JP3219150B2 - Performance information compression method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to compression of performance information.
In particular, the present invention relates to a compression method for reducing the data amount of performance information for designating a pitch and a sounding time of a sound constituting a musical tone.

[0002]

2. Description of the Related Art There is, for example, a MIDI signal as performance information for designating a pitch and a sounding time of a sound constituting a musical tone. MI
The DI signal is stored in a memory or the like or transmitted in the form of an SMF (Standard MIDI File). Such a MIDI signal is used in a system called a communication karaoke instead of a conventional karaoke apparatus of a system for reading data recorded on a storage medium such as an optical disk.

In this communication karaoke, if karaoke music, lyrics, related images (moving images or still images), etc. are transmitted as they are in the form of signal information output from an optical disk or the like, the amount of data is enormous. The information is transmitted as a MIDI signal from the station to a terminal provided at the place of use of each user, and the lyrics are transmitted as character codes, and data of desired music is distributed as needed. As such an example, a “JOYSOUND” system manufactured by Exing Co., Ltd. is known. A technique for efficiently processing the repetition portion of the melody score data composed of notes and signals is disclosed in
No. 1-91697, which is referred to as MID
It is not based on the I signal, but on the assumption that there is a score using symbols such as repetition.
It is not a technique for compressing existing performance information represented by MF. Further, Japanese Patent Application Laid-Open No. 4-147192 discloses a technique for patterning a repetition portion of a musical score. However, sufficient data compression cannot be expected.

[0004] The systems proposed so far, including the above-mentioned conventional communication karaoke system, are classified as shown in Table 1 below.

[0005]

[Table 1]

[0006]

As can be seen from Table 1, when a MIDI signal is used, the quality of the sound reproduced from a CD, an LD or the like is not the same, but if the quality is to be improved, the data is not good. It can be seen that the amount increases. If the amount of data is large, real-time communication becomes impossible unless a high-speed digital line such as ISDN is used, and the required memory capacity is large even for data storage and storage, leading to an increase in cost. Such a problem is an obstacle to the spread of the communication karaoke system to ordinary households. In addition to communication problems, a so-called package karaoke system, in which performance information of a large number of music pieces is stored in a large-capacity memory such as a hard disk, is required to reduce the data amount.

As music is said to be a repetitive art, performance is often repeated in a fixed pattern. However, in the performance information such as the MIDI signal, the permutation of successive events occurring on the time axis is handled as it is, so that the data amount is compressed only by a general digital signal compression technique. For this reason, in a communication karaoke system that distributes performance information or a karaoke system in a package format in which data of many songs are stored in a memory, etc.
There is a need for cost reduction and reduction in communication time by compressing and reducing the amount of data.

Therefore, the present invention focuses on the fact that the pitch and the sounding time of the sound constituting the musical tone are repeated in a fixed time unit and there is an overlapping portion, and the repetitive portion is detected.
In addition to deleting the signal, a signal indicating the position of the deleted repetition portion is inserted so that the compressed information can be correctly reproduced at the time of reproduction.

[0009]

SUMMARY OF THE INVENTION In order to achieve the above object, the present invention focuses on the fact that performance information is repeated in a fixed time unit and there is an overlap, detects such a repeat, and deletes it. At the same time, a signal indicating the position or the like of the deleted repeated portion is inserted so that the compressed information can be correctly reproduced at the time of reproduction. At the time of reproduction, the compressed performance information is decompressed using this signal, and the original information is expanded. This is to obtain performance information.

That is, according to the present invention, the performance information for designating the pitch of the sound constituting the musical tone and the sounding time is sequentially compared in the unit of a part from the head part of the part string divided into a plurality of parts. A first step of creating a file A in which the same part number is assigned to a part of substantially the same data in the part column, and when a part number is checked from the beginning of the file A and a duplicate number appears Deletes the duplicate part and the part number assigned to the part, and deletes the new file B
And comparing part numbers of the file A and the file B from the head part in part units. When a part number different from each other appears, a different part number in the file B is used. , A first jump mark indicating that a part is to be moved is inserted, and a part number in the file B that is identical to a different part number in the file A is immediately before the first jump mark. The third step of inserting the first tag as the movement destination, the part after the different part number in the file A and the part of the file B after the first tag is inserted are continuously performed. The part numbers are compared on a part-by-part basis, and when a part number different from each other appears, the part A second jump mark indicating that the second jump mark is to be moved is inserted, and the destination of the second jump mark is immediately before the same part number in the file B as a different part number in the file A. And a fourth step of inserting a second tag.

