JPH0738107B2 - MIDI signal recording / reproducing system and device - Google Patents

Description

The present invention relates to a digital signal processing device, and more particularly to a MIDI signal recording / reproducing system and device for controlling an electronic musical instrument.

[Prior Art] There is a MIDI standard for driving and controlling various electronic musical instruments with digital signals according to a predetermined format. This MI
Keyboard devices, electronic musical instruments, etc. have already been put into practical use according to the DI standard, and a large number of electronic musical instruments can be driven by operating a single keyboard device. In addition, the applicant has previously recorded 8-bit data based on the MIDI standard (hereinafter referred to as MIDI word) on a magnetic tape, and by reproducing this, one or more electronic musical instruments can be driven without operating a keyboard. We have developed a digital information recording and recording / reproducing system and applied for a patent. This is JP 62-1464.
As shown in Japanese Patent Publication No. 70, digital information relating to the performance of a musical instrument, that is, musical instrument control information such as pitch, scale and duration, is recorded in advance as a MIDI word in a helical scan type magnetic recording / reproducing apparatus. , Play this, add start bit and stop bit, and add 10 bit MI
The DI signal is given to each musical instrument to drive it.

[Problems to be Solved by the Invention] If the above magnetic recording / reproducing apparatus is used, an 8-bit MIDI word obtained by removing a start bit and a stop bit from a MIDI signal composed of a total of 10 bits including a start bit and a stop bit is used as it is. It can be recorded and played on magnetic tape, but it was not possible to record MIDI words in the subcode of a compact disc and play them on a compact disc player. This is because one MIDI word is composed of 8 bits, but one part of the subcode of a compact disc (hereinafter referred to as CD) is R, S except for 2 bits used for time code. … W
Because it consists of 6 bits consisting of, it is not possible to record the MIDI word as it is on the subcode channel of the CD. The transmission rate is 3,600 BPS (byte per second), that is, 28,800 bps, which is 31,250 bps.
The ps MIDI signal cannot be recorded as it is.

In order to solve such a problem, the applicant of the present invention, prior to the present invention,
Record the MIDI signal in the subcode channel of the CD,
We have developed a MIDI signal demodulation device that can reproduce the original MIDI signal on this CD player and have applied for a patent (Japanese Patent Application No. 63-294407).

In this way, a MIDI signal demodulation device for driving an electronic musical instrument by decoding a MIDI signal recorded in advance on a digital signal recording medium such as a CD or DAT may be played in real time at the time of playback depending on the arrangement of data when recording the MIDI signal. In some cases, playback may not be possible, or data that is erroneous due to MIDI format may be transferred.

Therefore, it is an object of the present invention to provide a MIDI signal recording / reproducing system and device capable of outputting MIDI output timing in accurate real time during MIDI signal demodulation.

[Means for Solving the Problems] The present invention has been made to achieve the above object, and when recording a MIDI signal in a digital signal recording medium, a NOP (command for instructing non-operation to an empty byte in MIDI data). No Operatio
n) Insert a command and write the number of MIDI bytes in one pack to the format's instruction section.
The empty byte part in which the NOP command is written during playback provides a MIDI signal recording / playback method that does not output.

Further, according to the present invention, demodulation means for demodulating a signal from a digital signal recording medium, correction means for performing error correction of subcode data based on the output of the demodulation means, and the digital signal in response to the output of the correction means. MI that decodes the identification signal of a predetermined frame in the pack of the subcode of the signal recording medium to identify whether the recorded data in the pack is MIDI data and detects the number of bytes of MIDI data in the pack.
DI data demodulating means and means for reproducing a MIDI signal based on data within a predetermined range of subcode bits, and outputting so as not to output an empty byte portion in which a NOP command for instructing non-operation at the time of reproduction is written The present invention provides a MIDI signal recording / reproducing device having

[Operation] Since the MIDI signal recording / reproducing system and device of the present invention adopt the above-mentioned system and have a configuration, the number of MIDI bytes and NOP of the instruction part of the format during demodulation
By outputting 1 byte of data when a command is detected, MIDI output timing is output in real time.

[Embodiment] First, before describing a specific embodiment of the present invention, a recording format of a MIDI signal in a subcode channel of a CD will be described.

