JPH11126065A - Playing data analyzing apparatus and medium recording playing data analysis program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To divide playing data into musically natural plural playing sections by dividing even sections which include modulation positions but are not divided concerning the playing sections where the playing data are sectionally divided. SOLUTION: The playing data is roughly divided into chorus sections by right after long notes, etc. A phrase section is determined from the chorus section by comparison of the average length of the tone length of the notes of the chorus section and the tone length of the note. The chorus section is divided into repeat sections from the top position of the melody and phrase sections of the chorus section. A tune candidate which is the candidate of the tune is detected in the repeat sections. If the tune candidate is detected, the tune section is determined as it is. If the tune candidate is not detected, the repeat section is finely segmented. When the tune candidate is detected in any of the finely segmented sections, the finely segmented section is determined as the tune section.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for analyzing performance data in electronic musical instruments, computer music, and the like, and to a performance data analyzer for dividing performance data into a plurality of performance sections based on information on notes of the performance data. And a medium recording an analysis data analysis program.

[0002]

2. Description of the Related Art Conventionally, in an electronic musical instrument or computer music, for example, a technique of analyzing a melody to add a chord to a melody or detecting a key and dividing performance data into a plurality of performance sections (phrases, passages, etc.). It has been known. For example, in Japanese Patent Application Laid-Open No. 8-292762, a melody is divided according to the following procedure. First, the average note length of the entire melody is determined, and the end of a note having a note length twice (or nearly twice) is defined as a phrase delimiter.

[0003] Next, the average note length is calculated for each of the divided phrases. If the condition of “average note length of phrase × 2 ≦ end note length of phrase” is not satisfied for each of the two successive phrases, 2 Remove breaks between phrases. Next, another phrase including a part having the same melody as a certain phrase is detected, and the same part and the remaining part of the melody in the detected phrase are subdivided as phrases. Then, measures that have the same melody in each phrase are detected, and the same measures are divided as phrases.

[0004]

In the above-described conventional processing, the performance section is divided based only on the length of a note or the identity of a melody, so that the delimiter position may not be musically appropriate. . For example, if there is a transposition in the middle of a song, it may be appropriate to divide the performance section at that position, but it may not be possible to detect this position with a division method based solely on the length of notes and the identity of the melody. is there.

According to the present invention, the performance data is divided into a plurality of musically natural performance sections by dividing each section divided by a desired method into sections which may include a modulation position. That is the task.

[0006]

According to a first aspect of the present invention, there is provided a performance data analyzing apparatus for performing a process for detecting a key candidate which is a candidate for a key in a first performance section of performance data. If the key candidate detection processing means does not detect a key candidate for the first performance section, the first performance section is subdivided and becomes a key candidate for the subdivided performance section. And subdividing processing means for setting the subdivided performance section to a second performance section when a key candidate is detected.

In the performance data analyzing apparatus according to the first aspect, if a key candidate is not detected in the first performance section, the first performance section may include at least a modulation part. There is. Accordingly, if the first performance section in which no key candidate is detected is subdivided, and the performance section in which a key candidate is detected among the subdivided performance sections is defined as a second performance section, this second performance section There is a high possibility that a section is divided by the modulation section by the section.

According to a second aspect of the present invention, there is provided a medium on which a performance data analysis program executed by a computer for analyzing performance data is recorded, wherein the first performance section of the performance data is recorded. Performing a process of detecting a key candidate that is a candidate for a key in the performance section, and a process of detecting the key candidate.
If no key candidate is detected for the performance section, the first performance section is subdivided, and if a key candidate that is a key candidate for the subdivided performance section is detected, the subdivided performance section is referred to as the first performance section. The performance data analysis program is executed by executing the steps of setting the second performance section by a computer. According to the execution of the performance data analysis program, the same operation and effect as in claim 1 can be obtained.

[0009]

Embodiments of the present invention will be described below. FIG. 2 is a block diagram of an electronic musical instrument to which the performance data analyzer of the present invention is applied and a network system connecting the electronic musical instrument. The CPU 1 uses a working area of a RAM 3 based on a control program stored in a ROM 2. To control the entire electronic musical instrument. At the time of playing the keyboard, the CPU 1 captures a note number, a key-on signal, and the like from the keyboard 4 via the detection circuit 4a, and sets a note number, a velocity, and a note-on / note-off for the tone generator circuit 5, thereby generating sound and processing sound. Perform processing. Various effects are given to the tone signal generated by the tone generator circuit 5 by the effect circuit 6, and a tone is generated by the sound system 7.