[0011] Also, data is sequentially compared in units of parts from the first part of a part sequence in which performance information for designating a pitch and a sounding time of a sound constituting a musical tone is divided into a plurality of parts, and substantially the same data is obtained. A first step of creating a file A by assigning the same part number to a data part, a second step of creating a file C having no initial contents, and a file C indicating a position in the file C. A third step of setting a pointer at the end of the file C, a fourth step of creating a part list having no initial contents, and a fifth step of setting a first part number of the file A to a variable i. A part number confirmation step of determining whether or not the variable i is in the part list; and a part number indicated by the variable i in the part number confirmation step. If not, the pointer position first determination step for determining whether the pointer is at the end of the file C, and the pointer position first determination step determines that the pointer is not at the end of the file C. Then, the jump mark JUMP (n) is inserted at the current position of the pointer, the pointer is moved to the end of the file C, and then the jump mark JUMP (n) is moved to the position of the pointer. Insert a tag TAG (n) and set the variable n to 1
When the pointer is determined to be at the end of the file C in the sixth step of incrementing by one and the pointer position first determining step, the file C
A seventh step of adding performance information of a part indicated by the variable i of the file A to the file A, adding the variable i to the part list, and moving the pointer to the end of the file C; A pointer position second determining step of determining whether the pointer is at the head of the part of the file C indicated by the variable i when it is determined in the confirmation step that the part number indicated by the variable i is present; If it is determined in the pointer position second determination step that the pointer is not at the beginning of the part indicated by the variable i of the file C, the jump mark JUMP is set to the position where the pointer is currently located.
(N) is inserted, the pointer is moved to the head position of the part indicated by the variable i of the file C, and then a tag TAG (n) is inserted at the position of the pointer, and the variable n is incremented by one. In the eighth step and the pointer position second determining step, when it is determined that the pointer is at the head of the part indicated by the variable i of the file C, the pointer is moved to the head position of the next part. After the ninth step, the seventh step and the ninth step, the next part in the file A is searched, and if it exists, the next part number of the file A is stored in the variable i. A next part number setting step of setting the part number from the part number confirmation step until the next part of the file A no longer exists. Providing performance information compression method which is characterized in that so as to repeatedly execute until constant step.

[0012]

[0013]

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a flowchart showing a first embodiment of a performance information compression method according to the present invention. Before describing the flowchart of FIG. 1, a communication karaoke system as an example to which the present invention is applied will be described with reference to FIGS.
This will be described with reference to FIG. FIG. 2 is a schematic block diagram showing an example of a communication karaoke system in which MIDI information as karaoke information is compressed by a performance information compression method of the present invention and transmitted from a station (host) to each terminal. FIG. 3 shows two types of transmission / storage files A and B in which both data not compressed by the station and data compressed in advance by the performance information compression method of the present invention are prepared according to the capabilities of the terminal and the line. FIG. 6 is a schematic block diagram showing a configuration example in which two files are used properly. The communication line (network) determines which file is read and transmitted.
Is determined by determining whether the line is a high-speed line such as ISDN or a general public line, or by detecting the extension capability of the terminal. In each of the systems shown in FIGS. 2 and 3, karaoke information for several thousand songs is stored in advance in a memory such as a hard disk on the terminal side as MIDI information compressed by the performance information compression method of the present invention. Can be. In this case, information on a new music or a music with a low request frequency that is not stored in the memory of the terminal can be transmitted from the station to the terminal using the communication line.

FIG. 4 shows an SM describing performance information of one piece of music.
FIG. 4 is a block diagram showing the general concept of a procedure for creating a compressed file by compressing F. The SMFs for a large number of music pieces are stored in a memory such as a hard disk, and among them, the SMF 1 for one music piece specified by a control system (not shown)
The signal is separated for each channel in a DI CH (channel) separation block 2. In other words, the instrument to be played (sound source)
The following processing is performed for each channel corresponding to. Block 3 detects (@ t + event) block repeat. That is, when there is a repetition (repeat) in a performance information pattern described later, this is detected. The next block 4 performs an R (repeat) encoding process, in which a process of deleting a duplicated portion detected in the block 3 so as not to be duplicated and adding a necessary signal is performed. The next merge processing block 5 combines the signals on which multi-channel R encoding processing has been performed as in the original SMF format. As a result, Δt (delta time) is unified and compressed. File 6 is created.

