JPH11288274A - Musical performance information converter - Google Patents

Abstract

PROBLEM TO BE SOLVED: To convert a music performance information so as to be adaptable to a musical instrument which can reproduce only less channel information by moving the phrase of the other channel considering release time to a section without the phrase of a detected channel. SOLUTION: An objective sound source conversion tool 101, a music performance file checker 102, a KARAOKE file formatter 103, a background still picture compilation tool 104, a delivery music formatter 105 and a music information input tool 106 convert music produced to correspond to a synthesizer SC-88 into that corresponding to a synthesizer with the small number of channels by SC-88. In such a case, the phrases of respective channels and a section with few phrases are detected. The phrase of the other channel which is in the same position as the section without the phrase of the channel and whose section is shorter than the period is moved to the section without the phrase of the channel. Thus, the other channels are reduced and the phrases are moved so that the release time of the original phrase does not change even if they are moved.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention converts musical instrument performance information, in which performance data of a plurality of channels is described in order to play a certain musical instrument, so as to be adapted to an instrument which can reproduce only channel information less than the number of channels. The present invention relates to a performance information conversion device.

[0002]

2. Description of the Related Art Generally, as one of the fields of electronic music,
In a field called desktop music using a personal computer, automatic performances using performance information are actively performed. Further, recently, electronic karaoke apparatuses for playing electronic music using a communication line, which are called communication karaoke, are becoming popular. The performance information of these music is MIDI
(Musical Instrument Digital Interface) is transmitted using a communication protocol called MIDI.
Is stored as a file called SMF (Standard MIDI File). Also,
In order to describe information that cannot be described by the SMF depending on the system, it may be saved in a file format dedicated to each sequencer.

[0003] The sound source actually reproduced based on the performance information has different specifications depending on the type and maker. Specifically, the number of simultaneous sounds of the musical instrument voice, the number of MIDI channels, the number of timbres, the types of timbres, and the like are provided. System exclusive messages, effects, etc. are different. Therefore, if a performance data file created for a sound source of a certain model is to be reproduced using a sound source of another different model, the performance is not performed correctly. More specifically, if the number of simultaneous sounds is insufficient, a sound dropout occurs, and if the number of channels is insufficient, the performance sound of the instrument described as the channel is not output at all. If the number of timbres or the types of timbres are different, the timbre is not produced, or the timbre is played with a completely different timbre.

As a conventional method of reducing the number of MIDI channels, humans repeatedly delete an insignificant channel musically (as an arrangement) while repeating an auditory test, or search for a vacant timing with respect to each other. A method of collecting a plurality of channels by manually moving the channels is adopted.

That is, as a conventional method for adapting musical instrument performance information in which performance data of a plurality of channels is described for a certain musical instrument to an instrument that can reproduce only channel information smaller than the number of channels, a MIDI sequencer is used. Some people make corrections manually, or recreate instrument performance information from the beginning so that people can listen to the song and conform to the specifications of the sound source. As a method of adapting from a lower model to a higher model, for example, a Roland SC-5
5 to SC88, GM sound source to Roland GS standard, GM sound source to Yamaha XG standard, etc.
Automatic conversion for lower-level models includes the number of tones, number of channels,
It is considered impossible due to the difference in polyphony.

[0006]

However, the above-described method of reducing the number of MIDI channels requires a long working time, and furthermore, when a plurality of channels are combined, a group of performance information is finely divided into time units. Since each channel is searched one after another from the beginning of the music and moved one by one to a vacant portion of each channel, there is a problem that the amount of work is large and it is not possible in practice.

The present invention has been made in view of the above-mentioned problems of the prior art, and is intended to adapt musical instrument performance information in which performance data of a plurality of channels is described for a certain musical instrument to an instrument that can reproduce only channel information smaller than the number of channels. It is an object of the present invention to provide a performance information conversion device which can automatically convert the performance information into a performance information.

[0008]

SUMMARY OF THE INVENTION In order to achieve the above object, the present invention detects a section having no phrase and a phrase in each channel and places the section in the section having no phrase in the channel at the same position as the section. In addition, by moving a phrase of another channel shorter than that section, the other channels are reduced, and the phrase is moved so that the release time of the original phrase does not change even if the channel is moved. That is, according to the present invention, from the sound source performance information in which each of the plurality of performance channels has a plurality of phrases, a section for detecting a section in which there is no phrase of each channel and a phrase in consideration of a release time for each timbre, In the section where the phrase of the channel does not exist, the same channel as that section and shorter than that section, and the phrase of the other channel is moved by considering the release time, thereby reducing the other channel. Means and
The present invention provides a performance information conversion device having the same.