The switch 8 is provided on the operation panel of the electronic musical instrument, and the CPU 1 takes in the input data from the switch 8 via the detection circuit 8a and performs various processes according to the operation. The switch 8 includes a tone color selection switch, an automatic performance start / stop switch, and the like. The display circuit 9 is composed of a liquid crystal panel or the like, and a so-called GUI (graphical user's interface) using a liquid crystal panel or the like of the display circuit 9 is used for the user's input operation and the like in addition to the switch 8.

For example, by graphically displaying a screen switch on the liquid crystal panel of the display circuit 9, operating the switch 8, setting the cursor of the liquid crystal panel on the screen switch, and operating the up / down switch of the switch 8, etc.
Various input processes, such as changing the value of data associated with the screen switch, can be performed. The input operation using this GUI also includes a performance data analysis mode selection operation,
There are a selection operation and an input operation of performance data (melody or the like) to be analyzed, and a setting operation in which a user sets division conditions. It should be noted that the timer 10 has a CPU for processing such as automatic performance.
1 generates an interrupt signal (clock signal).

The external storage device 11 includes a hard disk device (HDD), a floppy disk device (FDD), and a CD.
A ROM device, a magneto-optical disk (MO) device, a digital versatile disk (DVD) device, etc., which can be used for inputting and storing various data such as automatic performance data; A MIDI interface (I / F) 12 exchanges various data such as performance data with the MIDI device A, and a communication interface 13 is connected to the communication network B to transmit performance data and the like from the server computer C. Of various data is received.

In the performance data analysis mode, performance data to be analyzed is input from the external storage device 11 or the MIDI interface 12 and stored in a predetermined area of the RAM 3. The performance data can be input in real time by a performance operation of the keyboard 4, or can be input one by one by operating the keyboard 4 and the switch 8 (step record). At the time of inputting such performance data, note data (note length data, pitch data, rest data, rest length data, etc.) is generated by correcting (quantizing) variations in tone generation timing, etc. Configure as data. When the performance data is stored in the RAM 3, the CPU 1 analyzes the stored performance data and divides the performance data into sections in which a key can be specified, as will be described in detail later.

The performance data is data in which note data is stored in time series based on the number of clocks in units of a predetermined clock (for example, a clock obtained by dividing a quarter note into 96). The division position of the performance data is also represented by, for example, a clock value from the beginning of the music. Also, an arbitrary section in the performance data can be represented by a leading clock value and a terminating clock value of the section, or a leading clock value and the number of clocks from the beginning to the ending, and the like. , An arbitrary section can be stored. Further, it is also possible to identify an arbitrary section or group similar sections by storing predetermined data in a register associated with each section. In each process described below, the CP
U1 uses the register of the RAM 3 to perform the above-described section identification, classification, or division position determination.

FIG. 1 is a conceptual diagram illustrating the flow of performance data analysis in the embodiment. First, performance data is divided into chorus sections as follows. In other words, a long note or long rest is searched from the performance data, and the position immediately after that is set as the beginning of the next chorus section, or the position at which the tone is changed is set as the start of the next chorus section. The performance data is divided into chorus sections.

Next, the chorus section is divided into phrase sections as follows. That is, the rule is that the average value of the note length is calculated sequentially from the start position (head) of the chorus section, a note having a note length equal to or more than a constant multiple of this average value is searched, and the immediately following note is set as the start of the next phrase section. Divides the chorus section into phrase sections. For example, the duration of the fourth note is compared with the value of (first duration + second duration + third duration) × K / 3. Note that K is a constant of 1 or more, and may be set by a user, for example.

Next, a repeat section is determined from the beginning positions of the plurality of chorus sections and the phrase section. First, for a plurality of chorus sections, the first plurality of measures (for example, three measures) are compared, and the same number is assigned to a chorus section having the same melody (including parallel movement). Next, when the first bars of the chorus section are in the middle of the own section and the position is the head position of the phrase section, or when the first bars are in the middle of another section and the position is the phrase section. , The same number is assigned to each position as the section start position. Then, a plurality of sections with the same number are compared, and a part that does not match is set as a new section. The above processing is repeated until it becomes impossible to divide, and all sections are numbered, and each section is determined as a repeat section. Note that the same melody is given the same number.

As described above, a process is performed in which a key is divided into repeat sections as a "first performance section", and a key (referred to as a "key candidate") serving as a key of the section is detected for each repeat section. If a key candidate is detected (key detection OK), the repeat section is used as it is as a "second performance section", and if a key candidate is not detected (key detection NG), the repeat section is subdivided. And then further divided. In this manner, when a key candidate is not detected in a repeat section, there is a high possibility that a modulation portion exists in the repeat section, so that the section is subdivided.