FIG. 5 is a diagram showing the manner of change of the data sequence when a repeated file is detected from the description in the SMF, the duplicated portion is deleted, and a compressed file obtained as a result of compression is created. In FIG. 5, "EVENT" indicates the "event" of the MIDI signal, "T" indicates time information similar to "@t",
Fig. 6 shows the format of the information "T" or "R" shown in Fig. 5. This data has an 8-bit configuration as shown in the figure. 7 shows, as an example, a musical score having a repeated portion, in which the second bar and the third bar have the same pattern.

Returning to the flowchart of FIG. 1, a first embodiment of the performance information compression method of the present invention will be described.
This flow corresponds to blocks 3 and 4 shown in FIG. In step S1, △
t, and at step S2, {t (e) is the sum total}.
Δt (e) and the variable e indicating each event are set to 0
Initialize to At step S3, Δt (e) is read, at step S4, Δt (e) is incremented, and at step S5, it is determined whether Δt (e) has exceeded 1920. This number 1920 is the quarter note 1
The number of ticks (Ticks) per piece is 480, and 1
Measures are four quarter notes long (ie, the time signature is 4 /
This corresponds to the length of one measure in the case of 4). Tic
k is the minimum time unit in SMF. This step S5 divides the SMF, which is performance information, into parts (one bar in this example) in predetermined time units. Step S6 is repeatedly executed until one bar elapses in step S5, and an event map (EVENT MAP)
Create i. The event map i is a set of events for each part. In step S7, e is incremented by one. When one measure ends, 1920 is subtracted from the total of Δt in step S8, and the next measure (i
It is determined whether or not the content of the event map (i-1) of the current bar i matches the content of the event map i of the current bar i.
If they do not match, i is incremented by one in step S11. If they match, the last part of the duplicate data is deleted in step S10, and a repeat event (Repeat
Event). The repeat event corresponds to "R" shown in FIG. 5, and has the data format of FIG. In FIG. 6, (ID1, ID0) = (0, 0) and (ID1, ID0)
(1, ID0) = (0, 1) when T, (ID1, ID
0) = (1,0) and (ID1, ID0) = (1,1)
Is R. The simple repeat and the section designation repeat are performed on a channel basis, and the channel is determined based on an event immediately before the repeat byte. -Δt in FIG.
s is minus △ t-start,-△ te is minus △
This is the meaning of t-end, and-△ ts and-マ イ ナ ス te are represented by absolute values while taking minus. Step S
In step 10, the number of repeat detections is counted and described as the number of repeats during the repeat event in FIG. After step S10, step S11 is executed. Step S
After step 11, e is set to 0 in step S12, and the process proceeds to step S6. In step S13, it is determined whether or not an event map has been created for all data.
If so, the flow ends.

When the performance information of the music piece shown in the staff example of FIG. 7 is processed according to the flow of FIG. 1, a comparison of the event maps indicates that the repeat of the part having the same pitch and the same sounding time is repeated in the second bar. The third bar is detected. FIG.
In the embodiment of the present invention, when the change pattern of the performance information in the time axis direction continues for two or more parts, that is, when the contents of the event map completely match, it is determined that there is a repetition. Even when a natural match is detected, it is regarded as repetition, and different parts can be described as they are.

Further, the accuracy of detecting the presence / absence of matching between parts may be made coarse, that is, a predetermined error tolerance may be provided in the comparison between the pitch and the sounding time. By doing so, slight noise or misplay may cause
Although it is originally a repeated portion, it can be prevented from being recognized as a non-identical pattern. As a result, the number of parts recognized as repeated can be increased, and the compression efficiency can be increased. Furthermore, the fluctuation of the time axis during the performance,
Alternatively, the compression efficiency can be increased by widening an allowable range such as fluctuations in performance intensity so as not to affect the performance reproduction (at the time of reproduction). In addition to the comparison between parts, when a predetermined pitch repeatedly appears in the same part, for example, in the case of performance information of a percussion instrument such as a drum, this is detected and deleted so as not to be duplicated,
The same processing as the processing shown in FIG. 1 can also be described for each pitch. In the present invention, the detection is performed for each part in a predetermined time unit. However, this one part does not necessarily have to be the same as one bar, and may be a fraction or a multiple thereof. Furthermore, two or more types of part time are provided, and for example, while detecting one bar unit, one-half
It is also possible to detect repetition even in the unit of time.