[0009]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a configuration diagram showing a communication karaoke system to which a target sound source conversion tool as one embodiment of a performance information conversion device according to the present invention is applied. FIG. 2 is an explanatory diagram showing a terminal player of the user personal computer of FIG. 3 is FIG.
FIG. 4 is an explanatory diagram showing the flow of files in the music file authoring apparatus of FIG. 1, and FIG. 5 is a diagram showing before and after channels (represented as channels in the figure) are reduced. FIG. 6 is an explanatory diagram showing an example of MIDI data, FIG. 6 is a flowchart for explaining a channel reduction process according to the present invention, and FIG. 7 is an explanatory diagram showing a track window immediately after a file is opened (some windows are indicated as windows in the diagram). FIG. 8 is an explanatory diagram showing a track window immediately after the transition to the channel reduction mode. FIG. 9 is an explanatory diagram showing a channel reduction window immediately after the transition to the channel reduction mode. FIG. 10 is a flowchart for explaining the phrase detection subroutine of FIG. FIG. 11 shows the contents of the release time table shown in FIG. It is a diagram. 12 is an explanatory diagram showing a dialog screen in the channel reduction subroutine of FIG. 6, FIG. 13 is a flowchart for explaining the channel reduction subroutine of FIG. 6 in detail, FIG. 14 is an explanatory diagram showing a phrase configuration, and FIG. FIG. 16 is an explanatory diagram showing a reduction window after channel reduction, FIG. 17 is an explanatory diagram showing a track window after channel reduction, and FIG.
FIG. 19 is an explanatory diagram showing a track before channel reduction, FIG. 19 is an explanatory diagram showing a track in which channels are reduced in FIG. 18, and FIG.
0 is an explanatory diagram showing a state where the track of FIG. 19 is divided, FIG. 21 is an explanatory diagram showing a channel reduction window after channel division, FIG. 22 is an explanatory diagram showing a track window after channel division, and FIG. FIG. 24 is an explanatory diagram showing a state in which channels are merged, FIG. 24 is an explanatory diagram showing a channel reduction window after channel merging, and FIG. 25 is an explanatory diagram showing a track window after channel merging.

Here, in the present invention, a file for distributing music files and music data from the server 10 to the user (hereinafter referred to as a distribution file) has a unique format (UD).
F: UserDistribution File)
Are KMF (Karaoke Master File) and SIF (Song In)
formation File) and original still image files. KMF is the actual karaoke performance data and still image switching timing file (SPT: Still
Picture Timing file).
Compressed when filed in UDF.

The system shown in FIG.
1 shows a communication karaoke system using the karaoke system. The music file authoring device 12 constitutes a performance information conversion device according to the present invention, and stores a music material database with a music file (U
DF). The server 10 is a server for searching for a music file converted by the music file authoring device 12 and downloading it to the user personal computer 16 via the network 14. User is server 1
Server 1 from the song database registered in
When a desired song is searched for using the search function of No. 0 and a purchase file is designated, the UDF is downloaded from the server 10.

The user personal computer 16 has an HDD (hard disk drive) 18 for storing the downloaded music file, and terminal player software which operates on the user personal computer 16. As shown in FIG. 2, the terminal player software (user personal computer 16) includes a sequencer 24, an image processing unit 26, a music analysis unit 28, a screen display control unit 32, a music reproduction control unit 34, and a GUI (graphical user interface). Interface) 36. A display 38 is connected to the screen display control unit 32, and a speaker 42 is connected to the music reproduction control unit 34, and a speaker 44 is connected via an external sound source 40.

The GUI 36 is an interface for the user to give an instruction to the terminal player (user personal computer) 16 while viewing the display screen of the display 38 using a mouse or keyboard (not shown). The image processing unit 26 processes a large number of still image data recorded on the CD-ROM 20 and replacement image data stored in the HDD 18, and outputs the processed image data to the screen display control unit 32, thereby displaying the image on the display 38. Let it. The music analysis unit 28 analyzes the MIDI data stored in the HDD 18 and outputs the data to the music reproduction control unit 34, so that the music is analyzed by the internal sound source and the speaker 42 of the music reproduction control unit 34, and the external sound source 40 and the speaker 44. To play.