FIG. 3 is a flowchart of a section division / tone detection process in the performance data analysis program in the embodiment, and FIG. 4 is a flowchart of a subroutine thereof.
The operation of the embodiment will be described based on each flowchart.

In the performance data analysis mode, when the section division / tone detection processing of FIG. 3 is started, first, as described with reference to FIG. 1, a chorus section determination processing for dividing into chorus sections is performed in step S1. At step S2, a phrase section determination process for dividing into phrase sections is performed, and at step S3, a repeat section determination process for dividing into repeat sections is performed, and the process proceeds to step S4.

In step S4, in the first loop, key candidate detection processing is performed for the first repeat section, but after the second loop, key candidate detection processing is performed for the next repeat section. In this key candidate detection process, for example, the ratio of the total length of the scale notes of a certain key included in the repeat section to the total length of all the notes included in the repeat section is determined by a predetermined value (“pass line”). ")
If there is more than one such key, the key with the highest ratio is determined as a key candidate. In addition, the detection NG of the key candidate is, for example, a case where all the keys are less than the acceptable line. Note that detection processing other than the above may be performed. For a section having the same number as the section in which the key candidate has already been detected, the key candidate detection process may be omitted and the key candidate may be detected (detection OK).

When the processing in step S4 is completed, it is determined in step S5 whether a key candidate has been detected (key candidate detection OK). If not, the repeat section subdivision processing of FIG. 4 is performed in step S6. And return to step S4,
If detected, it is determined in step S7 whether the current repeat section is the last repeat section. If it is not the last repeat section, the process returns to step S4, and if it is the last repeat section, the process ends.

In the repeat section subdivision processing of FIG. 4, in step S11, the sense of end of the last sound of each phrase section in the current repeat section is determined. In the process for obtaining the sense of end, for example, a table of how much the two-tone progression gives a sense of end (according to human senses or music theory) is prepared in, for example, the ROM 2, and the final sound in the phrase section is prepared. Then, the sum of the feeling of ending of the other sounds is calculated (the sound may be weighted according to the length of the sound or the distance between the last sound and the other sounds as necessary), and this is used as the feeling of ending of the last sound.
The processing for obtaining the sense of end may be other than the above, or a combination of another method and the method using the table of sense of end may be used.

When the processing in step S11 is completed, in step S12, among the head positions of the phrase sections that do not include the first plural measures (for example, three measures) and the last plural measures (for example, three measures) in the current repeat section, The position where the sense of end is the highest is set as a new division position. Next, step S
In step S13, it is determined whether or not subdivision has been performed in step S12. If subdivision is not possible, in step S16, the most probable key (even if it is below the pass line) is selected in the key section (the current repeat period). Key), and returns to the original routine.

On the other hand, if the subdivision has been made in the determination of step S13, key candidate detection processing is performed for each of the divided sections in step S14, and whether a key candidate has been detected in step S15 (key candidate detection OK) is determined. If not, the process proceeds to step S16, and if detected, it is determined in step S17 whether there is another repeat section having the same number as the current repeat section number. If there is no repeat section having the same number, step S
In step S18, if there is a repeat section having the same number, the repeat section having the same number as the current repeat section is divided at the same position as the current repeat section in step S18, and the process proceeds to step S19. Step S19
Let's assign a new number to each of the divided sections,
Return to the original routine.

According to the above-described processing, the key segment candidate detected in the first divided repeat section is used as it is (step S5, step S6), and the key segment candidate not detected is replaced with the phrase When a key candidate is detected in each of the divided sections of the subdivided repeat section, the section in which the key candidate is detected (the same applies to a repeat section having the same number) A key adjustment section is set (step S17, step S18, step S19). If subdivision cannot be performed in this repeat section, or if a key candidate cannot be detected in a divided section even if subdivision is possible, a provisional key is determined in step S16.

The performance section (key section) divided as described above can be used for various purposes. For example, the divided performance sections may be used for automatic chord addition to performance data. Also, based on the divided performance sections, the automatic accompaniment sections (intro, main, fill-in,
It may be used to automatically determine the position of the ending or the like. Alternatively, based on the detection result of the key performed together with the detection of the performance section, automatic chord addition to the performance data, pitch conversion at the time of automatic accompaniment, automatic music or automatic music arrangement may be performed.

In the embodiment, the case where the repeat section corresponds to the "first performance section" of claim 1 and claim 2, but the "first performance section" is limited to this repeat section. It is not something to be done. For example, an arbitrary section of the performance data may be the “first performance section”, or may be another performance section.