In the first embodiment shown in FIG. 1, it is determined that the repetition is performed when the contents of the continuous parts match, but the melody over several measures is often repeated. . A method that can be compressed even in such a case will be described in a second embodiment and a third embodiment. FIG.
FIG. 3 is a diagram schematically showing an SMF before compression consisting of eight parts. In FIG. 23, the second to fourth parts and the fifth to seventh parts overlap. Therefore, in order to eliminate redundancy, as one possible method, if three overlapping parts are deleted, only five parts are obtained as shown in FIG. In this case, "if the playback is performed for the time of (t1 + t2 + t3 + t4) from the beginning, jump to the point of time lapse of t1 from the beginning"
Will be described. That is, a description is provided in which the sum of Δt of each part is calculated in advance, and the reproduction position is reset according to the elapsed time from the reproduction start time. However, according to this method, when it becomes necessary to re-edit the part, the description section must be edited similarly. Also, the more complicated the description, the more complicated the recalculation of time becomes.

In the above-mentioned Japanese Patent Application Laid-Open No. 61-91697, the abbreviations of a musical score are coded and stored together with note information, and control is performed in accordance with the abbreviations during reproduction. According to this, although the problem of redundancy and the problem of reediting are solved, the composition of modern music is complicated, and there is a case where the music composition cannot be correctly expressed by simply coding the abbreviations. As in the example shown in FIG.
D. S. (Darseño) It is common to skip repetition when returning and performing again, but it is impossible with the technique of the above publication to describe selective performance information indicated by annotations. Further, the technique of the above publication presupposes a musical score using symbols such as repetition as described above.

Therefore, in the performance information compressing method according to the second and third embodiments of the present invention, the performance information described in a serial manner including the same part which is described a plurality of times is compared with the performance information. By providing a jump mark and a tag that can be distinguished from each other, the repetition of the music is indicated, so that the description is made with a small amount of information. In addition, complicated performance information can be described by including a condition value (level) in the jump mark.

FIG. 13 is a diagram showing the configuration of a jump mark and a tag. Both are non-performance information ID, channel I
D and a level ID. The level ID identifies the same mark or indicates the order.
In both the second embodiment and the third embodiment, when performance information over a plurality of parts (measures in each embodiment) is repeated, the repetition is deleted and jump marks and tags are inserted. Is to make In the second embodiment, in addition to the original file A, a non-duplicate file B
Is required to be created once, and the number of procedures is larger than that of the third embodiment described later. However, since it is appropriate for understanding the procedure of compression, the second embodiment will be described first, and then the third embodiment having higher utility will be described.
14 and 15 are diagrams showing the concept of detecting and deleting the same part. When a song described by seven measures and abbreviations is developed as shown in a part of the score in FIG. 14, the result is as shown in the lower part of FIG. The numbers in the circles are written for convenience, and it is not known at this point which measure and which measure have the same performance information.

FIG. 15 is a diagram showing a method of detecting an overlapping portion from this data string when there is performance information as shown in the lower part of FIG. The performance information is arranged in units of measures on the time axis as shown in the upper part of FIG. Here, data comparison is performed for measures from the first measure x onward. If they match, the part numbers (in this case, 1) are assigned in ascending order.
If the performance information is not created by copying the part, fluctuations in the time axis direction and the like are included, so that data filing, quantization, and cognitive techniques are used.
A file containing the information with the part numbers added in this way is designated as A, and a file containing the information with the duplicated parts deleted is designated as B. FIG. 16 shows these two files A and B and a compressed file finally created by the second embodiment of the performance information compressing method of the present invention. Reading both files A and B, if a different part appears, increment the level n, insert a jump mark J (n) in file B, return to the part that appears next to file A, and return to tag T (n) Insert

The level n is for mutually discriminating each of the jump mark J and the tag T, and is incremented by 1, 2, 3,... As a natural number. A method of inserting a jump mark J and a tag T is shown below FIG. That is, the file A and the file B are compared with each other in part units from the first part, and if different part numbers appear,
If there is no part of file B to be compared with the part of file B, the level n for identification in each of the jump mark and the tag to be inserted is incremented, and the file B is incremented.
A level-incremented jump mark immediately before the different part number in the file A
The level-incremented tag is inserted immediately before the part number in the same file B as the different part number in the file B. The part numbers after the different part numbers in the file A and the part numbers immediately after the tags of the file B are inserted are compared in part units, and when the mutually different part numbers appear, the same operation as the above-described insertion operation is performed. Insertion takes place.