In this system, a CD-ROM 20 distributed to users is used. This CD-ROM2
0 stores the following terminal player software, a hidden key file for preventing unauthorized copying, a still image, a software synthesizer, and network connection software. Terminal player software This is recorded on the CD-ROM 20 as an application for the karaoke player, and is actually copied to the HDD 18 on the user side by the installer and used.・ Hidden key file CD-ROM for preventing unauthorized copying
This is a security file in which an arbitrary layer of the file system is scrambled, and the terminal player 16 reads this file at the time of startup. -Still images are a group of files for displaying the background screen of karaoke compressed by the JPEG method,
An optimum background is selected from a number of still image files stored in accordance with the tune and displayed on the display 38. The designated data of the still image file is described in the KMF in the distribution file.・ Soft Synthesizer SC-
It is a software-based synthesizer that plays back MIDI files compatible with 55mkII. Here, in the karaoke system of the present invention, as an example, MI produced for SC-88 having higher performance than SC-55mkII is targeted.
The DI file is converted so as to correspond to SC-55mkII.

In this embodiment, the user personal computer 16 is described as an example of a terminal. However, the present invention is not limited to the user personal computer 16, but may be a hardware sound source unit called a sound source module. A board-type sound source board built in the personal computer 16, a card sound source, a plug-in type synthesizer built in a WWW (World Wide Web) browser, a sound source built in a game machine, etc.
It is not limited to that form. Similarly, the sequencer is not limited to the sequencer as an application of the user personal computer 16, but instead is a sequencer realized as dedicated hardware, a MIDI sequencer realized in a game machine, or a CD + MI
A form such as a MIDI recorder in which MIDI data is linearly recorded on a medium, such as DI, may be used.

Next, the target sound source conversion system 101 for reducing channels will be described. Here, when a MIDI file is produced as performance information of a song, if the MIDI data, the character data of the lyrics, the color change data, etc. are produced from the beginning, the data production cost per song is high, and Significant investment is required to prepare the lineup.

Therefore, in a series of authoring systems 100 according to the present invention, a target sound source conversion tool 101, a music performance file checker 102, a karaoke file formatter 103, and a background still image editing tool 1 shown in FIGS.
04, a music piece that is produced by the distribution music formatter 105 and the music information input tool 106 so as to correspond to the software-based synthesizer SC-88 manufactured by Roland Co., Ltd.
The conversion is performed so as to correspond to 5mkII, so that the cost is minimized. For this reason, first, regarding the characters and color change data for displaying lyrics, a song file format (KMF format)
Then, a KMF formatter 103 for automatically converting to this format is required.

Here, the conversion of characters and color change data for lyrics is relatively easy, but the MI of SC-88 is relatively easy.
MIDI data produced to use a DI sound source does not consider compatibility with lower-level models such as the SC-55 at all. FIG. 5 shows the sound source performance of both, and the lower model (SC-55) has less numbers of the number of tones, the number of simultaneous sounds, the number of channels, the number of drum sets, the number of drum parts, and the number of sound effects. Particularly, the number of channels is 16 + 16 = 32 for the upper model and 16 for the lower model.

The 88-55 conversion tool (target sound source conversion tool) 101 shown in FIGS. 3 and 4 converts MIDI data produced for the SC-88 MIDI sound source into 16 channels using 32 channels. I do. Here, in a normal sequencer, one MIDI channel is assigned to one track, and only one timbre can be simultaneously played on one MIDI channel except for a rhythm channel. That is, in one channel, it is not possible to play the tone of the guitar at the same time as the tone of the piano, for example, so that another channel is required in this case.

[0020]

In the general processing of the present invention, when this 88-55 conversion application is started as shown in FIG.
Reading of a file (Step S101) Display of a track window (Step S102) Reduction of voices other than channel reduction, tone conversion (Step S10)
3) Switch to channel reduction mode (step S10)
4) ・ Phrase detection (Step S105) ・ Channel movement according to attribute (Step S106) ・ Display of channel reduction window, rewrite and redisplay of track window (Step S107) ・ Channel reduction (Step S1)
08) Channel division (step S109) and channel merge (step S110) are executed.

Here, as a method of starting the above processing, a menu format in which an operator selects “Channel reduction” while the menu of the 88-55 conversion application is displayed may be used, or an application dedicated to channel reduction may be used. May be launched. When this application is started, a MIDI file is read (step S101), and a track window as shown in FIG. 7 is displayed (step S102). In the track display window after the file is opened, the tracks are displayed in the order of the track chunks recorded in the SMF.