The format of the performance data in the embodiment is "event + absolute time" in which the time of occurrence of a note or the like (event) is represented by an absolute time in a song or a bar, but is not limited thereto. The format of “event + relative time” in which the occurrence time is represented by the time (duration indicated by the number of clocks) from the immediately preceding event may be used. Although the performance data is represented by note data, the performance data may be represented by data such as musical note codes, generation timings, gate times, and the like. Alternatively, a so-called “solid system” in which a memory area is secured for each minimum resolution of a tone generation event (clock in the embodiment) and data such as note numbers and velocities are stored in a memory area corresponding to the time when the tone generation event occurs. , Or any other format.

The performance data may be in a format in which data of a plurality of channels are mixed, or in a format in which data of each channel is divided for each track.

In the above-described embodiment, the case where the performance data analysis program is stored in the ROM 2 in advance has been described. For example,
A performance data analysis program is recorded on a CD-ROM, and the performance data analysis program is loaded from a CD-ROM device onto a hard disk. Then, the CPU 1 develops the performance data analysis program of the HD in the RAM 3 and controls the operation of the performance data analysis based on the program of the RAM 3 in the same manner as in the above embodiment. This gives R
The same operation as when the performance data analysis program is stored in OM2 can be performed by the CPU. This makes it possible to easily install, add, or upgrade the performance data analysis program.
Alternatively, the performance data analysis program may be recorded on a floppy disk, a magnetic disk (MO), or the like, and supplied to the RAM 3 or a hard disk.

Also, the performance data analysis program may be downloaded using the communication interface 13. In this case, for example, the communication network B is connected to a communication network B such as a LAN (local area network), the Internet, or a telephone line.
, The performance data analysis program is distributed from the server computer C, and is recorded on the hard disk, and the download is completed.

The present invention is not limited to the keyboard-type electronic musical instrument as in the above-described embodiment, but each of the sound source device, the sequencer, the effector, etc. is a separate device, and each device is provided by using communication means such as MIDI or various networks. It may be one that connects devices.

In the above embodiment, the present invention is applied to an electronic musical instrument. However, the performance data analyzer of the present invention can be constituted by a personal computer and application software.

In this case, the personal computer CP
U performs control using a working area of a RAM by an OS installed on a hard disk, for example, and supplies performance data and a performance data analysis program from an external storage device to a hard disk or the like as application software as in the above-described embodiment. The same operation as the above embodiment can be performed by the CPU. When performing the operation of selecting performance data or selecting performance data analysis, control can be easily performed by inputting data in accordance with operations of a keyboard, a mouse, and the like. Also, when a personal computer is used in this way, the information may be supplied via a network as described above.

The medium on which the performance data analysis program as described in the above embodiment is recorded, ie, RO
The destination storage device such as M, RAM, hard disk, CD-ROM, magneto-optical disk, DVD (digital versatile disk), or server computer of a communication network is stored in a medium on which the performance data analysis program according to claim 2 of the present invention is recorded. Equivalent to.

[0037]

As described above, according to the first aspect of the present invention,
According to the performance data analysis device described above or the execution of the performance data analysis program recorded on the medium according to claim 2,
By dividing each of the sections divided by the desired method into sections that may include the modulation position, the performance data can be divided into a plurality of musically natural performance sections.

[Brief description of the drawings]

FIG. 1 is a diagram showing an example of a flow of dividing a performance section in an embodiment of the present invention.

FIG. 2 is a block diagram of an electronic musical instrument to which the performance data analyzer of the present invention is applied and a network system connecting the electronic musical instruments.

FIG. 3 is a flowchart of section division / tone detection processing in the embodiment.

FIG. 4 is a flowchart of a repeat section subdividing process in the embodiment.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... CPU, 2 ... ROM, 3 ... RAM, 11 ... External storage device, 12 ... MIDI interface, 13 ... Communication interface.

Claims (2)

[Claims] 1. Key candidate detection processing means for detecting a key candidate that is a key candidate of a performance section of a first performance section of performance data, and a first performance section by the key candidate detection processing means. If no key candidate is detected, the first performance section is subdivided, and if a key candidate that is a key candidate of the subdivided performance section is detected, the subdivided performance section is divided into a second performance section. A performance data analysis device, comprising: segmentation processing means for dividing into sections. 2. A medium storing a performance data analysis program to be executed by a computer for analyzing performance data, wherein a process is performed for detecting a key candidate that is a candidate for a key of the first performance section of the performance data. And if the key candidate is not detected for the first performance section by the process of detecting the key candidate, the first performance section is subdivided to be a candidate for the key of the subdivided performance section. A step of setting the subdivided performance section to a second performance section when a key candidate is detected.