JP2007047535A - Fingering information generating device and fingering information generation processing program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a fingering information generating device that shortens the processing time by making the processing load reduced. <P>SOLUTION: Playing data MidiEvent of a track (playing part) to be processed are divided into phrase sections, a fingering generating thread is activated for each of the divided phrases, to perform fingering generation based upon playing data MidiEvent constituting the phrase, and the obtained fingering of the phrase is copied as fingering data cfig in the playing data MidiEvent constituting a phrase having the same phrase number phID. Those processes are carried out in parallel for respective tracks and the processing time is shortened by lightening the processing load. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a fingering information generation apparatus and a fingering information generation processing program suitable for use in an electronic musical instrument having a keyboard.

  2. Description of the Related Art There is known an apparatus that generates fingering information that guides a performer to use a finger (fingering) when operating a key. For example, in Patent Literature 1, two-finger fingering data representing ease of finger use when operating the keyboard for two adjacent sounds is stored in association with each combination of fingers. The two-finger fingering data corresponding to the performance data (sound string of the given melody) is read from within, and the fingering information is generated by determining the fingering that optimizes the accumulated value of the read two-finger fingering data An apparatus is disclosed.

Japanese Patent No. 2526954

Incidentally, in addition to the fingering generation method disclosed in Patent Document 1, there is also a method for evaluating all fingerings (fingering) that can be considered for a predetermined number of pieces of performance data and determining an optimum fingering among them. As is known, there is a problem that such a method requires a large processing load because it requires evaluation of all possible finger usages individually.
Therefore, the present invention has been made in view of such circumstances, and provides a fingering information generation device and a fingering information generation processing program capable of reducing processing load and shortening processing time. It is aimed.

  In order to achieve the above object, according to the first aspect of the present invention, there are provided storage means for storing performance data indicating each sound constituting the music and a performance part of each sound, and a performance stored in the storage means. The phrase dividing means for dividing the data into phrase sections in parallel for each performance part and the operation for generating fingering of the phrase sections divided by the phrase dividing means in parallel for each phrase section And a fingering generation means to be performed.

  In the invention according to claim 2, which is dependent on claim 1, the phrase dividing means sets a reference phrase divided into a predetermined number of notes from the beginning of the performance data of the performance part to be processed; Reference is made when the first search means for searching for a comparison phrase that matches the reference phrase from the reference phrase set by the setting means and when the first search means cannot search for a comparison phrase that matches the reference phrase. A second search means for searching for a comparison phrase that matches the reference phrase while shifting the position of the phrase, and when either of the first and second search means searches for a comparison phrase that matches the reference phrase, Section setting means for setting the searched comparison phrase as a phrase section is provided.

  The invention according to claim 3 dependent on claim 1 is characterized in that the fingering generating means assigns the generated fingering to the phrase section of the same type every time the fingering of the phrase section is generated. To do.

  According to the fourth aspect of the present invention, the storage means for storing the performance data indicating the performance parts of each sound and the performance data stored in the storage means are divided into phrase sections. Phrase dividing means for performing an operation for calculating the average number of sounds per section from the number of sounds included in each divided phrase section in parallel for each performance part; and the phrase section divided by the phrase dividing means If the number exceeds the average component number, the fingering generation using the performance data divided by the average component number is performed in parallel for each phrase section, and if the average component number is not exceeded, the corresponding phrase And a fingering generation means for performing a fingering generation operation using all the performance data that constitutes in parallel for each phrase section.

  According to the fifth aspect of the present invention, the performance data representing the performance part corresponding to each sound is divided into phrase sections, and the number of sounds included in each divided phrase section is represented. Phrase division processing for calculating the average number of sounds per section in parallel for each performance part, and if the phrase section divided by the phrase division processing exceeds the average number of constituent sounds The operation of fingering generation using performance data divided by number is performed in parallel for each phrase section, and if the average number of sounds is not exceeded, fingering is performed using all the performance data constituting the corresponding phrase. It is characterized by having a computer perform the fingering production | generation process which performs the operation | movement to produce | generate in parallel for every phrase area.

  According to the first aspect of the present invention, while representing each sound constituting the music and performing the operation of dividing the performance data representing the performance part of each sound into phrase sections in parallel for each performance part, Since the operation of generating fingering of the divided phrase sections is performed in parallel for each phrase section, the processing load can be reduced and the processing time can be shortened.