Next, after the reduction of voices and the like other than the channel reduction and the tone conversion (step S103) are executed, the mode is shifted to the channel reduction mode (step S104), and the track window as shown in FIG. 8 and the track window as shown in FIG. Display the channel reduction window. In the track display window and the channel reduction window immediately after the transition to the channel reduction mode, MIDI channels are displayed in order. In this state, channel reduction has not been performed yet.

Next, as a preparation for channel reduction, a phrase list is created for each MIDI channel to be subjected to channel reduction. This phrase list is a set list of phrase start events and phrase end events of a certain channel from the sequence data of the document before reduction.
Here, although shown in a bar graph form in FIGS. 8 and 9, they are actually displayed in color. When shifting to the channel reduction mode, the channels having the attributes of V {vocal melody (main)}, D {duet melody (sub)}, and C (chorus) are numbered in the order of V, D, and C. Move to a young channel.

The reason for this is to prevent the channels having these attributes V, D, and C from being dispersed as much as possible, thereby avoiding an unnecessary increase in the total number of tracks. The reason for avoiding this is that, when the total number of tracks increases, the load of subsequent conversion processing to a distribution file (UDF) and processing of the terminal player 16 increases. The reason why it becomes difficult to disperse when the channel number becomes younger will become clear when explaining the algorithm described later. In short, when moving each phrase between channels, the closer the channel is to the head, the more the phrase becomes. This is because the channel becomes the “filled side” and is buried in the gap of the channel near the head in order from the channel with the larger number (and from the B port).

When each channel is displayed in the channel reduction window shown in FIG. 9, the track name is removed because it is irrelevant, and the port names A and B, the channel numbers (the numbers following the port names A and B) are simply displayed. The number of phrases (the number in parentheses in the figure) is displayed.

Next, the phrase detection in step S105 will be described in detail with reference to FIG. First, S
The contents of the track described first in the MF are read (step S1). The SMF is composed of a header chunk and a track chunk, and the contents described here are all about the contents of the track chunk.

[0027]

The performance data track, especially the attribute information, is configured as follows. -Output MIDI port (1 byte) ASCII uppercase alphabetical order from "A". -Output MIDI channel (2 bytes) "01"-
"16". Decimal ASCII notation. -Attribute (1 byte) ASCII uppercase notation. V: Vocal melody (main) D: Duet melody (sub) R: Rhythm track C: Chorus (alternative) track N: Non Transpose
-Space (1 byte) corresponds to ASCII "20h". (H represents hexadecimal notation) Part name (0 to n bytes) Indicates the tone name, instrument name, or part name.

[0028]

Next, it is determined whether or not the track is a rhythm track R based on the attribute information in the track ID (step S2).
To read the attribute information in the next track ID,
It returns to step S2. Here, 'Use For Rhythm Par
All the attributes of the tracks set to t 'are R. If it is other than the rhythm track R in step S2, the process proceeds to step S3, where the number of simultaneous sounds = 0, the phrase number N = 1, the tone color address reset, the total sounding end time = 0tick
Reset s (tick) current time to 0 ticks.

Next, a MIDI event is read (step S4), and then the current time (hereinafter, the time indicates a position in the music represented by the total number of ticks from the beginning of the music) is updated (step S5). Next, the content of the MIDI event is determined (step S6). In the case of the bank select message and the program change (PC) message, the tone address of the tone generator assigned to each point is determined based on the output port number (step S8), and the next MIDI event is read. The process proceeds to preparation (step S7), and returns to step S4 to read the next MIDI event.

In the case of a note-off message, a hold-on message, and a hold-off message, a simultaneous sounding count value is calculated, and a total sounding end time is calculated based on a release time table T shown in detail in FIG. (Step S9) Then, the process proceeds to preparation for reading the next MIDI event (Step S7), and returns to Step S4 to read the next MIDI event. If it is the end of the track, the process branches to step S10 to read the attribute information in the next track ID, and returns to step S2.

In the case of a note-on message, the process proceeds to step S11 and the following steps. In step S11, if the simultaneous tone count value is 1 or more, the flow branches to step S9. On the other hand, if it is "0", the flow proceeds to step S12. . Then, in step S12, when the all sound ending time is "0", the process jumps to step S15, and when it is not "0", the process proceeds to step S13. Then, step S1
In 3, it is determined whether (current time)-(all sound generation end time)> (set phrase interval), and if No, the process branches to step S9. On the other hand, if Yes, the process proceeds to step S14. In step S14, the end time of the phrase N is set to the end time of all the sounds N = N + 1, and then the start time of the phrase N is set to the current time (step S1).
5) Then, proceed to step S9.