FIG. 6 is a diagram showing a storage state of data for one pack of the subcode channel of the CD, that is, data from frame 0 to frame 23. 6 from R to W for each frame
Bit data is stored, and MIDI data is stored from frame 4 to frame 19. Each frame has a 6-bit structure, and the stored MIDI data is a byte, that is, an 8-bit unit. 1-byte MIDI data is, for example, R to W of frame 4.
Is _composed of 6 bits represented by a ₀ and 2 bits represented by a ₀ ′ in R and S of the frame 5, that is, 8 bits in total. In embodiments hereinafter Similarly 1-byte data by a combination of four bits b ₀ 'remaining in the R~U of 4 bits of b ₀ and the frame 6 T～W frame 5 is constructed, the frame from the frame 4 12 bytes of MIDI data are stored over 16 frames up to 19.

FIG. 7 is a signal waveform diagram of 1 byte of MIDI data to be finally sent out. A start bit of logical 0 is added to the head of 8-bit MIDI data and a stop bit of logical 1 is added to the end.

FIG. 8 shows the structure of MIDI data in one pack to be sent out. A ₀ , a ₀ ′, b _0, etc. in FIG. 8 indicate the correspondence with the data in the subcode in FIG. To briefly explain the MIDI standard, MIDI data is 1
Sent at a rate of 3125 bytes per second. The subcode in Fig. 6 shows one pack consisting of 24 frames.
Since 300 packs are played from a CD per second, MIDI data is played at 6 bits x 16 frames x 300 packs = 28800bps. If this is expressed in bytes, 28800bps / 8 = 36
00 becomes 3125BPS of MIDI data (byte second)
Will exceed. Therefore, it is not possible to store 12 bytes of MIDI data per pack for all packs.
Not allowed by the standard. So, as a countermeasure, 3 with 300 packs
It is necessary to devise it beforehand when recording on the subcode of the CD so that it becomes 125 bytes. For example, 300 packs are divided into 12 packs of 25, and each 12 pack group
5 packs containing up to 11 bytes of MIDI data,
If you have 7 packs containing up to 10 bytes of MIDI data, the total number of MIDI bytes in this group will be up to 125 bytes, and the total of 300 packs will be 3125 or less.

However, most of the MIDI data is the note-on command and the note-off command, and each of these consists of 3 bytes, so it is better to consider 12 bytes which is a multiple of 3 as a reference. Therefore, it is preferable that 12 bytes are put in some packs and 11 bytes or less are put in the remaining packs in each of the above groups so that the total is 125 bytes or less.

By the way, in FIG. 6 above, n3 to n0 of the frame 1 indicating the instruction indicate the number (byte number) of MIDI byte data in the pack (this value is represented by a 4-byte binary in the range of 0 to 12), for example in the case of recording a 6-byte MIDI data, drawing _{a 0, a 0 '~c 1} , c 1' containing the data to, and n3~n3 = "0110". Also at this time a ₂ ,
All of a ₂ ′ to c ₃ and c ₃ ′ are 0. That is, the number of bytes is recorded sequentially from the head of each pack.

In this format, the time interval from a ₀ , a ₀ ′ to c ₃ , c ₃ ′ is 320 μsec for sending 1 byte of MIDI data.
Since it is necessary, 320 μsec × (12-1) = 3.52 [msec] (1).

Therefore, for example, if you input MIDI clock F8 _H , it is a song with a fast tempo, (Symbol that represents the number of beats in which one quarter minute is one beat) = 240
In some cases, the MIDI clock rate at this time is 240 (beats / minute) ÷ 60 (seconds / minute) x 24 (clocks / beats) = 9
6 (clocks / second) ... (2), and the time interval of 1 clock is 1/96 = 10.417 (msec / clock) ... (3).

Comparing the values of (1) and (3), (3) is about 3 of (1).
If you want to insert the MIDI clock into the CD subcode at the correct timing, it is impossible to record 6 bytes of MIDI data in each pack of the CD subcode channel in order from the beginning. The details will be described below.