  According to invention of Claim 4, 5, while representing each sound which comprises music, the performance data showing the performance part of each sound are divided | segmented into a phrase area, and the number of sounds contained in each divided phrase area If the divided phrase section exceeds the average number of constituent sounds while performing the operation of calculating the average number of constituent sounds per section in parallel for each performance part, the performance data divided by the average number of constituent sounds is The fingering generation operation is performed in parallel for each phrase section, and if the average number of sounds is not exceeded, the fingering generation operation is performed for each phrase section using all performance data constituting the corresponding phrase. Therefore, the processing time can be shortened by reducing the processing load by load distribution.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
A. First Embodiment (1) Configuration FIG. 1 is a block diagram showing an overall configuration of a fingering information generating device according to a first embodiment of the present invention. The fingering information generation device includes a CPU 1, a ROM 2, a RAM 3, an input unit 4, a display unit 5, and a MIDI interface 6. The CPU 1 executes a plurality of threads (software execution units) in a time-sharing manner on an OS that supports multi-threads, and apparently realizes real-time parallel processing. Specifically, a main routine (primary thread described later) ) Activates a fingering generation thread for each track (performance part) and performs fingering generation for each track in parallel.

  The ROM 2 stores various control programs and tables that are loaded into the CPU 1. The various control programs referred to here include a main routine (primary thread), a fingering generation thread, and a fingering sequence selection process. The RAM 3 converts a work area for temporarily storing various register / flag data used in the CPU 1 and music data in the SMF format acquired from the external electronic musical instrument 7 through the MIDI interface 6 into performance data MidiEvent described later. A performance data area to be stored and a track data area to store track data Track for managing the performance data MidiEvent stored in the performance data area for each part are provided. The main data structure stored in the RAM 3 will be described later. The input unit 4 generates an operation event corresponding to the user operation and supplies it to the CPU 1. The display unit 5 displays, for example, musical performance data MidiEvent as a musical score under the control of the CPU 1 and also displays fingering data corresponding to each sound constituting the musical performance data MidiEvent in the displayed musical score.

(2) Data Configuration Next, the configuration of the track data area and performance data area provided in the RAM 3 will be described with reference to FIGS. FIG. 2 is a diagram illustrating a configuration of track data Tracks [0] to [N] stored in the track data area of the RAM 3. One track data Track is data for managing performance data MidiEvent stored in the performance data area for each performance part, and is composed of a track number iTrack, a part number iPart, a pointer pNote, a pointer prev, and a pointer next representing the performance part. The The part number iPart represents a right-hand part when “0” and a left-hand part when “1”. The pointer pNote designates the performance data MidiEvent at the head of the track number iTrack. The pointer prev designates the previous performance data MidiEvent. The pointer next designates the next performance data MidiEvent.

  FIG. 3 is a diagram showing a configuration of performance data MidiEvent stored in the performance data area of the RAM 3. The performance data MidiEvent [0] to [N] represents each sound constituting the song, and has END data at the end thereof. The performance data MidiEvent includes a pronunciation start time ITime, a tone length lGate, a phrase number iPhrase, a phrase number iPorder, a pitch pitch, a fingering data cfig, a fingering evaluation value iCost, a chord discrimination value cIsHarm, a chord head pointer phTop, and a chord end It consists of a pointer phTail, a pointer prev, and a pointer next. The chord discrimination value cIsHarm indicates that the chord is not a chord when “0”, and indicates the chord when the value is “component sound −1” (“1” or more). The chord head pointer phTop designates the lowest pitch of the chord constituent sounds. The chord end pointer phTail designates the highest pitch sound among the chord constituent sounds. The pointer prev designates the previous performance data MidiEvent. The pointer next designates the next performance data MidiEvent.

B. Operation Next, the operation of the first embodiment will be described with reference to FIGS. Hereinafter, operations of the main routine (primary thread), the fingering generation thread, and the fingering sequence selection process will be described.

(1) Operation of main routine When an event for instructing the CPU 1 to execute the main routine is supplied from the input unit 4, the CPU 1 executes the main routine shown in FIG. 4 and proceeds to step SA 1 to read data / performance data. Generate. That is, in step SA1, music data in the SMF format is fetched from the external electronic musical instrument 7 via the MIDI interface 6, and the performance data MidiEvent [0] to [N] in the data format shown in FIG. Generate and store in the performance data area of the RAM 3, and generate track data Track [0] to [N] (see FIG. 2) for managing the performance data MidiEvent [0] to [N] individually for each performance part. And stored in the track data area of the RAM 3.

  In step SA2, it is determined whether or not all tracks (performance parts) have been referred to based on the track data Tracks [0] to [N]. If the reference has not been completed, the determination result is “NO”, and the flow proceeds to Step SA3. In step SA3, the pointer pNote for designating the first performance data of the track (performance part) that is the target of fingering generation is stored in the register meFrom, and the value “0” representing the end of the song is stored in the register meTo. In step SA4, the fingering generation thread is activated with the values of the register meFrom and the register meTo as arguments, and the process returns to step SA2.