FIGS. 17 and 18 are diagrams showing examples in which the performance information of a song having a complicated repetition can be compressed using the second embodiment. FIG. 17 shows an example of a song having such a complicated configuration. FIG. 18 is a diagram showing files A and B before compression and a compressed file.

Here, a description will be given of a method of modifying the SMF as the MIDI signal information file and inserting the jump mark and the tag by modifying the SMF. The SMF includes: -MIDI event: performance information-META event: non-performance information Since both events have time information @t, M
An ETA (meta) event can be inserted by specifying a relative time position with respect to the MIDI event. The META event can define the data format of the event. In a normal SMF, there is no concept of a part, but an identifier (P
By defining a meta-event with an ART ID, the part number can be described in the SMF.

Next, the definition of a part will be described. For example, each part corresponding to one bar is designated by a part number as shown in FIG. Thus each part starts with a META event, a series of MIDs in the SMF
This is an I event group. This META event, MET
Abbreviated as A (PART, x). META (PART, x) → @t: META ID: P
ART ID: PARTNO x. For comparison, the MIDI event Note on (keystroke) is N
note on → △ t: MIDI Note on I
D: MIDI channel: pitch: volume The structure of the SMF is as shown in FIG.

The insertion position of META (PART, x) is
Although it is optional, since it is used for detecting repetition, it is better to insert it in a unit less than the unit where repetition is expected. For example, when creating an SMF based on a score as shown in FIG.
The creator creates MIDI performance information for each of “A, B, C, D, E”, and according to the actual performance procedure, uses a method such as data copy / paste to perform “ABC D ABC”.
The MIDI performance information sequence that became "E" is stored in the SMF,
An SMF as shown in FIG. 21 can be created by creating a rule of how to insert a repeat symbol on a musical score and META (PART, x).

Alternatively, when the creator creates a normal SMF that has no relation to the creation procedure as described above, in order to automatically detect a repeated portion, one bar = 1 part on the side receiving the SMF. META (PA
RT, x) should be inserted. Bar breaks in the SMF are already defined in the SMF format. ME
Information such as TA (time signature)
By reading from the MF, it can be easily known. According to the above example, when META (PART, x) is inserted, the SMF becomes as shown in FIG. In either case, the detection of the repetitive portion is performed without any problem. Also, since META (PART, x) is inserted for detection, there is no problem with respect to the reproduction of the performance even if it does not exist in the file finally described with compression.

Next, a description will be given of a third embodiment of the performance information compression method according to the present invention. The META (PA
RT, x) is inserted as SMF A, and SM
The compressed file obtained by compressing the FA is referred to as SMFC. This compressed file SMFC has the standard MID
Since the format is different from that of the I file, it is not always appropriate to call it SMF.
And keep it.

Next, some definitions will be made. (A)
PART LIST (part list): a column of part numbers. (B) fpC (file pointer): compressed file SM
Indicates an arbitrary position in the contents of the FC.

The jump mark J and the tag T are described as follows using the META event. META (JUMP, Ch, y) META (TAG, Ch, z) These are described as a set in the compressed file SMFC as follows. META (JUMP, Ch, y) → @t: META I
D: JUMP ID: MIDI channel Ch: level y META (TAG, Ch, z) → @ t: META I
D: TAG ID: MIDI channel Ch: level z By adding a level to each of the jump mark and the tag, even a complicated performance procedure can be described in a uniform format and correctly reproduced. .

Here, the MIDI channel will be described. In MIDI, a MIDI channel (Ch) is added to a MIDI event so that information that a plurality of musical instruments are simultaneously played can be transmitted. For example, the event for the piano is Ch = 1, the bass is C
With h = 2, the receiving side separates events by channel and transfers each to an appropriate sound source. This concept is exactly the same in SMF. Although different performances are performed for each instrument, on the SMF, the events in the SMF A are separated for each channel in the repetition detection because they are mixed, and the compression description is performed for each channel independently. To the final SMF C,
It is desirable to record while compositing. Even after synthesis,
Each event has a MIDI channel so that the jump mark and tag of each channel can be identified.
For simplicity, SMF A is described herein as including only a single channel event.

