deformed). For example, there is one that is adopted as at least a part of the melody in this section. For example, “set the coarse / fine / dense / coarse condition which is the same as the coarse / fine condition“ dense / dense / rough ”set in the reference section” as the coarse / dense condition of this section ”,“ the melody set in the referenced section ” -"Rhythmic" as opposed to the rhythmic condition "Melodic" is set as the melodic rhythmic condition of this section. Or a pitch template "," a rhythm and / or pitch template selected in a reference section is used as a rhythm and / or pitch template in this section, and data of the template is compressed or expanded ", "The first half of the melody generated in the referenced section The type of reference is, for example, "Adopt as the first half of the melody", "Adjust the pitch of the melody generated in the reference section as a whole, and then modify the melody to the scale note as the melody in this section". Is set as

When the composition conditions relating to the template search are set to be the same or similar (only some conditions are set the same) as the type of reference, the template candidates extracted by the search means 23 are the same or similar. It will be. When a plurality of templates are extracted as candidates, the user can select an arbitrary template from the selected identical or similar template candidates. That is, the user has the freedom to select a template. In the case where the composition condition relating to the template search is set as the reference type, the template candidates extracted by the search unit 23 are in contrast to the reference section. When a template is adopted as the type of reference, the same template as that of the reference section is adopted, and the user has no flexibility in selecting a template. Instead, the user has a degree of freedom in setting composition conditions (chords, music rules, and the like) that are not related to template selection.
When the melody itself is adopted as the type of reference as it is, the user has no degree of freedom at all, but is suitable for creating a repeated section of the same melody. The case where the melody itself is processed and adopted as the type of reference is suitable for a case where a melody is transferred and a similar melody is created.

When the reference is fixed relative or absolutely, the information for specifying the passage to be referred to may be omitted from the reference condition, or when there is only one type of reference. May omit the information indicating the type of reference, or may omit the information indicating the phrase when the phrase in the target passage reflecting the reference result is fixed. Furthermore, it is also possible to use only the reference / non-reference as a reference condition. In the case where the previous passage is not referred to in the present embodiment, the information for specifying the passage to be referred to has the content to that effect. In addition, the user inputs the reference condition input means 24
Therefore, the reference condition is not limited to inputting the reference condition arbitrarily for each measure, but the reference condition may be input by selecting a reference condition pattern obtained by patterning a typical reference condition in advance.

FIG. 6 shows an example of melody data creation processing by the search means 23 and the melody creation means 25. The reference conditions assumed in the example shown in FIG.
The second section refers to the first section which is the preceding section, the third section refers to the second section which is the preceding section, and the fourth section refers to the first and third sections which are the preceding sections. is there. Under such reference conditions, first, the melody data of the first section is created without reference, and the melody data is stored in the melody storage means 26. Next, the melody data of the second section is created with reference to the composition conditions of the first section, the selected template, or the created melody data, and stored in the melody storage means 26.
Next, the melody data of the third section is created with reference to the composition conditions of the second section, the selected template, or the created melody data, and stored in the melody storage means 26. Then, the melody data of the fourth section is created by referring to the composition conditions of the first and third sections, the selected template, or the created melody data, and stored in the melody storage means 26. Note that if the created melody data is immediately output to the outside, the melody data of the fourth section may be directly output to the outside without being stored.

3: Operation of Embodiment 3-1: Outline Next, an outline of a melody data creation operation of the automatic music composition device having the above-described configuration will be described with reference to FIG. In addition,
Here, for the purpose of simplifying the description, a description will be given of a melody data creation operation relating to a specific case (a song composed of the first to fourth passages represented by the passage symbol AA'BA ').

As shown in FIG. 7, the search means 23 and the melody creation means 25 (hereinafter collectively referred to as search and melody creation means) perform a search from the first passage to the passage in which the section composition condition is set. The following processing is performed in order (steps SA1, SA7, SA8). The search and melody creation means first sets the target phrase (here, the first phrase A)
The composition conditions are read from the composition condition storage means (step S
A2) Subsequently, the same phrase symbol as the phrase symbol in the composition condition is included in the previous phrase (a phrase in which the composition condition is stored in the composition condition storage means included in the composition condition input means 22, or is included in the search means 23). It is determined whether the searched template exists in the search result storage means in the phrase in which the searched template is stored or in the melody storage means 26 (the phrase in which the melody data is stored) (step SA3). Since the target phrase is the first phrase, the result of the determination in step SA3, the result of determining whether or not a phrase symbol similar to the phrase symbol in the composition condition exists in the previous phrase (step SA4), and the The determination result (step SA5) as to whether or not a phrase symbol dissimilar to the phrase symbol in the previous phrase exists (step SA5)
O ", and the search / melody creating means creates the melody data of the target passage (first passage A) without referring to the melody data of the previous section (step SA6).