  Thereafter, the fingering generation thread for each track is activated until all tracks (performance parts) are referred to based on the track data Tracks [0] to [N]. When the fingering generation thread is activated for all the tracks, the determination result in step SA2 is “YES”, and the flow proceeds to step SA5. In step SA5, the process waits until the fingering generation thread for all the tracks is completed, and the determination result is “YES” in response to the completion of the fingering generation thread for all the trucks, thereby completing this process. As described above, in the main routine which is the primary thread, the fingering generation thread for each track is activated until all tracks (performance parts) are referred to based on the track data Tracks [0] to [N]. This processing is completed by completing the fingering generation thread.

(2) Operation of Fingering Generation Thread Next, the operation of the fingering generation thread will be described with reference to FIG. When the fingering generation thread is activated by the above-described main routine (see FIG. 4), the CPU 1 proceeds to step SB1 shown in FIG. 5, and the contents of the register meFrom, which is an argument from the main routine, that is, the target of fingering generation A pointer pNote specifying the first piece of performance data MidiEvent [pNote] of the track (performance part) is stored in the register me and the register met. Subsequently, in step SB2, the number of prefetchs FIGSTEPS is stored in the register i.

  In steps SB3 to SB5, the pointer (met.next) of the register met is advanced by the number of times of the prefetching FIG. STEPS by decrementing until the number of times of prefetching FIGSTEPS stored in the register i reaches “0”. If the pointer (met.next) of the register met is advanced by the number of times of prefetching FIGSTEPS, the determination result in step SB3 becomes “NO”, and the process proceeds to step SB6.

  In step SB6, it is determined whether or not the chord discriminating value cIsHarm (met.cIsHarm) in the performance data MidiEvent specified by the pointer (met.next) advanced by the number of pre-reads FIG. STEPS is “0”, that is, not a chord. . If it is not a chord, the determination result is “YES”, and the flow proceeds to step SB8 described later. On the other hand, if it is a chord, the determination result in step SB6 is “NO”, and the process proceeds to step SB7. In step SB7, the chord end pointer phTail (met.phTail) in the performance data MidiEvent specified by the pointer (met.next) advanced by the number of times of prefetching FIGSTEPS is stored in the register met.

  Next, in step SB8, it is determined whether the pointer of the register met and the pointer of the register meTo do not match. That is, it is determined whether or not the pointer (register met) that specifies the performance data MidiEvent that is the target of fingering generation has reached the end pointer (register meTo). If the pointer (register met) that specifies the performance data MidiEvent that is the target of fingering generation has not reached the end pointer (register meTo), the determination result is “YES”, and the fingering sequence selection process is performed via step SB9. (To be described later).

  As will be described later, in the fingering sequence selection process, an evaluation value representing “difficulty to play” is generated for every possible fingering (fingering sequence) based on the performance data MidiEvent specified by the pointer of the register met. Then, the fingering having the smallest evaluation value, that is, the finger usage that is easiest to play is selected from among them. When the fingering sequence selection process is completed, the process proceeds to step SB10, where the pointer next (me.next) in the performance data MidiEvent currently specified by the pointer stored in the register me is stored in the register me and the pointer is stored. At the same time, the pointer next (met.next) in the performance data MidiEvent currently specified by the pointer stored in the register met is stored in the register met to update the pointer.

  Next, in step SB11, the chord discrimination value cIsHarm (met.cIsHarm) in the performance data MidiEvent specified by the pointer (met.next) stored in the register met updated in step SB10 is “0”, that is, Determine if it is not a chord. If it is a chord, the determination result is “NO”, and after proceeding to Step SB12, the chord end pointer phTail (met.phTail) in the performance data MidiEvent designated by the pointer (met.next) is stored in the register met. Then, the process returns to step SB2. On the other hand, if it is not a chord, the determination result in step SB11 is “YES”, and the process returns to step SB2.

  Thereafter, the processing from step SB2 onward is repeated until the pointer of the register met reaches the end pointer stored in the register meTo. Then, when the pointer of the register met reaches the end pointer stored in the register meTo, the determination result in step SB8 becomes “NO”, and this processing is finished. As described above, the fingering generation thread performs fingering generation for each track (performance part) in parallel based on the contents of the register meFrom which is an argument given from the main routine.

(3) Operation of Fingering Sequence Selection Process Next, the operation of the fingering sequence selection process will be described with reference to FIG. When this process is executed through step SB9 (see FIG. 5) of the fingering generation thread described above, the CPU 1 proceeds to step SC1 shown in FIG. In step SC1, it is determined whether or not the evaluation has been completed for all fingering sequences. If the evaluation has not been completed for all fingering sequences, the determination result is “NO”, and the process proceeds to the next step SC2.