For example, Figure 5 shows MIDI data in packs 1 to 3, respectively.
Stores 11 bytes, 10 bytes, 10 bytes, Is an explanatory diagram of the problems that occur when inserting the MIDI clock of the CD subcode (in the figure, the shaded area is the MIDI byte containing the MIDI clock, "M" is the MIDI data, "-" is the empty byte, B1 ~ B11 indicates the MIDI byte column).

In the same figure (b), if the first MIDI clock is put at the beginning of the pack 1, the time when the next clock should be recorded is 200 ÷ 60 × 24 = 80 as in equations (2) and (3). (Clock / sec), 1 ÷ 80 = 12.5 (msec /
Clock) …… (4) after.

Also, since 320 μsec is required to transfer 1 byte of MIDI data, equation (4) is 12.5 (msec) ÷ 320 (μsec / byte) ≒ 39 (byte) ...
… (5), the time interval for 39 bytes is 320 μsec × 39 = 12.48 (msec)… (6) Therefore, the time difference between equations (4) and (6) is 0.02.
In msec, the audible range that humans can perceive is on the order of msec, so it does not matter, but if there is no MIDI data other than the MIDI clock in the pack 4, then according to the recording method of MIDI data described above. Since the MIDI byte fields B1 to B8 cannot be emptied, they will be recorded in B1.

However, the MIDI clock has an error of 9-1 = 8 MIDI bytes, that is, 320 μsec × 8 = 2.56 msec (7). On the other hand, MIDI byte field
If it can be put into B9, it will be held within 1 byte at the MIDI byte rate, that is, ± 0.5 bytes, and on the time axis it will be ± 320 μsec × 0.5 = ± 0.16 msec (8), encoding in byte units It becomes an error of time.

The figure (b) is an example in which the MIDI bytes are packed in each pack as described above, but MIDI is basically an asynchronous communication format, and normally there is a gap as shown in the figure (c). In this format, at this time, MIDI data is output from the beginning timing of each pack (when the MIDI output is 3.125 KHz timing and interrupt processing is performed, even if there is a gap, Become). Although the MIDI data is output from the beginning of each pack, the actual pack playback timing and the MIDI data output start timing may not always match and may have an offset due to the error correction end timing or the like.

Also in the case of FIG. 7C, as in the previous example, Then, the time to insert the second MIDI clock is (4)
It is 12.5 msec after the formula. Since the start position of pack 4 is 3.265 msec, 3.265 x 3 = 9.795 (msec) (9), the position to insert the second clock is 12.5 (msec) from the start position of pack 4- 9.795 (msec) = 2.705 (msec) …… (1
0) Since it is later, if you divide this by the MIDI byte rate, 2.705 (msec) ÷ 320 (μsec) ≈ 8 will leave 8 bytes worth of space. The time interval at this time is 9.795msec + 320μsec x 8 = 12.355msec (11) In other words, the MIDI clock should be recorded in the byte column B9 of the pack 4 as shown by the shaded area in Fig. 11 (c). ~
Since it is not possible to empty the empty byte of B8, byte field B1
Must be put in. Moreover, even if it is put at the beginning of the next pack 5, there is an error of 3.265 msec × 4-12.355 = 0.7 msec (12).

As described above, when recording various MIDI data, not limited to MIDI clock, to the CD subcode, it is necessary to insert them sequentially from the top of each pack. become. In other words, ± 5 MIDI bytes is 320 μsec × 5 = 1.6 msec (13), which is about 10 times as large as the encoding error in byte unit of equation (8).

Therefore, if the MIDI data is recorded in the pack at intervals, the MIDI data can be input at a more accurate timing. However, in this case, the definition of the number of bytes of MIDI shown from n3 to n0 of frame 1 in FIG. 6 becomes a problem.

That is, even if the value is 0, the MIDI data becomes meaningful data by adding a start bit and a stop bit to the 8-bit 0 data. In the MIDI format, except for system exclusive messages, the number of data that normally follows the status byte is fixed, but if 0 data is sent after sending a certain number of data bytes after the status byte, the receiving side will I will recognize this. Because there is a transmission method called running status in MIDI only for channel messages, if you want to send data of the same status continuously, send the status once at the beginning and send only the data byte without sending it after the second time. Because it is good. The system exclusive message has an optional data length following the status byte.

From the above, careless transmission of 0 data may result in incorrect data transfer due to the MIDI format.