In the present embodiment, similarity of phrase symbols means that the alphabets are the same and the dash symbols are different, such as "A" and "A '". A similar melody data is created in a section to which is added. In the present embodiment, the phrase symbols are dissimilar means that the alphabet portions are different, such as "A" and "B", and the phrase symbols with dissimilar phrase symbols are completely similar. No melody data is created.

For the subsequent second phrase A ', the search / melody creating means firstly performs the same operation as for the first phrase A:
The processing of steps SA2 to SA4 is performed. Since there is no previous phrase having the same phrase symbol A 'as the phrase and a previous phrase (first phrase) having a phrase symbol A similar to the phrase, the determination result of step SA4 is "YES". The search / melody creating means 25 creates melody data similar to the previous phrase by referring to the previous phrase having a similar phrase symbol (step SA9).

Next, the search / melody creating means 25 first performs the processing of steps SA2 to SA5 on the third phrase B in the same manner as for the first phrase A. Since there is no previous phrase having the same phrase symbol B as that phrase, there is no previous phrase having a phrase symbol similar to this phrase, and there is a previous phrase having a phrase symbol dissimilar to this phrase. The determination result of step SA5 is "YES", and the search / melody creating means creates melody data dissimilar to the previous phrase by referring to the previous phrase having the dissimilar phrase symbol (step SA10).

Next, the search / melody creating means performs the processing of steps SA2 and SA3 on the fourth phrase A ', as in the case of the first phrase A. Since there is a previous phrase having the same phrase symbol A 'as the phrase, the result of the determination in step SA3 is "YES", and the search / melody creating means determines that the previous phrase (the second phrase) has the same phrase symbol A'. The same melody data is created by copying the melody data of the section (step SA11).

In the above description, in order to avoid complicating the description, the melody data is created by referring to the composition condition, template, and melody of the previous section based on the similarity / dissimilarity of the phrase symbols, or It is assumed that the melody data is created without doing this, and the passage to be referred to and the type of reference when it is referred to are determined.
The presence / absence of reference, the phrase to be referred to, the type of reference, and the like are determined in consideration of information other than the phrase symbol (to be described in detail later).
Of course, the presence / absence of reference, the reference destination phrase, the type of reference, etc. may be determined based only on information other than the phrase symbols. When the phrase symbols are dissimilar, melody data may be created by non-reference instead of creating melody data by dissimilar reference.

3-2: Specific Example FIG. 8 is a diagram showing the melody data creation processing of the music composed of the first to fourth passages represented by the passage symbol AA'BA 'along the data flow. As shown in the figure, in the present embodiment, first, the melody data (A melody data A) of the first phrase is based on the overall composition condition, the phrase composition condition of the first phrase (A section composition condition), and the number of input lyrics syllables. ) Is generated
It is memorized. Next, an A section composition condition referred to in accordance with the overall composition condition, the phrase composition condition of the second phrase (A 'section composition condition), the number of input lyric syllables, and the reference condition (A section reference condition) in the second phrase, Based on the A section template and the A melody data, melody data (A 'melody data) of the second phrase is generated and stored.

Next, A section composition referred to in accordance with the overall composition condition, the phrase composition condition of the third section (B section composition condition), the number of input lyric syllables, and the reference condition in the third section (A section reference condition). Based on the condition, the A section template, and the A melody data, melody data of the third phrase (B melody data) is generated and stored, and based on the number of input lyric syllables and the A 'melody data of the second phrase. Melody data (A 'melody data) of four phrases is generated.

3-3: Melody Data Creation Operation in Specific Example Next, a more specific melody data creation operation according to the above specific example will be described. FIG. 9 is a diagram showing conditions set at the time of composition for the first phrase A and the second phrase A '. As the conditions, refer to the overall composition condition, the phrase composition condition, the number of input lyric syllables, and the previous phrase. Conditions are indicated.