That is, in step S3 and subsequent steps, the phrases are determined in order. At this time, the phrase is first detected, so that the phrase starts only when note-on information is present. Although not shown in the figure, note-off information is obtained when the velocity is "0" even in the case of a note-on message. Then, as a basic process, if there is at least one sounding state, it is determined that the phrase has started, and the number of simultaneous sounds becomes "0". Then, the last note-off message is output in a state where the sound is not held. When a predetermined time interval (the default value is 480 ticks) is available before the next sounding, one phrase ends.

Here, "release" is a term related to pronunciation of an electronic musical instrument. In the present invention, when the sound is still left as a lingering sound even when the keyboard is released, this is defined as "release", and this volume gradually decreases and is reduced to a level that cannot be recognized by hearing. At the time, or when the use as a sound resource in the sound processing is released, processing for terminating the phrase is performed. Similarly, hold-on and hold-off messages are necessary for recognizing a phrase. Even if a note-off (stop sounding) message appears, if the hold-on message is valid after the note-on Keeps the tone hold level until a hold-off message appears. Note that the MIDI tone colors are mainly provided with A, D, S, and R envelopes, but “release” is represented by R.

The hold-on message includes the above-mentioned A, D,
It works so as to maintain the level of S of S and R, that is, the level of “sustain”, and in a state where the hold message is on, the release state is not immediately set even if a note-off event appears. It is normal to go into the release state from the moment when the hold-off message appears after the appearance of the event. If there is a sound source that performs other processing, the release time may be determined according to the length of time that the sound is attenuated after the appearance of the note-off event.

In the release time table T, for example, as shown in FIG. 11, a release time (release margin variation in units of msec) is stored for each tone such as a piano. And FIG.
In step S9 indicated by 0, the release time corresponding to the timbre of the moving phrase is read from the table T, and this is added to the time of the note-off event of the phrase to make the entire sound generation end time. Therefore, of course, the length of each phrase with the release time taken into account is usually longer (and sometimes the same) than not taken into account. Regardless, it is possible to prevent a problem in which the sound is interrupted and the sound is interrupted because it is erroneously determined to be another phrase.

The release time table T shown in FIG. 11 is stored independently as a table file in the same folder (directory) as the application so that the application can refer to the table as another table. Therefore, when the operation of correcting the release time is performed, only the release time table T can be corrected without changing the application itself. However, the form of the application is not limited to this, and the application may be integrated with the table T. The table T is stored in a separately designated folder (directory) instead of in the same folder (directory) as the application. Is also good.

The release time table T shown in FIG. 11 may be used when counting and reducing the number of simultaneous sounds. That is, in the 88-55 conversion processing, the number of simultaneous sounds is changed from 64 voices to 28 as shown in FIG.
Voices, but according to the concept of voices,
Just counting the number of notes that are on at the same time does not actually count the sound that remains as release after the note is off, so if you switch to another sound assignment during release, the sound will be released and attenuated The problem of interruption occurs. Therefore, when the number of simultaneous sounds is reduced by the 88-55 conversion tool 101, the table T
By counting the number of simultaneous sounds in consideration of the release time of, the number of simultaneous sounds can be reduced while maintaining the performance information optimally.

Here, the release time of the table T shown in FIG. 11 is indicated by msec, but since the actual times of the phrases are all expressed by Ticks, the table T can be similarly stored by Ticks. Desirably, the value of Ticks varies in length according to the tempo of the song, so table T
According to the tempo of the song when memorizing in msec and reading it out
It may be converted to Ticks. Also, since the release time is not related to the tempo of the song, if the release time of the sound source changes according to the current tempo or previous sound, it is possible to output the optimal value according to the condition. You just need to prepare tables and functions that can be used.

In the actual table T, the pitch is basically measured at six points C1, C2, C3, C4, C5 and C6, and the longest velocity value is stored.
However, the release of a certain pitch may be extremely long depending on the timbre, so that such a timbre is stored in an appropriately shortened manner. Of course, a table whose value dynamically changes according to all intervals may be used, or a function that outputs a release time when an interval is input may be used instead of the table. Although the velocity is measured in the range of 100 to 110, a table or a function that dynamically changes may be used.
Also, some timbres, such as bass sounds, do not sound in the range of C1 to C6, so for timbres with a narrow range, representative samples may be extracted at smaller intervals.