Therefore, according to the present invention, when a MIDI signal is recorded on a sub-code channel of a CD, a 1-byte command that is a NOP (NO Operation) is used to enter an empty byte in MIDI data, and
NO in 1 pack in the format instruction section
By writing the number of MIDI bytes including P, the MIDI output timing can be output in accurate real time.

Next, specific examples will be described.

FIG. 1 is a MIDI data input timing diagram for explaining the MIDI signal recording / reproducing system of the present invention. The figure (a) is aa
An example of MIDI data (6 bytes long) of a 3-byte completed message called a and bbb is shown, and FIG. 7B shows an example of MIDI data when this message example is packed in order from the beginning of a pack of 6 bytes. . At this time, the number of bytes, which is the number of MIDI data, is 6. However, since the decoder reads the byte value and reads that number from the beginning of the pack, in the case of (a), the data sent out is only 6 bytes of 0, a, a, a, 0,0 in order. I will end up. Further, at this time, if data is reproduced by skipping 0, for example, even if there is 0 data in a or b of 3 bytes, that byte will not be sent. On the other hand, in the case of (b), if the number of bytes is not specified, 0 data of bytes 7 to 12 will be transmitted, and erroneous data transfer will be performed.

Therefore, as shown in (c) of the same figure, a 1-byte command N of MIDI that is a NOP is inserted in the empty byte part of the original MIDI data shown in (a), and this command N is added to the instruction part of the format. By writing the number 10, the above-mentioned drawback can be solved.

However, there is no command that has the meaning of NOP in the current MIDI standard, and statuses that are not yet used in MIDI should not be used in the MIDI standard, and even if some unused status is used later The system reset command FF _H is currently used because the status may be defined to have another meaning. The reason for using this system reset command FF _H is as follows.

-The master reset is a MIDI standard, and it is desirable to use it conservatively and only as a command for panel operation. Therefore, a CD automatically sent from the disc.
-It cannot exist in MIDI data.

-Since it has already been defined as the system reset command FF _H , this value will not be assigned to other uses.

・ The blanks after the last data of each pack are all 0 data, so it is easy to distinguish (FF _H is all 1 data).

・ The standard says, "It is rather dangerous to use a lot of reset messages. It is considered safe to not recognize it in the case of ordinary keyboard instruments." Putting it on a CD as a reset is dangerous because it is not compatible with the connection with each instrument.

-The active sensing command FE _H may be used as a NOP command, but once inserted, it must be encoded so that the data always exists within 300 msec.

To avoid the above restrictions, the system reset command FF _H is used as the NOP command.

FIG. 2 is a block diagram showing an embodiment of the MIDI signal recording / reproducing apparatus of the present invention. The MIDI signal recording / reproducing apparatus is an optical pickup 12 for reading CD data from a CD 11, and amplifies the read data via a photodetector preamplifier 13 for input, and EFM demodulates the audio data and subcode data as main data, respectively. Data error correction / interpolation circuit 15 and subcode data error correction circuit 17
Via the PLL EFM demodulation circuit 14, the audio data error correction / interpolation circuit 15 that outputs the digital / analog converted audio data to the D / A conversion circuit 16, and the subcode data error correction circuit 17 MIDI data demodulation circuit 18, which demodulates MIDI data from subcode data,
And a MIDI signal modulation circuit 19 which modulates a demodulated and output MIDI signal and sends out a MIDI output.

Here, the subcode data error correction circuit 17 may be the same as that used in the CD graphics decoder,
Error detection and correction is performed using the parity bit group P _{0 to} P _{3 of} frames 20 to 23 and Q ₀ and Q ₁ of frames 2 and 3.This error detection is performed by the CPU, ROM, RAM and other software. Although it can be realized, it can also be realized by hardware by a discrete circuit. Since it is also used in common with the CD graphics decoder, it can be used as a CD graphics and MIDI signal recording / reproducing device. If it is also used as the CD graphics decoder, enter the frame 1 and the frames 4 to 19 in FIG. Data is graphics data or MI
It is necessary to identify whether it is DI data. Frame 0 in FIG. 6
The 3 bits indicating the mode of are used for this purpose, and are set to 001 for graphics and 011 for MIDI. Similarly, the latter 3 bits of frame 0 are an item and indicate the type of graphics or the like. The number of items is 000 when MIDI data is included.