Next, the compression procedure of the third embodiment will be described with reference to FIGS. There are two passes in the flow of the compression procedure. Figure 1 shows the first pass
As shown in the figure, the same part in SMF A is detected and M
The rewriting of ETA (PART, x) is performed. In order to regard two parts as the same, MIDI filtering and MIDI quantization may be used at the stage of performing the comparison. This is particularly effective when the content of the SMF is a record of the performer's actual performance or the like. That is,
Even if you try to play the same part, the actual performance cannot be the same MIDI event group.
I-filtering and MIDI quantization are performed to improve the detection rate of the same part, thereby improving the compression rate. MIDI filtering is to delete MIDI events corresponding to extremely short sounds that are not intended by the player. The MIDI quantization is to improve the detection rate by requantizing the temporal fluctuation of sound generation into a large time slot. Also,
The same applies to the volume. The above is a method used for comparison, and is not used for the content of the file finally generated. The second pass is from SMF A to SMF A, as shown in FIG.
C is generated.

In FIG. 8, "j ++" and "i ++" are steps for incrementing variables "j" and "i". The variable i indicates the original part to be compared, and the variable j indicates the part to be compared. In this flow, when the contents of part i and part j match, it is determined that there is a repetition and META (PA
RT, j) is changed to META (PART, i).

Next, a specific compression method will be described with reference to FIG. In the third embodiment, it is not necessary to use the file corresponding to the file B shown in FIG. 16 described in the second embodiment. That is, SM corresponding to file A in FIG.
A compressed file SMF C is finally obtained from F A without using file B.

In step S21, a file C (SMF C) having no initial contents is created, a pointer of the file C indicating a position in the file C is set at the head of the file C, and a part list having no initial contents is created. At this time, since there is no content in the file C, the position of the pointer indicating the position in the file C can be said to be at the beginning of the file C or at the end of the file C. Then, in the next step S22, the first part number of the file A is set to the variable i, and in the next step S23, it is determined whether or not the variable i is in the part list. If there is no part number indicated by the variable i, it is determined in step S24 whether or not the pointer fpC is at the end of the file C.
If the pC position is other than the end of the file C, a jump mark (n) is inserted at the position where the pointer fpC is present at step S25, and then the pointer fpC is moved to the file C.
And then insert a tag TAG (n) at the position of the pointer fpC, and increment the variable n by one. If the pointer fpC is at the end of the file C, the information of the part indicated by the variable i of the file A is added to the file C in step S26, the variable i is added to the part list, and the pointer is Move to the end.

If there is a part number indicated by the variable i in the part number confirmation step, it is determined in a step S29 whether or not the pointer fpC is at the head of the part of the file C indicated by the variable i. In the next step S30, if the position of the pointer fpC is other than the head of the part indicated by the variable i of the file C, a jump mark JUMP (n) is inserted at the position where the pointer fpC is present, and then the pointer fpC is stored in the file C. The part is moved to the head position of the part indicated by the variable i of C, and then the tag TAG (n) is inserted at the position of the pointer fpC, and the variable n is incremented by one. If the pointer fpC is the head of the part indicated by the variable i of the file C, the pointer fpC is moved to the head position of the next part in step S31. This 2
After one of the two moving steps S28 or S31, the next part in the file A is searched in a step S32, and if it exists, the next part number of the file A is set to a variable i. Thereafter, steps from the part number confirmation step S23 to the next part number setting step S32 are repeatedly executed until the next part in the file A disappears.

FIG. 10 shows SM before the start of pass 1 shown in FIG.
FIG. 10 is a diagram showing F A, SMF A obtained after the end thereof, and SMF C obtained after the end of pass 2 shown in FIG. 9. In the figure, A, B, C... Indicate the contents of each part, and numerals 1, 2, 3,... Indicate the part numbers. SMFC shows a state where a jump mark and a tag are added.

Next, a description will be given of a method of decompressing the performance information compressed by each of the above-described embodiments so as to be reproduced on, for example, a communication karaoke terminal. FIG. 11 is a block diagram showing an embodiment of a performance information reproducing apparatus (expansion apparatus) used to realize the performance information decompression method. The storage device 10 is, for example, a large-capacity hard disk, in which a large number of compressed files SMFC shown in FIG. 10 are stored. Among them, the SMFC of the desired music is read out according to an address instruction of a control device (not shown) and loaded into the memory 12. Thereafter, prior to playback, the SMF
A search for the tag in C is performed. That is, the reader 14 operates the memory 1 according to the instruction of the address counter 16.
Read the contents of No. 2 and search for tags. Separation unit 1
Reference numeral 8 is for separating and extracting a MIDI event, a jump level value, and a tag level value from the output signal of the reader 14, that is, the SMFC. Every time a tag is detected as a result of the search, the tag address memory 26 stores the tag level value given from the separating unit 18 and the address given from the address counter 16 in a form as shown in FIG. Go. Thus, one SMF C
To complete the table shown in FIG.

Thereafter, to start reproduction, first, the address counter 16 and the level counter 22 are initialized.
At the time of reproduction, the tag level value from the separator 18 is not written into the tag address memory 26 (or the tag level value is not output from the separator 18). M output from the separator 18
The IDI event is given to each corresponding sound source (not shown) (for each channel). A jump level value is given to the level comparator 20 as one of the output signals of the separator 18. The level counter 22 outputs the count value to the level comparator 20. The level comparator 20 compares the two input levels, and outputs a signal for incrementing the level counter 22 by one when they match, and also outputs a gate signal to the gate 24. give. The gate 24 supplies the output count value of the level counter 22, that is, the level value (coincidence level value) of the detected jump mark to the selector 28 when receiving the gate signal. The selector 28 reads the address of the tag corresponding to the jump level from the tag address memory 26 and loads the tag address into the address counter 16. The address counter 16 sequentially reads data from the memory 12 according to the loaded address.

Thus, one S in the memory 12
The decompression operation is performed by sequentially outputting MIDI events while reading the contents of the MFC C, and the reproduction of one music is completed by reading out all the data of the SMF C.

In the above description of the configuration shown in FIG. 11, a method is used in which the position of the tag is searched prior to reproduction and the address is loaded in advance. However, if the processing capability of the decompression device is sufficient. For example, without reading tags in advance, reproduction can be performed simultaneously while searching for tags. In such a configuration, the tag address memory 26 shown in FIG. 11 becomes unnecessary.

[0045]

As described in detail above, according to the performance concession compression method of the present invention, the data amount of performance information for designating the pitch and sounding time of a tone constituting a musical tone, such as a MIDI signal, is reduced by the melody. By eliminating the repetition, the cost can be significantly reduced, thereby reducing the communication cost and eliminating the problem of the memory capacity. Further, in a communication karaoke system or the like, when image information and character information are transmitted together with original performance information or stored in a memory, the amount of performance information data can be reduced, so that the handling of such additional information becomes easy. .

[Brief description of the drawings]

FIG. 1 is a flowchart illustrating a performance information compression method according to a first embodiment of the present invention.

FIG. 2 is a block diagram of a communication karaoke system as an example of a performance information compression method and a performance information expansion method applied to the present invention.

FIG. 3 is a block diagram showing another configuration example of the communication karaoke system as an example of the performance information compression method and the performance information expansion method applied to the present invention.

FIG. 4 is a block diagram showing the concept of a performance information compression method according to the present invention.

FIG. 5 is a diagram illustrating a technique of a performance information compression method according to the present invention.

FIG. 6 is a signal format diagram used in the first embodiment of the performance information compression method of the present invention.

FIG. 7 is a partial score of a music piece to be compressed by the performance information compression method according to the first embodiment of the present invention.

FIG. 8 is a flowchart showing a part of a third embodiment of the performance information compression method of the present invention.

FIG. 9 is a flowchart showing another part of the third embodiment of the performance information compression method of the present invention.

FIG. 10 is a diagram illustrating a compressed file creation process in a third embodiment of the performance information compression method of the present invention.

FIG. 11 is a block diagram showing an example of a reproducer for realizing a performance information decompression method applied to the performance information compression method of the present invention.

FIG. 12 is a diagram showing the contents of a tag address memory in FIG. 11;

FIG. 13 is a diagram showing a data configuration of a jump mark and a tag in a performance information compression method according to a second embodiment of the present invention.

FIG. 14 is a diagram illustrating a compression procedure of a performance information compression method according to a second embodiment of the present invention.

FIG. 15 is a diagram illustrating a compression procedure of a performance information compression method according to a second embodiment of the present invention.

FIG. 16 is a diagram illustrating a compression procedure of a performance information compression method according to a second embodiment of the present invention.

FIG. 17 is a diagram illustrating a compression procedure of a performance information compression method according to a second embodiment of the present invention.

FIG. 18 is a diagram illustrating a compression procedure of a performance information compression method according to a second embodiment of the present invention.

FIG. 19 is a diagram illustrating a compression procedure of a performance information compression method according to a second embodiment of the present invention.