3-3-1: Operation for Creating Melody Data of First Section FIG. 10 is a diagram showing the flow of processing during the operation of creating the melody data of the first half of the first section A. In FIG. The parameter indicated by c corresponds to the overall composition condition, the section composition condition, and the number of input lyrics syllables. Since the first phrase is a section in which no reference condition is set, the processing shown in this figure is a non-reference creation processing (step SA6 in FIG. 7).

As shown in FIG. 10, in the operation of creating the melody data of the first passage, the tempo (10
0) and genre (American folk song), rhythm pattern (coarse and fine, fine and coarse) and performance tone (melodic) of section composition conditions, and a relative pitch pattern is selected based on the performance tone. Selected. The selected rhythm pattern is adjusted to the number of input syllable syllables (6-3), and the selected pitch pattern is determined by the start sound (A4), end sound (F4), and amplitude (9 semitones) of the section composition condition. The pitch of each hit point is determined based on the absolute pitch. However, at this point, the pitch pattern is merely sampled at the rhythm pattern hit points to determine the pitch of each hit point, and the value of each pitch is linear and not quantized to a musical scale. Therefore, the pitch correction process is performed after the pitch of each hit point is quantized into a scale tone according to the key (A minor) of the overall composition condition.

The process of adjusting the rhythm pattern to the number of input lyric syllables is a process of increasing, decreasing, or maintaining the number of hit points in the selected rhythm pattern. The process of increasing the number of hit points without significantly affecting the melody is performed. For every measure type (pickup measure, phrase start measure, phrase middle measure, ...), all the hit points and their priorities are associated in advance, and among the hit positions where no hit exists in the selected rhythm pattern, This is realized by inserting the hit points in order from the hit point position having the highest priority. The process of reducing the number of hit points without greatly affecting the melody is realized by deleting the hit points in the selected rhythm pattern in order from the hit point position having the lowest priority in the hit point positions where the hit points exist.

The process of converting a pitch pattern into an absolute pitch is a process that is necessary because the pitch pattern on the template is normalized in the amplitude direction and the time direction. Examples include deformation due to expansion / compression of amplitude, deformation according to the first pitch and last pitch, and deformation along the time axis according to a rhythm pattern.

In this scale correction process, the relative chords (Im to IVm) of the section composition condition are absolutely coded according to the key of the overall composition condition, and the universal music rule (before the jump) of the overall composition condition Is the chord sound) and each music rule of the section composition condition (the beginning and end of the phrase are chord sounds and others are optional.
The sound between the adjacent chord sounds separated by a degree is changed to the elapsed sound, and the strong beat is performed based on the chord sound, and the final melody data corresponding to the chord sound is output. The melody in the second half of the first phrase A is generated in the same manner (however, the phrase composition conditions and the number of input lyric syllables are slightly different from those in the first half of the first phrase A).

3-3-2: Operation for Creating Melody Data of Second Section FIG. 11 is a diagram showing the flow of processing during the operation for creating melody data of the second section A '. Are denoted by the same reference numerals. Since the second phrase is a section in which the reference condition (d in the figure) is set and refers to the similar first phrase A, the processing shown in this figure is similar reference creation processing (step SA9 in FIG. 7). It has become. Further, as is clear from the reference condition d in FIG. 11 (the first half of the section has the same rhythm pattern and the same pitch pattern as the first phrase A, and the second half is new), the melody data creation processing of the second phrase is performed in the first half. Is different between the second phrase and the second phrase. Therefore, in FIG.
The parameters used only for the melody data creation processing in the latter half of the phrase are indicated by dotted lines to distinguish them from other parameters. In the phrase composition condition shown in b in the figure and the number of input lyric syllables shown in c, the value shown in parentheses () is a value for the latter half phrase.

In the operation of creating the melody data of the first half phrase of the second phrase, the rhythm pattern and the pitch pattern selected at the time of creating the melody data for the first half phrase of the first phrase are selected according to the reference condition, and the former is selected as the number of input lyric syllables ( The pitch of each hit point is determined on the basis of the result according to 6-3) and the latter. The pitch of each hit point is quantized to a scale sound corresponding to the key (A minor) of the overall composition condition, and then subjected to scale correction processing. This scale correction process
The relative chord (Im-IVm) of the section composition condition is absolutely coded according to the key (A minor) of the overall composition condition, the universal music rule of the overall composition condition (chord sound before jumping) and Melody data is output based on the music rules of the section composition conditions (the sound between adjacent chord sounds three times apart is changed to an elapsed sound).