Furthermore, the release time is the length of time from the time of note-off until the volume on the auditory sense becomes "0". In actuality, release may continue at a lower volume until then. . Therefore, a point in time when the sound volume attenuates to an optimum sound volume in accordance with the balance of the sound volume level with the whole sound may be determined to be the release end.

By the way, the values stored in the table T are stored based on the result of measuring the timbre of the converted sound source (SC-55), not the sound source (SC-88) before conversion. Therefore, when the number of channels is reduced to correspond to a sound source other than the SC-55, the timbre of the sound source is measured. Also, in the table T shown in FIG. 11, one value is stored for one tone color. However, since the actual release time differs depending on the range and velocity, one value is stored.
A value may be stored for each of a plurality of different tone ranges (or pitches) or velocities for one tone color. Also, SC
The release time may be obtained based on the SC-55 tone contained in the -88.

With the above processing, preparation for channel reduction is completed, and then the process proceeds to actual channel reduction processing. This processing is performed in three stages, namely, channel reduction (step S
108), channel division (step S109), and channel merge (step S110). The processing of each stage is executed by a channel reduction command, a channel division command, and a channel merge command, respectively.

(1) Channel reduction command In response to the channel reduction command, a dialog screen for each of 16 channels of MIDI used for two ports A and B is displayed as shown in FIG. In this embodiment, the merging is performed excluding the rhythm channel R, and thus the 'Use For Rhythm P
The channel for which 'art' is specified is gray in the dialog (channels “10” and “1” of port A in the figure).
1)). The operator selects the reduction target channel in this display state and presses the execution button. Normally, all channels are selected. However, if there is a channel not particularly desired to be merged, the check mark “x” of that channel is deleted and the execution button is pressed.

[0044]

Next, the channel reduction processing will be described with reference to FIGS. First, in step S21, the reduction target channel number CNOr = the least significant channel merge target channel number CNOm = 1, the phrase start time Tb (CNO,
PNO) = 0 ticks Phrase end time Te (CNO, P
NO) = 0 ticks Phrase number P of the channel to be merged
NOm = 1 Phrase number of the channel to be reduced PNOr =
Set to 1.

In the following step S22, Tb (CNO
r, PNOr)> Te (CNOm, PNOm) + Tps
However, it is determined whether or not Tps is a time required for phrase setup, and if YES, a flag (End-o) is set.
k) = 1 (step S25), then increment the phrase number PNOm of the channel to be merged (step S26), and then return to step S22.

If NO in step S22,
It is determined whether or not the flag (End-ok) = 1 (step S23). In the case of 1, the process proceeds to step S24 and the following, and in the case of 0, the process proceeds to step S29. Step S2
In the case of 4 or less, first, the flag (End-ok) is reset to 0, and then Te (CNOr, PNOr) + Tps <T
b (CNOm, PNOm) is determined (step S
27) If YES, move the phrase to P
C is added, the phrases CNOm and CNOr are renumbered (step S28), and the process proceeds to step S29. On the other hand, if NO in step S27, the process directly proceeds to step S29. Here, PC indicates a MIDI standard program change, which is a message relating to tone color switching.

In step S29, the channel to be reduced CN
It is determined whether or not it is the last phrase of Or. If NO, the phrase number PNOr of the channel to be reduced is incremented (step S30), and the process returns to step S22. When processing is performed up to the last phrase of the channel to be reduced CNNo, the process proceeds to YES in step S29, and it is determined whether or not channel to be reduced CNNo = channel to be merged CNOm + 1 in step S31. The number PNOr = 1 is set to the merge target channel CNOm = CNOm + 1 (step S32), and then the process returns to step S22.

[0048]

If the channel to be reduced CNNO = the channel to be merged CNOm + 1 in step S31, the process proceeds to step S33, and the channel to be reduced CNNO =
CNOr + 1 Merge target channel CNOm = 1 Phrase start time Tb (CNO, PNO) = 0 ticks Phrase end time Te (CNO, PNO) = 0 ticks Phrase number PNOm of merge target channel CNOm = 1 Phrase number PNOr = 1 of reduction target channel CNOr Set (step S33), and then return to step S22.