FIG. 20 is a diagram illustrating a compression procedure of a performance information compression method according to a second embodiment of the present invention.

FIG. 21 is a diagram illustrating a compression procedure of a performance information compression method according to a second embodiment of the present invention.

FIG. 22 is a diagram illustrating a compression procedure of a performance information compression method according to a second embodiment of the present invention.

FIG. 23 shows an SMF before compression according to the performance information compression method of the present invention.
It is a figure which shows typically.

FIG. 24 shows the SMF of FIG. 23 of the performance information compression method of the present invention.
FIG. 3 is a diagram for explaining one technique for compressing.

FIG. S. FIG. 9 is a diagram illustrating an example of a musical score including (Darsegno).

[Explanation of symbols]

 1 SMF (Standard MIDI File) 2 MIDI Channel Separation Block 3 Block Repeat Detection Block 4 R (Repeat) Encoding Processing Block 5 Merge Processing Block 6 Compressed File 10 SMF C Storage 12 Memory 14 Reader 16 Address Counter 18 Separator 20 Level comparator 22 Level counter 24 Gate 26 Tag address memory 28 Selector

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int. Cl. ⁷ , DB name) G10H 1/00 102 G10H 1/00 G10K 15/04 302 H03M 7/30

Claims (2)

(57) [Claims] 1. Performance data for designating a pitch and a sounding time of a sound constituting a musical tone are sequentially compared in a unit of a part from a head part of a part row divided into a plurality of parts. A first step of creating a file A in which parts of the same data are given the same part numbers, and in the file A, checking the part numbers from the beginning,
When a duplicate number appears, a second step of deleting the duplicate part and the part number assigned to the part to create a new file B; , And when a different part number appears, a first jump mark indicating that the part is to be moved immediately before the different part number in the file B is inserted, and A third step of inserting a first tag that is a destination from the first jump mark immediately before a part number in the file B that is the same as a different part number in the file A; Of the file B and the part of the file B after the insertion of the first tag are successively transmitted in part units. It performs a comparison of the port number, when the part number has appeared to be different from each other,
Inserting a second jump mark indicating that a part is to be moved immediately before a different part number in the file B;
And a fourth step of inserting a second tag to which the second jump mark is moved immediately before a different part number in the file A and a same part number in the file B. Compression method. 2. Performance data for designating a musical interval and a tone generation time of a tone constituting a musical tone are sequentially compared in units of parts from a head part of a part row divided into a plurality of parts, and substantially the same data is obtained. A first step of creating a file A by assigning the same part number to the data parts, a second step of creating a file C having no initial contents, and a file C indicating a position in the file C. A third step of setting a pointer at the end of the file C, a fourth step of creating a part list having no initial contents, and a fifth step of setting the first part number of the file A to a variable i. A part number confirmation step of determining whether or not the variable i is in the part list; and a part number indicated by the variable i in the part number confirmation step exists. When there is no pointer, the pointer position first determination step for determining whether the pointer is at the end of the file C, and the pointer position first determination step determines that the pointer is not at the end of the file C. The jump mark JU
MP (n) is inserted, the pointer is moved to the end of the file C, and then the tag TAG to which the jump mark JUMP (n) is moved is located at the position of the pointer.
A sixth step of inserting (n) and incrementing the variable n by one; and determining that the pointer is at the end of the file C in the pointer position first determination step. A seventh step of adding performance information of a part indicated by the variable i of the file A, adding the variable i to the part list, and moving the pointer to the end of the file C; When it is determined in step that there is a part number indicated by the variable i, a pointer position second determination step of determining whether the pointer is at the beginning of the part of the file C indicated by the variable i; When it is determined in the pointer position second determination step that the pointer is located at a position other than the beginning of the part indicated by the variable i of the file C A jump mark JUMP (n) is inserted at the position where the pointer is currently located, the pointer is moved to the head position of the part indicated by the variable i of the file C, and the tag TAG ( n) is inserted, and the variable n is incremented by one. In the pointer position second determination step, it is determined that the pointer is at the head of the part indicated by the variable i of the file C. In some cases, after the ninth step of moving the pointer to the head position of the next part, and after the seventh step and the ninth step, the next part in the file A is searched for and exists. Setting a next part number of the file A to the variable i. The next part number of the file A Until no, performance information compression method which is characterized in that so as to repeatedly execute until the next part number setting step from the part number verification step.