On the other hand, in the operation of creating the melody data of the second half phrase, a non-reference creation process similar to the process performed for the first phrase is performed, and finally, the universal music rule of the overall composition conditions ( A chord sound before jumping), a chord sound of an absolutely coded section composition condition, and a musical rule of the section composition condition, and a quantized pitch of each hit point obtained based on a rhythm pattern and a pitch pattern. Is corrected and output as melody data.

The reference condition set in the third phrase B is as follows:
The condition is that “melody data is dissimilar to the melody data of the first phrase A”, and the process of creating the melody data of the third phrase is a dissimilar reference process (step SA1 in FIG. 7).
0). Specifically, by making the value of a parameter specified in the reference condition or a predetermined parameter (for example, coarse / fine / coarse or melodic / rhythmic) different from the parameter corresponding to the phrase composition condition of the first phrase. Melody data dissimilar to the first phrase is created. Since the fourth phrase A 'has the same melody data as the second phrase A', the second phrase A '
The melody data obtained for the passage is copied as it is.

FIG. 12 shows a melody for one piece of music obtained as a result of the above-mentioned melody data creation processing.
In FIG. 12, the melody of the first to fourth passages is shown in order from the top of the figure. As is clear from this figure, the first half phrase of the first passage A and the first half phrase of the second passage A 'are Have the same melody, and the second half phrase of the first phrase A and the second half phrase of the second phrase A 'have different melodies. The third passage B has a melody different from that of the first passage A, and the fourth passage A 'has the same melody as the second passage A'. in this way,
Melody data corresponding to the reference condition is created.

3-3-3: Another Example of Melody Data Creation Operation FIG. 13 shows a case in which a condition that “the melody of the preceding section is processed and adopted as the melody of the target section” is set as the preceding section reference condition. It is a figure which shows the flow of a process at the time of the melody data creation operation. The processing shown in this figure corresponds to the similar reference creation processing (step SA9 in FIG. 7). First, the preceding section reference condition in the composition conditions in the target passage is read (step SB1). In this example, since the condition that “the melody of the previous section is processed and adopted as the melody of the target section” is set as the previous section reference condition, the melody of the previous section to be referred to and stored in the melody storage unit 26 is set. The data is read (step SB2). Then, the read tone data is processed (step SB3), and the processed melody data is written to the melody storage means 26 as melody data of the target passage (step SB4). Through the above processing, a melody obtained by processing the melody of the previous section is generated, and written to the melody storage unit 26 as the melody of the target section.

FIG. 14 is a flowchart showing step SB3 in FIG.
FIG. 9 is a diagram showing a first example of detailed processing of FIG. The same numerical value is added to the read pitch data of all the melody data of the preceding section to be referred to (step SB31). That is, the pitch of all the melody data is uniformly shifted by the same pitch. The pitch to be shifted is included in the preceding section reference condition, and is, for example, a minor third pitch. As a result, for example, the melody data of the previous section as shown in FIG.
It is converted into a melody as shown in FIG. Such melody generation is effective when performing transposition or transposition.

FIG. 15 is a flowchart showing the operation of step SB3 in FIG.
FIG. 9 is a diagram showing a second example of the detailed processing of FIG. The same numerical value is added to the read pitch data of all the melody data of the preceding section to be referred to (step SB32). That is, the pitch of all the melody data is uniformly shifted by the same pitch. The pitch to be shifted is included in the preceding section reference condition, and is, for example, a minor third pitch. Then, each pitch data of the melody data after the shift is corrected to a scale tone in the melody key closest to the pitch data before the correction. If there are two scale notes closest to the pitch data before correction (one each above and below), one of them is selected according to a predetermined rule. The predetermined rule is, for example, that the upper part (higher side) is preferentially selected, the lower part (lower side) is preferentially selected, one of them is randomly selected, and the pitch between the preceding and following sounds is reduced. Select the
and so on. As the melody key, the overall key set under the overall composition condition is used. As a result, for example, FIG.
Is converted into a melody as shown in FIG. When a similar melody is generated in the same key, a similar melody may be shifted to a similar shape.
The melody generation of No. 5 is effective in such a case.