That is, in this process, when a phrase list is created for each MIDI channel, the source (reduction target channel) and destination (merge target channel) MIDI channels (and ports A and B) are determined in order and looped. Is created, and if a moveable phrase is found from the source to the destination in each loop, the phrase is moved. At this time, the lowest channel is set as a reduction target channel, and an empty area of the top merge target channel is searched for. When a free area is found, the phrase is moved to the free area. At this time, the numbers of the channels to be reduced and the channels to be merged are assigned from the beginning (renumbering). Also,
If all the phrases of the reduction target channel cannot be moved to the top merge target channel, an empty area is searched from the second merge target channel. In this way, an empty area is searched up to the channel immediately before the channel to be reduced, and when the processing is completed up to this point, the channel to be reduced is moved up by one, and an empty area is searched from the top merged channel.

As shown in FIG. 26, when the phrase is detected, the polyphony becomes "0", the last note-off (or hold-off) message is output, and the time until the next first note-on message is output. When detecting the interval, the “phrase setup” is also taken into consideration. In addition, the phrase detection may be performed after subtracting the start time of each phrase by the time of “phrase setup” in consideration of the time of “phrase setup”. In this case, the polyphony becomes “0”, the last note-off (or hold-off) message is output, and the detection of the time interval up to the next first note-on message requires this “phrase”. You don't need to consider the “setup” time.

Here, as shown in FIG. 15, when the phrase PNOr of the channel CNOr to be reduced is moved to the free area of the channel CNOm to be merged, the position on the time axis cannot be moved, and Direction). This is because moving the position on the time axis results in different performance contents. At this time, the determination as to whether or not it fits in the free area always takes the time Tps required for phrase setup into consideration. This time T
If it is possible to move in consideration of ps, a message regarding tone change is created and the phrase setup time Tps
Add inside.

In step S22 in FIG. 13, the leading phrase PNO1 of the reduction target channel CNOr
Is located after the end position of the phrase of the merge target channel CNOm to which the phrase setup time Tps has been added. Here, the flag (End-ok) in FIG. 13 becomes “1” when at least one phrase corresponding to the above exists, and the step S
23, if the flag (End-ok) = 1, it is determined in step S27 whether or not the phrase immediately before the position to be inserted is located after the phrase PNOr of the reduction target channel CNOr, that is, whether or not the phrase is located at the end position. To check.

FIGS. 16 and 17 show a reduction window and a track window after channel reduction, respectively. As shown in FIG. 16, since the phrase is embedded from the head channel A-01 by executing the channel reduction command as described above, the number of phrases (shown in FIG. 43) of the head channel A-01 is the largest. FIG. 17 shows the track window in the state shown in FIG. 16, in which the position of each phrase in each track is displayed in color. Here, the track windows shown in FIGS. 8 and 17 are the same, but have different track names and colors. In the case of a single destination channel, the track name shown in FIG. 17 is changed to the destination channel as it is, and in the case of a plurality of destination channels, “*” is displayed. Since the display colors are classified for each destination channel, the color of the track having a larger number of destination channels is more classified.

(2) Channel division In the above-described channel reduction, a MIDI channel (and port) of a phrase is used.
, The data is not moved to another track, and as a result, the track having the moved phrase becomes a multi-channel track in which one track has data on a plurality of channels. At this point, the track composition does not change,
The state of the sequence data as music data is maintained. However, since the multi-channel is prohibited in the finally required data form, the multi-channel track created by the above-described channel reduction is divided for each MIDI channel. Here, each track at this time specifies a certain part on the music, but a certain part may be represented by a plurality of tracks.

FIG. 18 shows a track before channel reduction, and FIG. 19 shows a case where channel X in FIG. 18 is reduced. FIG. 20 shows a case where track A in FIG. 19 is divided into tracks A-1 and A-2. 21 and 22 show a channel reduction window and a track window after channel division, respectively.

(3) Channel Merge When a multi-channel track is divided for each MIDI channel,
Multiple tracks with DI channels (and ports) are created. Here, since the data format finally required has an upper limit of the number of tracks, the same M number is set so that the waste of the tracks is eliminated and the upper limit of the number of tracks is not exceeded.
A plurality of tracks having IDI channels (and ports) are combined into one track (merged). FIG. 23 shows a case where the divided tracks A-1 and A-2 are merged into tracks B and C, respectively, as shown in FIG. 24 and 25 show a channel reduction window and a track window after channel merging, respectively. Here, when merging tracks like this,
Since the relationship between the track and the part is no longer maintained, the data format of the sequence data as a music score collapses, but the original state is maintained as data for the computer to play.