  In step SC2, the evaluation value of the fingering sequence is generated and stored in the register iCost. The evaluation value of the fingering sequence is a value representing “difficulty in playing” obtained from the relationship between the coordinates of the finger assigned to be fingered corresponding to the performance data MidiEvent and the coordinates of the key to be played. Also, the fingering assignment referred to here is a process of assigning a finger whose finger position is not more than the coordinate of the key to be played and that has not been pressed, among the fingers from the thumb to the little finger. Point to. Next, in step SC3, it is determined whether or not the evaluation value of the fingering sequence stored in the register iCost is smaller than the minimum evaluation value stored in the register iCost1. If the evaluation value of the register iCost is greater than the minimum evaluation value of the register iCost1, the determination result is “NO”, and the flow proceeds to step SC5.

  On the other hand, if the evaluation value of the register iCost is smaller than the minimum evaluation value of the register iCost1, the determination result in Step SC3 is “YES”, and the process proceeds to Step SC4. In step SC4, the evaluation value of the register iCost is updated and registered as the fingering evaluation value iCost in the corresponding performance data MidiEvent, and the fingering data representing the finger assigned to the fingering is updated in the corresponding performance data MidiEvent. Register as finger data cfg. Thereafter, the process proceeds to step SC5, the evaluation value of the register iCost is stored in the register iCost1 and the minimum evaluation value is updated, and then the process proceeds to the next fingering sequence.

  Thereafter, the above-described steps SC1 to SC5 are repeated until all fingering sequences have been evaluated. When all the fingering sequences have been evaluated, the determination result in step SC1 is “YES”, and this process is completed. Thus, in the fingering sequence selection process, an evaluation value representing “difficulty to play” is generated for every possible finger use (fingering sequence) based on the performance data MidiEvent specified in the fingering generation thread. Then, the fingering of the smallest evaluation value, that is, the fingering that is most likely to be played is selected from among them, and the fingering data cfig representing the fingering and the fingering evaluation value iCost are registered in the corresponding performance data MidiEvent.

  As described above, in the first embodiment, in the fingering generation thread activated for each track (performance part), every finger use (fingering sequence) that can be considered based on the performance data MidiEvent of the corresponding track. An evaluation value representing “difficulty in playing” is generated, a fingering with the smallest evaluation value is selected, that is, the fingering that is most likely to be played is selected, and fingering data cfig and the fingering evaluation value representing the fingering are selected. As a result of performing the process of registering iCost to the corresponding performance data MidiEvent in parallel, it is possible to reduce the processing load and shorten the processing time.

C. Second Embodiment Next, operations of a main routine (primary thread), a track processing thread, and a phrase division process according to a second embodiment different from the first embodiment will be described with reference to FIGS. Note that the configuration of the second embodiment is the same as that of the first embodiment, and a description thereof will be omitted.

(1) Main Routine Operation When an event for instructing the CPU 1 to execute the main routine is supplied from the input unit 4, the CPU 1 executes the main routine shown in FIG. 4 and proceeds to step SD 1 to read data / performance data. Generate. That is, in step SD1, music data in the SMF format is fetched from the external electronic musical instrument 7 via the MIDI interface 6, and the performance data MidiEvent [0] to [N] in the data format shown in FIG. Generate and store in the performance data area of the RAM 3, and generate track data Track [0] to [N] (see FIG. 2) for managing the performance data MidiEvent [0] to [N] individually for each performance part. And stored in the track data area of the RAM 3.

  Subsequently, in step SD2, it is determined whether or not all tracks (performance parts) have been referred to based on the track data Tracks [0] to [N]. If the reference has not been completed, the determination result is “NO”, and the flow proceeds to step SD3. In step SD3, a track processing thread to be described later is activated for the track (performance part) that is the target of fingering generation, and the process returns to step SD2. Thereafter, the track processing thread for each track is activated until all the tracks (performance parts) are referred to based on the track data Tracks [0] to [N]. When the track processing thread is activated for all tracks, the determination result in step SD2 is “YES”, and the flow proceeds to step SD4.

  In step SD4, the process waits until all track processing threads are completed, and the determination result becomes “YES” in response to completion of all track processing threads, thereby completing this process. As described above, in the main routine which is the primary thread, the track processing thread for each track is activated until all tracks (performance parts) are referred to based on the track data Tracks [0] to [N]. The process ends when it is completed.

(2) Operation of Track Processing Thread Next, the operation of the track processing thread will be described with reference to FIG. When the track processing thread is activated by the aforementioned main routine (see FIG. 7), the CPU 1 proceeds to step SE1 shown in FIG. 8, and divides the performance data MidiEvent of the track (performance part) to be processed into phrase sections. A phrase division process (described later) is executed. Next, in step SE2, a pointer pNote for designating the top performance data MidiEvent [pNote] of the track (performance part) that is the target of fingering generation is stored in the register me.