FIG. 16 is a flowchart showing the operation of step SB3 in FIG.
FIG. 11 is a diagram showing a third example of the detailed processing of FIG. The read tone length data of all the melody data of the preceding section to be referred to is enlarged or reduced by the same numerical value (step SB34).
In other words, the duration of all melody data is expanded uniformly.
to shrink. The enlargement / reduction magnification is included in the previous section reference condition, and is, for example, twice. As a result, for example, the melody data of the previous section as shown in FIG.
Is converted to the melody as shown in. A melody in which a melody obtained by multiplying the length of a melody in a certain section by a predetermined length appears in another section is often used in music of a classic genre or the like, and the melody generation in FIG. 16 is effective in such a case.

4: Summary As described above, in this embodiment, a rhythm or pitch template is prepared for each passage, and a melody is generated for each passage based on the template selected according to the set composition conditions. At the same time, in a certain passage, the preceding passage is referred to, and the composition conditions set in the front passage, the selected template, and the generated melody are applied to the passage. Therefore, the number of templates is at least large, and the number of combinations is enormous, and the user can compose a desired music piece from the enormous number of combinations. Also, in a certain passage, the melody of the passage can be made the same, similar, or contrasted with the preceding passage by referring to the preceding passage, so that a template can be freely selected for each passage. It is possible to make the flow of the music more natural than that of music that only needs to be performed. Furthermore, since the reference condition can be set for each section shorter than a phrase such as a phrase, the degree of freedom of composition can be improved.

5: Modification In the dissimilar reference creation processing, not only dissimilarity but also melody data may be created so as to be compared with the melody data of the reference destination. For example, the melody data may be created by using a melody element having a value (eg, rhythmic) opposite to the value of the melody element of the reference destination (eg, melody). Instead of setting the reference condition for each section (section), the same reference condition may be set for all sections. However, since there is no reference destination in the head section, it is necessary to make non-reference creation only in the head section.

The template is not limited to one prepared for each phrase, but may be prepared for any unit, such as a phrase unit or a bar unit, as long as it is shorter than the whole music. Further, the present invention is not limited to storing the rhythm pattern and the pitch pattern as a set, but may store them separately, and may store only the rhythm pattern template or only the pitch pattern template. Further, the storage format of the template may be any format.

In the above-described embodiment, an example in which the present invention is realized by a general-purpose computer has been described. However, the present invention can also be realized in the form of a so-called electronic musical instrument. In the form of an electronic musical instrument, for example, a keyboard and various switches perform functions equivalent to those of the above-described keyboard and pointing device.

The electronic musical instrument is not limited to a keyboard instrument, but may be a stringed instrument, a wind instrument, a percussion instrument, or the like. Further, the present invention is not limited to an electronic musical instrument having a built-in sound source device and an automatic performance device, but may be provided separately and connected to each device using communication means such as MIDI or various networks. Of course, it goes without saying that the present invention may be applied to an automatic performance piano.

In the above-described embodiment, the application software is stored in a magnetic disk, an optical disk, or a magneto-optical disk on a general-purpose computer, and is read out and executed on the general-purpose computer to realize various functions. Although shown, it goes without saying that another storage medium such as a semiconductor memory may be used as a storage medium for application software. Further, a communication interface may be provided to download and execute application software via a communication network. The following is an example of downloading application software from the network side.

The communication interface is connected to a communication network such as a LAN (local area network), the Internet, or a telephone line, and is connected to a server computer via the communication network. When the control program and various data are not stored in the storage medium of the self-operating music device serving as a client, the self-composition music device transmits a command requesting the program or data to the server computer via the communication interface and the communication network. . Upon receiving this command,
The server computer distributes the requested program and data to the automatic music composition device via the communication network. Then, the automatic music composition apparatus receives the distributed program or data via the communication interface and stores the program or data in the storage medium, thereby completing the download.

[0055]

As described above, according to the present invention, a template for a section shorter than a song is set for each section, and composition processing of the entire song can be performed using the template. That is, it is possible to compose a variety of music pieces while suppressing the number of templates. Further, the melody data of the target section can be created by referring to the composition conditions of the previous section, the selected template, and the generated melody data, and the preceding section reference conditions (similarity / contrast / duplication, etc.) can be appropriately set. By setting, melody data with a natural flow that can be realized as a music piece can be created. Therefore, compared to the case where templates are prepared for each song, it is possible to effectively support composition of various songs while suppressing the number of templates to be prepared. Furthermore, since the user can input the previous section reference condition for each section, it is possible to more effectively support the composition of a desired music piece.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a hardware configuration of an automatic music composition device according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a functional configuration of the device.