[0057]

As described above, according to the present invention, a phrase of each channel and a section having no phrase are detected,
In a section where there is no phrase in the channel, move the phrase of another channel that is at the same position as the section and shorter than that section to reduce other channels and release the original phrase even if it moves The instrument performance information, which describes the performance data of multiple channels for a given instrument, is automatically adjusted so that it matches the instrument that can reproduce only less channel information than this number of channels. Can be converted to

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing a communication karaoke system to which a target sound source conversion tool as one embodiment of a performance information conversion device according to the present invention is applied.

FIG. 2 is an explanatory diagram showing a terminal player of the user personal computer of FIG. 1;

FIG. 3 is a configuration diagram showing a music file authoring device in the system of FIG. 1 in detail;

FIG. 4 is an explanatory diagram showing a file flow in the music file authoring device of FIG. 1;

FIG. 5 is an explanatory diagram showing an example of MIDI data before and after channel reduction.

FIG. 6 is a flowchart illustrating a channel reduction process according to the present invention.

FIG. 7 is an explanatory diagram showing a track window immediately after a file is opened.

FIG. 8 is an explanatory diagram showing a track window immediately after transition to a channel reduction mode.

FIG. 9 is an explanatory diagram illustrating a channel reduction window immediately after transition to a channel reduction mode.

FIG. 10 is a flowchart illustrating a phrase detection subroutine of FIG. 6 in detail.

FIG. 11 is an explanatory diagram showing contents of a release time table of FIG. 10;

FIG. 12 is an explanatory diagram showing a dialog screen in a channel reduction subroutine of FIG. 6;

FIG. 13 is a flowchart illustrating the channel reduction subroutine of FIG. 6 in detail.

FIG. 14 is an explanatory diagram showing a configuration of a phrase.

FIG. 15 is an explanatory diagram supplementing processing in a channel reduction subroutine of FIG. 13;

FIG. 16 is an explanatory diagram showing a reduction window after channel reduction.

FIG. 17 is an explanatory diagram showing a track window after channel reduction.

FIG. 18 is an explanatory diagram showing a track before channel reduction.

FIG. 19 is an explanatory diagram showing a track in which the channels in FIG. 18 are reduced.

FIG. 20 is an explanatory diagram showing a state in which the track in FIG. 19 is divided.

FIG. 21 is an explanatory diagram showing a channel reduction window after channel division.

FIG. 22 is an explanatory diagram showing a track window after channel division.

FIG. 23 is an explanatory diagram showing a state where the channels of FIG. 20 are merged.

FIG. 24 is an explanatory diagram showing a channel reduction window after channel merging.

FIG. 25 is an explanatory diagram showing a track window after channel merging.

FIG. 26 is an explanatory diagram showing the structure of a phrase.

[Explanation of symbols]

 101 Target sound source conversion tool

Claims (5)

[Claims] 1. A means for detecting, from sound source performance information in which each of a plurality of performance channels has a plurality of phrases, a section in which there is no phrase of each channel, and a phrase in consideration of a release time for each timbre, Channel reducing means for reducing the other channel by moving a phrase of another channel which is located at the same position as the section and is shorter than the section, and in consideration of the release time, in a section having no phrase of A performance information conversion device having: 2. The method according to claim 1, wherein the channel reducing unit moves a phrase of another channel when a section having no phrase of the detected channel is equal to or longer than a section set in advance in consideration of the release time. The performance information conversion device according to claim 1. 3. A table for storing a margin time in consideration of a release time for each timbre, wherein the channel reducing means is configured to store a margin time for each timbre stored in the table at a time of a note-off event of a phrase of the other channel. 3. The performance information conversion apparatus according to claim 1, wherein the movement is performed after the margin time is added to set the entire sound generation end time. 4. The table stores a margin time for each of a plurality of intervals of each tone color,
4. The performance according to claim 3, wherein a margin time for each interval stored in the table is added to the time of the note-off event of the phrase of the other channel, and the time is set after setting the end time of all sounds. Information conversion device. 5. The table stores a margin time for each of a plurality of dynamics of each tone color, and the channel reduction unit stores the margin time in the table at the time of a note-off event of the phrase of the other channel. 5. The performance information conversion device according to claim 3, wherein the movement is performed after adding a margin time for each strength and setting the end time of all sounds.