  In step SE3, it is determined whether or not the value of the register me is not “0”, that is, whether or not the song end has been reached. If the end of the song has not been reached, the determination result is “YES”, and the flow proceeds to step SE4. In step SE4, the phrase number iPhrase (me.iPhrase) in the performance data MidiEvent designated by the pointer of the register me is stored in the register phID. Hereinafter, the content of the register phID is referred to as a phrase number phID. Subsequently, in step SE5, the pointer next in the performance data MidiEvent designated by the pointer of the register me is stored in the register met.

  Then, in steps SE6 to SE7, while incrementing the pointer of the register met, the phrase number iPhrase (met.iPhrase) in the performance data MidiEvent specified by the incremented pointer matches the current phrase number phID. Judge whether to do. In other words, the performance data MidiEvent that ends the phrase is searched for.

  While the phrase number iPhrase (met.iPhrase) matches the current phrase number phID, the determination result at Step SE6 is “YES”, and Steps SE6 to SE7 are repeated. When the phrase number iPhrase (met.iPhrase) does not match the current phrase number phID and the performance data MidiEvent at the end of the phrase is detected, the determination result of step SE6 becomes “NO”, and the process goes to step SE8. move on.

  In step SE8, the pointer of the register me (pointer that specifies the beginning of the phrase) is stored in the register meFrom, and the pointer of the register met (pointer that specifies the performance data MidiEvent that ends the phrase) is stored in the register meTo. In step SE9, thread activation is instructed using the pointers of the registers me and meTo as arguments. Next, in step SE10, after the pointer next in the performance data MidiEvent designated by the pointer of the register met is stored in the register me, the process returns to the above-described step SE3.

  When thread activation is instructed in step SE9, the fingering generation thread (see FIG. 5) is executed via step SE11, and based on the performance data MidiEvent constituting the phrase specified by the pointers of the registers me and meTo. Perform fingering generation. In step SE12, the fingering data of the phrase obtained in the fingering generation thread in step SE11 is copied as fingering data cfig in the performance data MidiEvent constituting the phrase having the same phrase number phID.

  On the other hand, in the process of searching for the beginning and the end of the phrase number phID in steps SE3 to SE10 described above, when the value of the register me becomes “0” and the end of the song is reached, the determination result in step SE3 is “NO”. The process proceeds to step SE13 and waits until all fingering generation threads (step SE11) activated in units of phrases are completed, and the determination result is “YES” in accordance with the termination of all threads, and the process ends. .

  As described above, in the track processing thread, the performance data MidiEvent of the track (performance part) to be processed is divided into phrase sections, and the fingering generation thread is activated for each divided phrase to constitute the phrase. Fingering generation is performed based on MidiEvent, and the fingering data of the obtained phrase is copied as fingering data cfig in performance data MidiEvent that constitutes a phrase having the same phrase number phID.

(3) Operation of Phrase Division Processing Next, the operation of the phrase division processing will be described with reference to FIG. When this process is executed via step SE1 (see FIG. 8) of the track processing thread, the CPU 1 proceeds to step SF1 shown in FIG. 9 and resets the register phID temporarily storing the phrase number to zero. Subsequently, in step SF2, a pointer pNote specifying the first performance data MidiEvent [pNote] of the track (performance part) to be processed is stored in the register meRef.

  Next, in step SF3, a pointer for designating a predetermined number of pieces of performance data MidiEvent from the first piece of performance data MidiEvent [pNote] is stored in the register metRef. Then, the process proceeds to step SF4, and it is determined whether or not the value of the register metRef is not “0”, that is, whether or not the song end has been reached. If the end of the song has not been reached, the determination result is “YES”, the process proceeds to step SF5, and the coincidence flag HIT is set to “0”. Next, in step SF6, a pointer that designates the performance data MidiEvent after the performance data MidiEvent designated by the pointer of the register metRef and for which the phrase has not been detected is stored in the register meSrc.

  In step SF7, a pointer designating a predetermined number of pieces of performance data MidiEvent from the performance data MidiEvent designated by the pointer of the register meSrc is stored in the register metRef. In step SF8, it is determined whether or not the value of the register metSrc is not “0”, that is, whether or not the end of the song has been reached. If the end of the song has not been reached, the determination result is “YES”, and the flow proceeds to step SF11.

  In steps SF11 to SF12, a reference phrase consisting of each sound from the performance data MidiEvent specified by the pointer of the register meRef to the performance data MidiEvent specified by the pointer of the register metRef and the performance data MidiEvent specified by the pointer of the register meSrc. To the performance data MidiEvent designated by the pointer of the register metSrc, it is determined whether or not the comparison phrase consisting of each sound matches. Note that the match / mismatch is determined based on whether or not the “pitch”, “sounding timing”, and “sound length” of the sounds constituting the reference phrase and the comparison phrase match.

  If the reference phrase and the comparison phrase do not match, the determination result in step SF12 is “NO”, and the comparison phrase is updated in steps SF13 to SF14. That is, in step SF13, a pointer that specifies the performance data MidiEvent that is subsequent to the performance data MidiEvent specified by the pointer of the register metSrc and for which the phrase has not been detected is stored in the register meSrc. Subsequently, in step SF14, a pointer designating a predetermined number of pieces of performance data MidiEvent from the performance data MidiEvent designated by the pointer of the register meSrc is stored in the register metRef. Thereafter, the process returns to step SF8 described above.

  Thus, in steps SF8 to SF14, the pointers of the register meSrc and the register metSrc are updated until a comparison phrase that matches the reference phrase appears. If a comparison phrase that matches the reference phrase is found before reaching the end of the song, the determination result in step SF12 is “YES”, and the flow proceeds to step SF15. In step SF15, it is determined whether or not the coincidence flag HIT is not “1”, that is, whether or not a comparison phrase that matches the reference phrase has already been found. When a comparison phrase that matches the reference phrase is found for the first time, the determination result is “YES”, the process proceeds to step SF16, the phrase number phID is incremented, and the match flag HIT is set in step SF17. Set to “1”.

  On the other hand, if a comparison phrase that matches the reference phrase has already been found, the determination result in step SF15 is “NO”, the process proceeds to step SF17, and the match flag HIT is set to “1”. Next, in step SF18, the phrase number phID is registered in the phrase number iPhrase (meSrc.iPhrase) in the performance data MidiEvent designated by the pointer of the register meSrc, and then the process proceeds to the above-described step SF13. Therefore, when a comparison phrase that matches the reference phrase continues, the same phrase number phID as the same group is registered in the phrase number iPhrase (meSrc.iPhrase) in the corresponding performance data MidiEvent.

  In the process of updating the pointers of the registers meSrc and metSrc until the comparison phrase that matches the reference phrase appears, if the pointer of the register metSrc reaches the end of the song, the determination result in the above-described step SF8 becomes “NO”. Thus, the process proceeds to step SF9. In steps SF9 to SF10, the position of a new reference phrase is set. That is, in step SF9, a pointer designating the first performance data MidiEvent for which no phrase has been detected is stored in the register meRef as the head of a new reference phrase, and in a subsequent step SF10, a specified number of data is recorded from the head of the new reference phrase. A pointer designating the previous performance data MidiEvent is stored in the register metRef.

  Thereafter, the process proceeds to step SF4 described above, and it is determined whether or not the value of the register metRef is not "0", that is, whether or not the end of the music has been reached. If the end of the song has not been reached, the determination result is “YES”, and step SF5 and subsequent steps are executed to search for a comparison phrase that matches the newly set reference phrase. Then, when setting a new reference phrase, if the rear end of the reference phrase has reached the end of the song, the determination result in step SF4 is “NO”, and the present process ends.

  As described above, in the phrase division process, a reference phrase obtained by dividing the performance data MidiEvent of the track (performance part) to be processed into a predetermined number of sounds is set, and a comparison phrase matching this reference phrase is searched for. If there is no matching comparison phrase, the matching phrase is searched while shifting the position of the reference phrase, and if there is a matching comparison phrase, the phrase number iPhrase (meSrc.iPhrase) in the first performance data MidiEvent of the comparison phrase is found. ) To register the phrase number phID. Thereby, the performance data MidiEvent of the track (performance part) to be processed can be divided into phrase sections.

  As described above, in the second embodiment, the performance data MidiEvent of the track (performance part) to be processed is divided into phrase sections, and a fingering generation thread is activated for each divided phrase to form a phrase. Fingering generation is performed based on the performance data MidiEvent, and the fingering data of the obtained phrase is copied as fingering data cfig in the performance data MidiEvent constituting the phrase having the same phrase number phID. As a result of executing such processing in parallel for each track, it is possible to reduce the processing load and shorten the processing time.

C. Third Embodiment Next, the operation of the track processing thread according to the third embodiment will be described with reference to FIG. As in the second embodiment described above, when the track processing thread is activated by the main routine (see FIG. 7), the CPU 1 proceeds to step SG1 shown in FIG. 10 and performs the performance of the track (performance part) to be processed. A phrase dividing process for dividing the data MidiEvent into phrase sections is executed. Next, in step SG2, the counters iP and iN are reset to zero. Subsequently, in step SG3, a pointer pNote for designating the top performance data MidiEvent [pNote] of the track (performance part) that is the target of fingering generation is stored in the register me.

  In step SG4, it is determined whether or not the value of the register me is not “0”, that is, whether or not the song end has been reached. If the end of the song has not been reached, the determination result is “YES”, and the flow proceeds to step SG5. In step SG5, the phrase number iPhrase (me.iPhrase) in the performance data MidiEvent designated by the pointer of the register me is stored in the register phID. Hereinafter, the content of the register phID is referred to as a phrase number phID. Subsequently, in step SG6, the pointer next in the performance data MidiEvent designated by the pointer of the register me is stored in the register met.

  In steps SG7 to SG9, while incrementing the pointer of the register met, the counter iN counts the number of data incremented, while the phrase number in the performance data MidiEvent designated by the incremented pointer It is determined whether iPhrase (met.iPhrase) matches the current phrase number phID. In other words, the performance data MidiEvent at the end of the phrase is found out while counting the number of performance data (phrase constituting sounds) constituting the phrase by the counter iN.

  While the phrase number iPhrase (met.iPhrase) matches the current phrase number phID, the determination result in Step SE6 is “YES”, and Steps SG7 to SG9 are repeated. If the phrase data iPhrase (met.iPhrase) does not match the current phrase number phID and the performance data MidiEvent at the end of the phrase is detected, the determination result in step SG7 is “NO”, and the process proceeds to step SG10. move on.

  In step SG10, the pointer of the register me (pointer for designating the beginning of the phrase) is stored in the register meF [iP] designated by the value of the counter iP for counting the number of phrases, and designated by the value of the counter iP. In the register meT [iP], the pointer of the register met (a pointer that specifies the performance data MidiEvent that is the end of the phrase) is stored. Furthermore, in step SG10, the number of phrase constituent sounds counted by the counter iN is stored in a register iN [iP] designated by the value of the counter iP. Next, at step SG11, the value of the counter iP is incremented and stepped, and the counter iN is reset to zero. In step SG12, the pointer next in the performance data MidiEvent specified by the pointer of the register met is stored in the register me, and then the process returns to the above-described step SG4.

  Thereafter, the above steps SG4 to SG12 are repeated until the pointer of the register me reaches the end of the music piece. As a result, a pointer that specifies the head of each phrase divided in step SG1 is stored in the registers meF [0] to [iP], and a pointer that specifies the rear end of each phrase is stored in the registers meT [0] to [iP]. In addition, the number of phrases constituting each phrase is stored in registers iN [0] to [iP]. When the pointer of the register me reaches the end of the music piece, the determination result in step SG4 is “NO”, and the process proceeds to step SG13.

  In step SG13, the average constituent sound number per phrase is calculated based on the phrase constituent sound number of each phrase stored in each of the registers iN [0] to [iP], and the calculated average constituent sound number is stored in the register iNave. To do. Next, in step SG14, the pointer i is reset to zero, and in the subsequent step SG15, it is determined whether or not the pointer i is smaller than the value of the counter iP (hereinafter referred to as the phrase number iP). If the pointer i is smaller than the number of phrases iP, the determination result is “YES”, and the flow proceeds to step SG16.

  In step SG16, it is determined whether or not the register iN [i] designated by the pointer i, that is, the phrase constituting sound number of the phrase designated by the pointer i is larger than the average constituting sound number stored in the register iNave. When the phrase constituent sound number iN [i] of the phrase designated by the pointer i is larger than the average constituent sound number iNave, the determination result is “YES”, and the process proceeds to Step SG17. In step SG17, a thread for generating fingering is activated for the performance data MidiEvent divided by the average component number iNave. Then, the process proceeds to step SG18, the pointer i is incremented and stepped, and then the process returns to step SG15 described above.

  On the other hand, when the phrase constituent sound number iN [i] of the phrase designated by the pointer i is smaller than the average constituent sound number iNave, the determination result in step SG16 is “NO”, and the process proceeds to step SG19. After starting a fingering generation thread for all performance data MidiEvent constituting the specified phrase, the pointer i is incremented in step SG18, and then the process returns to step SG15.

  In this way, depending on whether or not the phrase component sound number iN [i] exceeds the average component sound number iNave, all the threads or phrases constituting the finger or generation of the performance data MidiEvent divided by the average component sound number iNave When one of the threads that generate fingering for the performance data MidiEvent is activated and the pointer i becomes larger than the number of phrases iP, the determination result in step SG15 is “NO”, and the process proceeds to step SG20. In step SG20, the process waits until all the activated fingering generation threads are terminated, and the determination result is “YES” in accordance with the termination of all threads, and the process is terminated.

  As described above, in the third embodiment, the performance data MidiEvent of the track (performance part) to be processed is divided into phrase sections, and the average number of constituent sounds iNave per phrase is determined by referring to all the divided phrases. calculate. Then, for each divided phrase, it is determined whether or not the average constituent sound iNAve is exceeded. If the phrase exceeds the average constituent sound iNAve, the performance data MidiEvent divided by the average constituent sound iNAve is operated. If a phrase that does not exceed the average component number iNave is activated, a thread that generates fingering is activated for all performance data MidiEvent that constitutes the phrase. It is possible to shorten the processing time by reducing the above.

  In the present embodiment, the average constituent sound number iNAve per phrase is calculated with reference to all the divided phrases, and when the calculated average constituent sound number iNAve is exceeded, the average constituent sound number iNave is divided into the average constituent sound numbers iNave. The thread for generating fingering for the performance data MidiEvent is activated, and the thread for generating fingering for all the performance data MidiEvent composing the phrase is activated when the average component number iNave is not exceeded. However, the present invention is not limited to this. In the case of a phrase exceeding the predetermined number of sounds (number of performance data), a thread for generating fingering for the performance data MidiEvent divided into the predetermined number of sounds is activated, and for phrases that do not exceed the predetermined number of sounds, all of the phrases constituting the phrase Activating a thread to generate fingering for performance data MidiEvent It may be as that aspect. In this way, the process of calculating the average constituent sound number iNave becomes unnecessary, and the processing load can be further reduced.

It is a block diagram which shows the structure of 1st Embodiment by this invention. It is a figure which shows the structure of track data Track. It is a figure which shows the structure of performance data MidiEvent. It is a flowchart which shows operation | movement of the main routine by 1st Embodiment. It is a flowchart which shows operation | movement of the fingering production | generation thread | sled by 1st Embodiment. It is a flowchart which shows the operation | movement of the fingering series selection process by 1st Embodiment. It is a flowchart which shows operation | movement of the main routine by 2nd Embodiment. It is a flowchart which shows operation | movement of the track processing thread | sled by 2nd Embodiment. It is a flowchart which shows the operation | movement of the phrase division | segmentation process by 2nd Embodiment. It is a flowchart which shows operation | movement of the track processing thread | sled by 3rd Embodiment.

Explanation of symbols

1 CPU
2 ROM
3 RAM
4 Input section 5 Display section 6 MIDI interface 7 Electronic musical instrument

Claims (5)

Storage means for storing performance data representing the performance parts of each sound, as well as representing each sound constituting the song;
Phrase dividing means for performing the operation of dividing the performance data stored in the storage means into phrase sections in parallel for each performance part;
A fingering information generating device, comprising: fingering generating means for generating the fingering of the phrase section divided by the phrase dividing means in parallel for each phrase section. The phrase dividing means is a setting means for setting a reference phrase divided into a predetermined number of notes from the beginning of the performance data of the performance part to be processed;
First search means for searching for a comparison phrase that matches the reference phrase after the reference phrase set by the setting means;
A second search means for searching for a comparison phrase that matches the reference phrase while shifting the position of the reference phrase when the first search means cannot search for a comparison phrase that matches the reference phrase;
And a section setting means for setting the searched comparison phrase as a phrase section when either of the first and second search means searches for a comparison phrase that matches a reference phrase. Item 10. The fingering information generation device according to Item 1. 2. The fingering information generation device according to claim 1, wherein the fingering generation unit assigns the generated fingering to the same kind of phrase section every time the fingering of the phrase section is generated. Storage means for storing performance data representing the performance parts of each sound, as well as representing each sound constituting the song;
The performance data stored in the storage means is divided into phrase sections, and the operation of calculating the average number of constituent sounds per section from the number of sounds included in each divided phrase section is performed in parallel for each performance part. Phrase dividing means;
When the phrase section divided by the phrase dividing means exceeds the average number of constituent sounds, the fingering generation operation is performed in parallel for each phrase section using the performance data divided by the average number of constituent sounds. And a fingering generating means for performing a fingering generation operation using all performance data constituting the corresponding phrase in parallel for each phrase section when the number of constituent sounds is not exceeded. Finger information generator. Represents each sound that makes up a song and divides the performance data representing the performance part corresponding to each sound into phrase sections, and calculates the average number of constituent sounds per section from the number of sounds included in each divided phrase section Phrase division processing to perform the operation to be performed in parallel for each performance part,
When the phrase section divided by the phrase dividing process exceeds the average number of constituent sounds, the operation of generating fingering using the performance data divided by the average number of constituent sounds is performed in parallel for each phrase section, and the average When the number of constituent sounds is not exceeded, the fingering generation process is performed on the computer in which the fingering generation operation is performed in parallel for each phrase section using all performance data constituting the corresponding phrase. Fingering information generation processing program.

Images