FIG. 3 is a diagram showing a data structure of a template in the same device.

FIG. 4 is a diagram showing types of pitch patterns used in the device.

FIG. 5 is a diagram showing an example of a pitch pattern employed in the device.

FIG. 6 is a diagram showing an example of melody data creation processing by the same device.

FIG. 7 is a flowchart showing an outline of a melody data creation operation by the apparatus.

FIG. 8 is a diagram showing an outline of a melody data creation process of a song composed of first to fourth phrases represented by phrase symbols AA'BA 'in the same device.

FIG. 9 is a diagram showing a comparison between conditions for first and second phrases represented by phrase symbols A and A ′.

FIG. 10 is a diagram specifically showing melody data creation processing for a first phrase represented by a phrase symbol A in the apparatus.

FIG. 11 is a diagram specifically showing melody data creation processing for a second phrase represented by a phrase symbol A 'in the same device.

FIG. 12 is a diagram showing a melody represented by melody data created by the apparatus.

FIG. 13 is a diagram showing a flow of processing in a melody data creation operation under the section reference condition of “processing a melody of a previous section and adopting it as a melody of a target section” in the apparatus.

FIG. 14 is a diagram illustrating a first example of a detailed process of step SB3 in FIG. 13;

FIG. 15 is a diagram showing a second example of the detailed processing.

FIG. 16 is a diagram showing a third example of the detailed processing.

[Explanation of symbols]

Reference Signs List 1 CPU, 2 RAM, 3 ROM, 4 external storage device, 5 display circuit, 6 bus, 7 keyboard, 8 pointing device, 9 sound source, 10 sound system, 11 interface, 21 ... Template database, 22 ... Composition condition input means, 23 ... Search means,
24: Reference condition input means, 25: Melody creation means, 2
6. Melody storage means.

Claims (7)

[Claims] 1. A template database having a plurality of templates including feature information for a section shorter than a song, and a template including feature information matching feature information input by a user as composition conditions for a target section. Search means, and melody creating means for generating melody data of the target section based on the template extracted by the search means, wherein the search means includes a melody in a previous section located in front of the target section. An automatic music composition device, wherein composition conditions are referred to as composition conditions in the target section, and a template is extracted based on the referenced composition conditions. 2. A template database having a plurality of templates including feature information for a section shorter than a song, and a template including feature information matching the feature information input by the user as a composition condition for the target section. And a melody generating means for generating melody data of the target section based on the template extracted by the searching means. An automatic music composition apparatus, wherein a template extracted in a section is referred to as a template in the target section, and melody data in the target section is generated based on the referenced template. 3. A template database having a plurality of templates containing feature information for a section shorter than a song, and a template containing feature information matching the feature information input by the user as composition conditions for the target section. And a melody generating means for generating melody data of the target section based on the template extracted by the searching means. With reference to the melody data generated in the section,
An automatic music composition apparatus for generating melody data of the target section based on the referenced melody data. 4. The apparatus according to claim 1, further comprising setting means for setting at least one of a preceding section to be referred to and a reference mode when referring to the target section.
4. The automatic composition device according to any one of claims 1 to 3. 5. A template that includes feature information that matches feature information input as composition conditions for a target section by a user from a template database that includes a plurality of templates that include feature information for sections shorter than the song. And a second step of generating melody data of the target section based on the extracted template. In the first step, the melody data before the target section is located A recording medium storing a program for referring to composition conditions in a section as composition conditions in the target section and extracting a template based on the referenced composition conditions. 6. A template including feature information that matches the feature information input by the user as composition conditions for a target section from a template database having a plurality of templates including feature information for sections shorter than the song. A first step of extracting; and a second step of generating melody data of the target section based on the extracted template. In the second step, the melody data is located in front of the target section. A recording medium characterized by recording a program for referring to a template extracted in a previous section as a template in the target section and generating melody data in the target section based on the referenced template. 7. A template including feature information that matches the feature information input by the user as composition conditions for a target section, from a template database having a plurality of templates including feature information for sections shorter than the song. A first step of extracting; and a second step of generating melody data of the target section based on the extracted template. In the second step, the melody data is located in front of the target section. With reference to the melody data generated in the previous section,
A recording medium recording a program for generating melody data of the target section based on the referenced melody data.