Also, the MIDI data demodulation circuit 18 is for converting one pack of subcode shown in FIG. 6 into 12-byte MIDI data shown in FIG. 8, in which case the number of bytes displayed in bits n3 to n0 of frame 1. To detect.

Although the subcode data error correction circuit 17 and the MIDI data demodulation circuit 18 can be realized by a general-purpose microcomputer, FIG. 3 shows a processing flow of the MIDI data demodulation circuit 18. In Fig. 3, MODE and ITEM are the mode and item of frame 0, BYTES is the number of bytes in frame 1, FA is the frame address of the subcode, MA is the memory address for storing MIDI data, and MB is the byte number of MIDI data. , And n is the quotient of (FA-4) divided by 4. The first judgment step MODE = 3, ITEM = 0? Is the detection of the MIDI mode, the next judgment step BYTES = 0? Is the detection of the existence of one or more MIDI bytes, and the next judgment step (FA -4) MOD4? Is a step of obtaining a remainder obtained by dividing (FA-4) by 4, and the following flow is determined depending on whether the remainder is 0 to 3. By this flow, one pack, that is, frame 4 to frame 19 in FIG. 6 is exchanged, 12-byte MIDI data in FIG. 8 is created, and then the same process is performed for the next pack.

The 12-byte MIDI data created in this way is sent to the built-in buffer memory in the MIDI signal modulation circuit 19 shown in FIG. 2, P / S converted and bits are added in synchronization with the MIDI sending clock, and MIDI is sent. The byte is sent out to the outside as a 10-bit serial MIDI signal with a start bit and a stop bit added as shown in FIG. In this case as well, it is transmitted in synchronization with the MIDI output clock. FIG. 4 shows an interrupt process that occurs at regular intervals, and the data in the second judgment step = FF? Is the NOP in the pack when there is data in the buffer memory in the MIDI signal modulation circuit 19.
NO in the step to determine if the command is used
When a P command is detected, that part is not output and MIDI data is reproduced and output.

In the above-mentioned embodiment, the MIDI signal is recorded in the CD and reproduced. However, the present invention is not limited to the CD, and may be recorded in a digital signal recording medium such as DAT in advance and reproduced. Applicable. FIG. 9 shows an embodiment in which a MIDI signal is recorded in the DAT. FIG. 9 (a) is generally known.
The frame format of DAT is shown in (b) for one subcode area of one track therein, (c) shows the format in one block in (b), and BA Even is in even addresses. , BA Odd indicates the format within an odd address, and these are all known. Further, Format (I) and (II) of (d) in FIG. 9 are DAT.
It is a figure which shows a mode that the MIDI signal was recorded in the pack format of, and the ITEM and PARITY section in the figure are known.

The value of ITEM is undefined in the current DAT format (published by the DAT Conference in July 1987).
One of 1000 to 1110, for example 1000, is used as the MIDI mode. (I) SUB ITEM identifies the contents of B2 to B7 in more detail in MIDI mode. For example, MIDI
Set to 0000 when inserting data. ADRS is 00000-1
There are 19 addresses of 0010 and represent the position of MIDI data in one frame. Here, 5 bytes can be stored in 1 pack,
One frame contains 5 bytes x 19 (address) = 95 bytes of MIDI data, and 1 frame is 30 msec as is well known. Therefore, if you insert it in this way, the number of MIDI data that can be inserted per second is Therefore, the maximum transfer capacity of MIDI will be exceeded. Therefore,
As with CDs, it is necessary to limit the amount of recorded MIDI data.

FIG. 9 (e) shows the case where 375 bytes of MIDI data are inserted in 4 frames when this limitation is added. As mentioned above, DAT is 1sec / 30msec frame per second, And conforms to the MIDI standard.

The actual data at this time is the first 3 frames, that is, the frame address = 4n to 4n, with the 4 frames as one unit.
MIDI data of 94 bytes when +2 and 93 bytes when the last one frame, that is, frame address = 4n + 3, is inserted so that the total becomes 94 × 3 + 93 = 375 bytes.

In addition, for addresses 0 to 18 in each frame, all of addresses 0 to 17 in the unit of 4 frames are 5
Bytes of MIDI data shall be inserted, and for address 18, 4 bytes will be inserted when frame address = 4n to 4n + 2, and 3 bytes of data will be inserted when frame address = 4n + 3.

In addition, Format (II) in Figure 9 (d) uses 6 bytes per pack to reduce the number of packs required for MIDI.
It contains MIDI data. The meaning of ADRS is Format (I).
Is the same as. At this time, 0 to 15 (0000 to 11
The data of 11) is given, and like the Format (I), MIDI
Frame address 4n to meet the transfer rate of
94 bytes for ~ 4n + 2, 95 for frame address 4n + 3
Byte MIDI data is put in, and the final address in each frame, that is, address 15, is the frame address = 4n to 4n
Data of 4 bytes is put in +2, and data of 3 bytes is put in when frame address = 4n + 3.

As in the above example, it is possible to put MIDI data in the DAT, but in this case, the NOP command described above can also be applied. In the case of DAT, as shown in FIG. 9 (a), since the data is recorded in a compressed form of 1/2 on the time axis, it depends on the address as shown in FIG. 9 (d). Continuous position information within one frame is required. At this time, if the discrete MIDI data on the time axis is demodulated into a continuous format on the time axis in this way, there is a method of not outputting the MIDI data during the time when nothing is sent. May be needed. At this time, while demodulating the MIDI data in the DAT subcode pack using the NOP command described above,
It can be configured not to send a MIDI signal when the NOP command is detected.

[Effects of the Invention] As is clear from the above description,
When recording a MIDI signal in a subcode channel of a digital signal recording medium such as a CD or DAT in the MIDI signal recording / reproducing system and device, insert it into the empty byte in the MIDI data using the NOP command, and in the format instruction section. The number of MIDI bytes in one pack is written, and the empty byte part where the NOP command is written is not output based on these detections at the time of playback, so the MIDI clock is set to the sub timing of CD or DAT at the correct timing. It can be inserted in the code, and thus the MIDI output timing can be output accurately in real time.

[Brief description of drawings]

FIG. 1 is an explanatory diagram of a MIDI signal recording / reproducing system of the present invention, and FIG.
FIG. 4 is a block diagram showing an embodiment of a MIDI signal recording / reproducing apparatus of the present invention, FIG. 3 is a flow chart showing a process of a microcomputer constituting the MIDI data demodulating circuit 18 of FIG. 2, and FIG. A flow chart showing the processing of the microcomputer which constitutes the MIDI signal modulation circuit 19 section of the figure,
FIG. 5 is an explanatory view for explaining a problem in outputting MIDI data which is a premise of the present invention, FIG. 6 is a view showing one pack of R to W channels of subcode of CD, and FIG. Should MI
Waveform diagram of 1 byte of DI signal, FIG. 8 is a diagram in which the MIDI data in the 1-pack subcode of FIG. 6 is a 12-byte MIDI signal, and the waveform diagram of FIG. 3 is followed. 9 is DA
It is a figure which shows the Example at the time of recording MIDI data in T. 14 ... PLL EFM demodulation circuit, 17 ... subcode data error correction circuit, 18 ... MIDI data demodulation circuit, 19 ... MIDI signal modulation circuit.

Claims (2)

[Claims] 1. When recording a MIDI signal on a digital signal recording medium, a NOP which commands a non-operation to an empty byte in MIDI data.
(No Operation) command is inserted, the number of MIDI bytes in one pack is written in the format instruction part, and the empty byte part where the NOP command is written during playback is not output. Play method. 2. A demodulation means for demodulating a signal from a digital signal recording medium, a correction means for correcting an error of subcode data based on the output of the demodulation means, and the digital signal in response to the output of the correction means. Decoding the identification signal of a predetermined frame in the subcode pack of the recording medium,
MI that identifies whether the recorded data in the pack is MIDI data and detects the number of bytes of the MIDI data in the pack
DI data demodulating means and means for reproducing a MIDI signal based on data within a predetermined range of subcode bits, and outputting so as not to output an empty byte portion in which a NOP command for instructing non-operation at the time of reproduction is written A MIDI signal recording / reproducing device having: