JPS6232793B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention relates to an automatic accompaniment generation method that encodes the musical score of an accompaniment section consisting of single notes, chords, and rests of a digital electronic musical instrument, stores it in a performance information memory, and faithfully reproduces the musical score during playback. . Conventionally, digital sequencers have been used as accompaniment devices for digital electronic musical instruments.
Known digital sequencers use a read/write memory to store a number of words, each word being encoded and representing an accompaniment note that corresponds to a note played on a keyboard, and placing the data words in memory in a temporal sequence. The accompaniment sound was obtained by changing the reading speed using a rhythm generator. However, these accompaniment tones are composed of single tones, and the memory is not used effectively. Therefore, the present applicant proposed Japanese Patent Application No. 18845-1988 entitled "Automatic Sequence Generator for Polyphonic Synthesizer." It can play not only single notes but also chords, and is configured to increase memory usage efficiency. That is, this sequence prepares a sequencer memory for storing data that defines the notes to be played. As shown in Figure 1c of the specification of the above application, the contents of this sequencer memory are 2 bits for division (keyboard type), 3 bits for group (octave), 4 bits for pitch name, 8 bits for sound duration on, and 8 bits for off. Stored. The contents include word data for turning on or off accompaniment sounds corresponding to key operations. In this manner, the sequencer memory encodes and indicates only the on/off states of pronunciations and rests, and there is no need to indicate each note of the melody, thereby reducing memory bits. However, a large number of bits, for example, 23 bits, excluding the division 2 bits as mentioned above, are required to constitute word data in the sequencer memory. In this case, the sequence counter that counts the on and off sound times is
This is 8-bit data in 256-decimal format, and each on and off data is stored in each of the 8 bits at the end of the word data. Sequence counter 256 decimal 1
If the count corresponds to the sound of a quarter note, if the song is in 4/4 time, only 256/4 = 64 bars can be recorded. On the other hand, in order to record even more bars, it is necessary to increase the number of on/off bits for the sequence counter and the sequencer memory, and as the number of bits in the sequencer memory increases, the related components become larger. Furthermore, even if three-note chords, such as accompaniment notes, turn on and off at the same time, the on/off data will be duplicated with the pronunciation data and stored in the rear 8-bit portion of each word in the sequencer memory. Become. Therefore, this recording method is effective when the notes make complex movements, such as in a melody, but becomes a highly redundant recording method when the notes make monotonous movements, such as accompaniment notes. An object of the present invention is to provide an automatic accompaniment generation system that allows accompaniment sounds consisting of polyphonic notes and rests to be automatically played faithfully to the musical score using a small number of memory bits. In order to achieve the above object, the automatic accompaniment generation method of the present invention, when recording performance information, generates recorded key press information from one or more octaves and note name bytes that should be generated at the same time, which are formed from the key press information of the keyboard switch. For each set, the sound time information of one measure/beat byte consisting of the number of measures and the number of beats is set by operating the control switch, and the recorded key press information and sound time information are used to control the performance order indicating various notation symbols. The information is transferred and stored in the performance information memory according to the musical score of the accompaniment notes, and during playback, the key press information is created from the recorded key press information and held until the sound time information is read from the performance information memory. , a total sum of the number of beats found from the product of the bar number code, which is a component of the bar/beat byte of the read pronunciation time information, and the beat number of the rhythm generator, which is another component. Based on the number of beats, the duration of the sound generation is determined and the sound generation of the stored recorded key press information is performed, and when the performance order control information is read out,
The present invention is characterized in that the address of the performance information memory storing the key press information to be generated next is searched for, and then the next key press information is sounded. The present invention will be described in detail below with reference to examples. The recording method of the present invention does not encode the start and stop of pronunciations and rests, but instead expresses how many measures and beats last, and the same chord can be played for one measure or several measures like accompaniment notes. This is an effective recording method when most of the music is played over multiple periods. Also, since the bar/beat byte data indicating how many bars and beats last is independent from the octave and note name byte data in which key press information is stored, no duplication is recorded. Generally, the number of chords used in a song is around 100 at most, depending on the flow of the song. Also, the same chord is often repeated for about 8 to 10 measures.
For example, if the performance information memory of the present invention is
If the bit is composed of 256 bytes and a chord is composed of 3 notes, 4 bytes are required from 3 octaves, a note name byte, and 1 bar/beat byte, so 256/4 = 64 chords can be recorded. . The length of each chord consists of a 4-bit bar number code and a 2-bit beat number code, so it can be set up to a maximum of 16 bars and 3 beats. In the embodiment described later, 12 of the input switches are used for a maximum of 12 measures. Furthermore, since the accompaniment is configured to be performed in order according to the notation symbols, the accompaniment part of the musical score that is actually used can be recorded and reproduced as is. Therefore, the performer does not need to play the accompaniment section and can concentrate on playing the melody section. FIG. 1 is an explanatory diagram showing the configuration of an embodiment of the present invention according to the above-mentioned principle, and the portion surrounded by broken lines refers to the keyboard switch information assignment circuit of Japanese Patent Application No. 54-38865. This part will be briefly explained. In the figure, a keyboard circuit 10 receives a 5-bit signal from an assigner 20, including a keyboard code (2 bits) that separates the upper keyboard, lower keyboard, and foot keyboard, and an octave code (3 bits) that separates the pitch range.
As a result, 12 bits corresponding to key information (C, C#, D, . . . , B) corresponding to an arbitrary octave of an arbitrary keyboard are sent to the assigner 20. The assigner 20 includes an information processing unit (CPU) 21, a CPU clock generator 22 that drives it, and a program 2.
3, an assignment memory 24, and an event memory 25. The CPU 21 uses one of its internal registers to create a keyboard code and an octave code, increments the value of this register, and outputs it to the keyboard circuit 10 through the output port (1). Hereinafter, this register will be referred to as the OD register. The keyboard circuit 10 immediately sends the key information of the specified keyboard and the specified octave to the assigner 20. The assigner 20 inputs this from the input port (1) and compares it with the key information of the corresponding keyboard and octave in the event memory 25 that stores the key information left from the previous scan to check whether there is any change. The change in this case is hereinafter referred to as an event. If the open/closed state of the keyboard switch is different from the previous scan, it means that an event has occurred, and conversely, if there is no change in the open/closed state, it means that no event has occurred. If there is no event, the value of the OD register is incremented again and output to the keyboard circuit 10, and the process moves to the next step. If an event occurs, check if the same key code has already been written to the assignment memory 24, or if it is an event from ON to OFF, or from OFF to OFF.
Check whether the event is ON, etc., and if it is an event with a key code that is not written in the assignment memory 24, write this key code to the channel of the assignment memory 24 that is in the released state and turn it ON/OFF. The bit is an ON symbol,
Also, if the event is for a key code that has already been written, the ON/OFF bit of this key code is inverted. This means that if you select and write an event for a key code that has already been written to a channel in which the key is released, if the decay duration of the musical tone is set to be long, the same key will be pressed and released. This is because if repeated quickly, the same key code will be stored in all channels, and these attenuated tones will overlap, resulting in a musical tone that differs from the actual musical tone. Such consideration is not necessary when the decay duration is short. After scanning the upper key, lower key, and foot key, the necessary contents of the assignment memory 24 are transferred to other blocks, such as the tone envelope generator 31 and tone frequency indicated by dotted lines in the figure. It is transferred to the generator 32, musical waveform generator 33, etc. In this proposed example, when writing a key code to a channel in a key-released state, a priority number is considered as a criterion for determining which channel to write to. This priority number (hereinafter referred to as PN) is stored in the channel of the assignment memory where this key code is stored, as soon as the key is released and the musical tone begins to decay. Write PN. Next, it is necessary to advance the priorities of other channels in preparation for key presses, so the priority is set by subtracting the PN from the channel where the sound has started and which has a larger value than the remaining PN. Let's move on. A channel with a smaller PN value is determined to be a channel with more advanced attenuation, and a new key code is written preferentially. This system does not require a release end signal, and only needs to be controlled by providing a priority number area assigned to each channel in the assignment memory, thereby simplifying the configuration. Furthermore, when a new key is pressed while all channels are simultaneously sounding, the channel with the most advanced attenuation is always assigned the minimum priority number 1, making it easy to search and simple in configuration. The present invention provides this keyboard switch information allocation circuit.
For the CPU 21, input switch groups S0 to S14
Various musical notation symbols etc. on the musical score are inputted through the input port (2), and a signal indicating the type of beat of the rhythm selected by the rhythm generator is inputted through the input port (3) from the aforementioned keyboard circuit 10. The octave and pitch name signals are sent to the CPU 21 in parallel via the input port (1).
Enter. An interrupt is then applied to the CPU 21 via a logic circuit that performs predetermined control at the timing of each beat rhythm pulse from the rhythm generator during playback. The CPU 21 is provided with a program 59 and a performance information memory (PIM) 60 for these processes. To explain in detail below, S0 to S14 are input switch groups, and Table 1 shows one example of these input switch groups for each group. Most of them are switches corresponding to musical notation symbols on the musical score, and are represented by switch codes ss (ss ₁ to ss ₄ ). Input switch S0 inputs the third parameter (described later) that indicates how many measures and beats the same pronunciation or rest continues.
Confirm that the bar/beat byte information in the table has been entered.
Notify CPU21. For bar number information, by pressing any of the switch groups SA, the switch code ss is stored as a bar number code in b ₀ to b ₃ of the bar/beat byte. Therefore, the maximum number of consecutive measures is up to 12 measures. Generally, the number of consecutive measures of the same pronunciation in a song is at most 8 to 10 measures. Beat rate information can be displayed by pressing any switch group SA1.
ss is stored as a beat number code in b ₄ and b ₃ of the bar/beat byte. Therefore, the maximum number of continuous beats is up to 3 beats. Generally, it is sufficient if the types of beats are 2, 3, or 4 beats, and since the rhythm generator has only these types, it is sufficient to store up to 3 beats. Next, the input switches S1 to S12 generally input information on music notation symbols that are often used in songs. These are musical notation symbols that indicate changes in the playing order of the piece, ie, starting, stopping, repeating, skipping, etc., and the details are as described in the music book. However, the notation symbol shown in S3 is “Repeat and
This is a notation symbol used to achieve the "Fadeout" effect. This is a music notation symbol that sets S1 at the front end of range A of a certain song and S3 at the rear end, and gradually expresses This is the effect of ending a song by reducing the volume with the keyboard pedal and making it disappear.By pressing any of these input switches S1 to S12, the switch code ss will be changed to the notation symbol byte in Table 3. The notation symbol byte (UB) is stored as a notation symbol code in b ₀ to b ₃ of =
If it is used, UB is set to “1”. However, in the case of musical notation symbols that do not require such operations, UB="0" always remains. Next, input switch S13 is a switch group.
This is a correction switch for correcting incorrect input using SA. Further, S14 is a switch for specifying recording, off (that is, manual playable state), or playback. These are switched by terminals P1, P2, and P3, respectively. Also, the third
The octave and pitch name bytes in the table contain the octave codes OCT ₁ to OCT ₃ of the key press information transferred from the assignment memory 24 and the pitch name code.

【table】

【table】

【table】

[Table] Codes NOTE ₁ to NOTE ₄ are stored. In this embodiment, the generation of automatic accompaniment is limited to the lower keys only, so there are no bits for the keyboard codes DIV ₂ and DIV ₁ to represent each keyboard. The assignment memory 24 to which the transfer is to be made is also the assignment memory address A ₇ to A ₁₂ in Table 5 of the specification of the proposed example _.
NOTE ₄ , consists of OCT ₁ to OCT ₃ . Next, there is an OR circuit 41 that inputs input signals in parallel through the input switch group S0 to S14, and a monostable multivibrator 4 that operates with its output.
2 and its output signal as one input, S0 to S1
4 is converted into a pulse signal corresponding to each input switch by an AND circuit 43 which takes the input signal from 4 as the other input. This pulse signal is input to the input port (2)
The CPU 21 detects the presses of S0 to S13 via the CPU 21. The rhythm selection controller 44, rhythm pattern pulse generator 45, rhythm sound source 46, and tempo setting controller 47 constitute a rhythm generator. The type of rhythm (waltz, march, rock, rumba, begin, etc.) is selected by the rhythm selection controller 44, and the rhythm tempo is controlled by the tempo setting controller 47 accordingly.Rhythm pattern pulses are generated from the rhythm pattern pulse generator 45. is generated, and the rhythm sound source 46 is driven to obtain a desired rhythm sound. Then, code signals indicating the types of rhythms illustrated in b ₀ to _{b 2} of Table 2 are sent from the rhythm selection controller 44 via the input port (3).
It is sent to the CPU 21 and judged. For example, when waltz is selected by the rhythm selection controller 44, "011" is input to the input port (3),
When Begin etc. is selected, "100" is input to the input port (3). The AND circuit 48 logically multiplies the pulse signal of the first beat of the rhythm pattern pulse from the rhythm pattern pulse generator 45 and the rhythm start signal synchronized within the rhythm pattern pulse generator 45 by turning on and off the rhythm start switch. The tempo display device 50 displays the rhythm visually by lighting up a light emitting diode or the like. In addition, the pulse signal of each beat of the rhythm pattern pulse, the rhythm start signal, and the D-type flip-flop (DFF) described later.
51 and the AND circuit 49,
The signal is passed through an AND circuit 58 and inputted to the interrupt terminal INT ₃ of the CPU 21 so as not to be synchronized with the output of an OR circuit 56 (described later) by a NOT circuit 57. On the other hand, the signal from the terminal P3 that specifies playback of the input switch S14 is input to the D terminal of the D-type flip-flop (DFF) 51, and the pulse signal of the first beat of the rhythm pattern pulse is input to the T terminal for synchronization. AND the output with the input port (2)
The edge detection circuit 5
2, a single pulse is generated by the edge, and the AND circuit 54 performs a logical product with the rhythm start signal.
In addition, the rhythm start signal is input to the edge detection circuit 53 to generate a single pulse due to the edge, and the DFF 51
The AND circuit 55 performs logical product with the Q output of
OR circuit 5 outputs from AND circuits 54, 55, 58'
6 to the interrupt terminal INT ₂ of the CPU 21. Further, a signal from the terminal P1 specifying recording of the input switch S14 is passed through an edge detection circuit to generate a single pulse due to an edge, and is directly input to the interrupt terminal _INT1 of the CPU21. In the above configuration, the operation will be explained with reference to the operation waveform diagrams shown in FIGS. 2a to 2i. In the figure, a case is illustrated in which march (double beat) is selected by the rhythm selection controller 44. Figure a shows the signal at the P3 terminal of the input switch S14 during playback, Figure b shows the pulse signal of the first beat, and Figure c shows the pulse signal of each beat. The playback signal at the P3 terminal of switch S14 in figure a is DFF5 in figure g.
As shown in the output No. 1, synchronization is established with the pulse signal of the first beat in FIG. Next, the rhythm start signals ○a, ○b, and ○c in the same figure d show the case where the generation timings are different from the signal of the P3 terminal of the switch S14 in the same figure a, and the following figure d shows the case where the generation timing is different from the same figure e to i○b, ○b, ○C indicates the corresponding waveform. In the normal usage of ○i, switch S14 P3
After turning on the terminal, turn on the rhythm start switch, turn off the rhythm start switch when playback is finished, and then switch the switch S14 to the P2 terminal (off). ○In the usage of B, switch S1
If you turn on the rhythm start switch before turning on the P3 terminal of DFF51, the rhythm sound will start immediately, but playback will start from the moment the Q output signal of DFF51 becomes "1". On the other hand, in the usage of ○c, if switch S14 is switched to the P2 terminal (off) before turning off the rhythm start switch, the rhythm sound will continue, but the playback will be interrupted by the Q of DFF51.
It stops when the output signal becomes "0". In the figure h shows the output signal of the OR circuit 56, ○a,
The first pulses k ₁ , k ₂ , k ₃ of ○B and ○C are shown in g of the same figure.
This is a single pulse generated by the rising edge of the DFF 51, pulses _k1 and _k3 are the output signals of the AND circuit 58', and _k2 is the output signal of the AND circuit 58' and the AND circuit 54.
The output signals of the two are overlapping. Other pulses are single pulses and falling edges of the rhythm start signals ○a, ○b, and ○c shown in figure d.
This is a single pulse caused by the falling edge of DFF51. Each of these pulses is sent to the CPU 21 as an interrupt signal INT ₂ , and the Q of the DFF 51 shown in g in the same figure is
Between the pulses specified by the signal INT ₂ during the output signal period, the output signal of the AND circuit 58 corresponds to each beat pulse shown in the figure c, as shown in the figure i. A plurality of pulses ○i', ○b', and ○c' are generated and sent to the CPU ₂₁ as an interrupt signal INT3. In addition, AND in e of the same figure
The waveforms of the output signal of the circuit 48 and the output signal of the AND circuit 49 shown in FIG. Note that the processing of the interrupt signals INT ₂ and INT ₃ in the CPU 21 will be described in detail later with reference to FIG. The above describes the operation during playback, but when recording, turn on the input switch S14 to the P1 terminal.
Sends an interrupt signal to the interrupt terminal INT ₁ of the CPU 21, and records the recording information of the three types of accompaniment in Table 3 according to the musical score.
It is stored in the performance information memory (PIM) 60 via the CPU 21. Note that the pitch name code is based on Japanese Patent Application No. 54-38865. Table 4 shows PIM60 using the accompaniment piece in Figure 3 as an example.
This shows the procedure for storing the data in the file. Recording is performed in an orderly manner, including note names and notation symbols, and these are arranged at addresses A ₀ to A ₃₇ . In other words, even if the same pronunciation continues from the 3rd measure to the 4th measure, if there is a 1〓 at the boundary, it will be separated at this boundary.

【table】

[Table] (for addresses _A4 to _A12 ). The same pronunciation continues from measure 5 to measure 6, but there is nothing between them, so 1
It can be recorded as one (addresses A ₂₀ to A ₂₄ ). In general, in most cases, like the latter, it can be recorded as a single piece; for example, there are also other parts in measures 1 and 2 and measures 7, 8, and 9, and it is very often used as an accompaniment piece. Therefore, by applying this method, the memory can be used effectively as a whole and the memory capacity can be reduced. Also, the order of the octave and pitch name bytes is different between addresses A ₄ - A ₆ and A ₉ - A ₁₁ , but this is stored in addresses A ₇ - A ₁₂ of the assignment memory 24 by pressing the key during recording. This is due to the different order of the key codes. Also, in the case of a rest like address A ₁₃ , there is only a bar/beat byte, and there is no corresponding octave or note name byte. Figure 4 shows the patent application filed in 1973, which is surrounded by the dotted line in Figure 1.
This is a flowchart of the assigner 20 shown in No. 38865, and FIG. 5 shows a comparison of the flowchart when a program 59 and a performance information memory 60 are added thereto. That is, in FIG. 4, after starting, initial values are set in the event memory 25, assignment memory 24, etc., and then upper key, lower key, and foot key processing are performed.
In contrast to repeating the transfer of necessary contents of the assignment memory 24 to other blocks, in the present invention, "playback" indicates the contents of the automatic accompaniment program 59 as shown within the dotted line in FIG. and a “record” judgment routine is added. According to this, the flow is similar to that shown in FIG. 4 unless it is playback or recording. In the case of recording, after lower key processing, recording processing is performed, and then foot key processing is reached. For playback, lower key processing is excluded. Therefore, during playback, the lower key cannot be played manually. Since the playback process is executed by interrupt processing, it is not included in this flow. In FIG. 5, if the lower key processing and upper key processing are interchanged, recording and reproduction will be performed for the upper key. The same applies to foot key processing. FIG. 6 is a flowchart showing the recording procedure in the embodiment of FIG. In the figure, when the P1 terminal of input switch S14 is turned on, the corresponding information is transmitted from input port (2).
The information is input to the CPU 21, and as described in FIG. 5, the process proceeds through the recording process determined as YES in the "record" determination routine newly provided in the flowchart of the assigner 20. At the same time, an output signal is generated from the edge detection circuit connected to _the P1 terminal, and as shown in FIG. If it is determined, (00) # is stored in the performance information memory address counter (PIMC). Next, press the lower key of the key you want to record. However, when it is a rest, you do not need to press anything. Key codes are stored in addresses A ₇ to _{A 12} of the assignment memory 24 as appropriate. Nothing is stored when it is a rest. Now, enter the bar/beat byte information.
Press the input switch S0 to notify the CPU 21. Therefore, the judgment routine becomes YES,
If there is no correction in S13, (00)# is stored in the write counter (WC). The next □D is a delay time routine, which prevents the program processing from proceeding unnecessarily during the interval Δt in which the output of the AND circuit 43 in FIG. 1 is "1". That is, △ generated by one press of S0 to S13
t ends while executing □D, and the AND circuit 43
The output of returns to “0”. Therefore, the "SA" determination routine is reached after Δt. Next, in order to input the bar number code,
Select one from SA and press it. This switch code is stored in the bar/beat register, and it is checked in the "SA or S13" determination routine via □D, and if an error is made, the process returns to step. Next, to input the beat number code, select one from SA1 and press it. Store this switch code (ss ₁ to _{ss 4} ) in the measure/beat register, and
To distinguish that it is a beat byte, set the b ₇ and b ₆ bits of the bar/beat register to b 7 = b ₇ as shown in Table 3.
“1” and b ₆ = “0”. Next, in the process, the 1 channel address is placed in the assignment memory address counter (AMCM) and the contents of the assignment memory (AM) are read. From this content, as explained in the proposed example, it is determined whether the ON/OFF bit = 1, and if YES, the octave code and pitch name code are transferred and stored in the performance information memory (PIM). In order to distinguish between octave and pitch name bytes, the _b7 bit of PIM is set to "0" as shown in Table 3. After that, the performance information memory address counter (PIMC) and write counter (WC) are sequentially incremented to advance the AMAC channel address, and this is repeated until the final channel address.
In the process, the contents of the bar/beat register are
Transfer to PIM as bar/beat bytes. In this process, a pair of octave codes, an octave consisting of a note name code, a note name byte, and a measure/beat byte are transferred to the PIM. If you have selected it by mistake, press S13. Then, in the next flow, the process is executed and the PIMC
The WC is subtracted, the count is returned to the number before the PIMC transferred, and the transferred octave, note name byte, and bar/beat byte are invalidated. You will then have to return and start over. That is, after pressing S0, the measure number code is input using SA, and then the beat number code is input using SA1, and the process returns. After this, if it is desired to input the notation symbol bytes shown in S1 to S3 and SM2 in Table 1, the corresponding input switch is pressed, and NO is determined in the determination routine "S0". That is, S0 to S13
Of these, those that are neither S13 nor S0 are S1~
It can be seen that the switch is S12 and corresponds to the music notation symbol. Even in this case, if you press the switch by mistake, you only need to press S13 and then press the correct switch. If the determination routine "S0" is NO, a switch code for the corresponding notation symbol is created and this switch code is transferred to the PIM as a notation symbol byte. and PIM to distinguish between musical notation bytes.
Set the _b7 and _b6 bits to "1" and the _b3 bit to "0". After that, increment PIMC and WC and return. When recording is finished, press input switch S14.
Return to P2 terminal (off). As a result, an output signal is generated from the edge detection circuit connected to the P1 terminal again, and is input to the INT ₁ terminal of the CPU 21 to execute an interrupt, but since the switch S14 is now returned to the P2 terminal, as shown in Fig. 7. The remaining unrecorded items are determined as NO in the “record” determination routine.
Increment PIMC and store (00)# in the PIM area. This is because (00)# in PIM is used for determination as data outside the recorded area, as will be explained later in sub-notation symbol processing, and the previously recorded content remains. This prevents erroneous judgments. Note that this example assumes that PIM consists of 256 bytes, so the remaining
If you write all (00) # in the PIM area, it will inevitably become
PIMC returns to (00)#. FIG. 8 is a flowchart showing the reproduction procedure of the embodiment shown in FIG. First, the input switch S14 is turned on to the P3 terminal, and the rhythm start switch is turned on. The timing of turning on is as described in detail with reference to FIG. In other words, if input switch S14 is turned on to terminal P3, DFF5
1's Q output signal becomes "1", this signal is input to the CPU 21 from the input port (2), and as shown in FIG. YES to routine
, and the lower key processing is removed. This makes it impossible to play the lower keys manually. As shown in FIG. 2, the output signal of the OR circuit 56 consisting of a single pulse from each edge of the Q output signal of the DFF 51 and the rhythm start signal is sent to the interrupt terminal of the CPU 21.
An interrupt is executed by inputting to INT ₂ ,
Between the single pulses in the output period of DFF51
The output signal of the AND circuit 58 is the interrupt terminal of the CPU 21.
It is input to INT ₃ and interrupt of each beat pulse signal is executed. As is clear from ○A′, ○B′, and ○C′ in Figure 2, INT ₂ input occurs at the start and end of playback as described above, except for the k ₁ and k ₃ parts described later. . In Fig. 8, “00” is stored in the performance information memory address counter (PIMC) by the INT ₂ input, and the determination routine for “DFF Q = “1”” and the determination routine for “rhythm start signal = “1”” are executed. pass through. This determination routine distinguishes between the start and end of INT ₂ playback. In other words, if either of the determination routines is determined as NO, it is the INT ₂ input signal at the end of playback.
Jump to the flow of the process described below. If both of the judgment routines are judged as YES, since this is the INT ₂ input signal at the start of playback, the process is executed first, and the contents of PIM are judged as musical notation symbol bytes. In this case, even if there is a repeat start symbol or the like as a notation symbol at the beginning of the PIM, it is removed because it is unnecessary at the start of playback. The determination in the "notation symbol byte" determination routine is based on whether the b ₇ and bF bits are both "1". If this is NO, the process reads the PIM and transfers the octave and note name bytes to the lower key channel of the assignment memory (AM) until the bar/beat byte appears. The bar/beat byte is determined based on whether the _b7 bit is "1" and the _b6 bit is "0". If YES, the process proceeds. In addition, in the case of a rest, the octave,
Since there is no note name byte, the bar/beat byte appears immediately, and it immediately goes to the process without transferring anything to AM. In the process, the AM lower key channels that were not transferred in the process are returned to their initial state up to the final channel. This is to prevent unnecessary pronunciation due to preprocessed lower key codes that remain in the AM channel without being rewritten during transfer in the process. This is the end. In the process, the same pronunciation is made by multiplying the bar number code of the bar/beat byte by the time signature type input from the input port (3) to the CPU 21, and adding the beat number code with the digits aligned. Store in the beat register as the total number of beats. For example, if the bar number code is 3 and the beat number code is 2 in 3/4 time, the total number of beats is 3 x 3 + 2 = 11 beats. Then, in the sub musical notation symbol processing, PIMC is incremented and PIM is read out, and if the result is a musical notation symbol byte, the program returns after processing as shown in Table 5. If it is not a notation symbol byte, it is an octave, pitch name byte, or measure/beat byte, so it returns without doing anything. Then, as shown in FIG. 5, upon return, after the foot key processing, the desired playback sound is obtained when the necessary contents of the assignment memory 24 are transferred to other blocks 31, 32, and 33. . On the other hand, when the INT ₃ input is executed, the beat register is decremented. Then, if the beat register becomes (00) # according to the determination routine, the sound generation must be completed, so the next process is executed. (00)
If it does not result in #, the pronunciation is not yet complete and returns without doing anything. When executing the next process, read the contents of PIM. This is an octave, note name byte or measure/
If it's a beat byte, we have information about the next pronunciation or rest, so we go to the process and execute the same processing as before. When there is a rest, as shown in the example in Table 4 A ₁₃ , the bar/beat bite comes suddenly, so the pronunciation stops immediately as the process begins. If what is read is not an octave, note name byte, or bar/beat byte, but a 〓 or Fine notation that has been retained by sub-notation processing, this means that playback has ended, and the beat Set the register to (01)# and head to the process. In the process, all AM channels from channel 1 to the final channel are returned to their initial state and sound generation is stopped. Any INT ₃ input that occurs after this will go to the process and not to the process. The process executed by the INT ₂ input that occurs at the end of a playback is to obtain an initial state in preparation for the next playback. In the process, a notation symbol byte is selected in a "notation symbol byte" determination routine. And “1”
After all the UBs that were previously set are returned to the initial state of "0", the PIMC becomes (00)# and the initial state is also set. In this embodiment, it is assumed that PIM consists of 256 bytes, so if all PIMs are examined, PIMC will inevitably return to (00)#.
Then, the process returns all AMs to their initial state. Therefore, when the playback ends, the sound generation stops. As a result, if you turn off the rhythm start signal and then turn it on again, even in the middle of playback or after playback has ended, (after turning input switch S14 on to P2 terminal, it will turn on again to P3 terminal). ) Playback can be restarted from the beginning. In addition, for the process, switch S14 is set to P3.
The signal in Figure 2, which is the signal generated when the Q signal of DFF51 becomes “1” after the terminal is turned on, is
k ₁ , k ₃ (out of k ₁ , k ₂ , k ₃ , k ₂ is with
YES, but k ₁ and k ₃ are determined to be NO and the process is executed).
If you turn on the switch S14 to the P3 terminal while pressing some keys on the lower keyboard, even if you release the keys afterwards, the lower key processing will be out of the flow of the assigner 20, so the AM on/off bit will be changed. Prevents pronunciation from becoming unstoppable without being reversed.
In the case of k ₂ , automatic accompaniment is executed immediately, so there is no problem. The sub notation symbol processing procedure in FIG. 8 described above is shown in Table 5 below. This is program controlled using UB in Table 3. FIG. 9 is an example of a musical score showing notation symbols to which Table 5 is applied, and musical notes are omitted for simplicity.
It is a general rule in performance that a performance section that is repeated during the first pass is not repeated during the second pass. Note that solid lines indicate sections to be played, and dotted lines indicate sections to be skipped. Now, if the music notation symbols are arranged as shown in the diagram, from ○I to ○T, it starts from the treble clef of ○A, passes through the performance section, reaches the bottom of ○C, and calls this UB "1".

【table】

[Table] Jump to the bottom of UB=0 in ○Ro and set UB to “1” (Table 5 No. 1). Next, there is a 1〓 on the way after passing through the performance section from the bottom of ○B, but this is invalidated and after reaching the bottom of ○G, it returns to the original position of ○A (Table 5 No. 1, No.
2). Next, from the position of ○A, through the performance section, ○1 of E
〓 is reached, and since 〓 of 〇 is “1”, it jumps to 〓 of 〓, 2〓 (Table 5, No. 2). Next, after passing through the performance section from ○chi, it reaches the bottom of ○ru, and sets UB to "1", and the previous ○nu's 〓 is reached.
Jump to and set UB to “1” (Table 5 No. 1). Similarly, after passing through the performance section from ○nu, we reach the bottom of ○o, setting UB to "1", and then returning to the bottom of ○ri, setting UB to "1" (Table 5, No. 1). Next, it starts from ○ri, passes through the performance section, reaches DS, sets UB to "1", and then jumps to 〓 of ○ni (Table 5, No. 4). After reaching t ₀ 〓 of ○ through the performance section from 〓 of ○ D, it jumps to 〓 Coda of ○ (Table 5, No. 5). After the performance section from ○ka, ○ta's 〓
〓
It reaches 〓〓 and jumps to the previous UB=0 〇yo〓, then repeats the performance section during this time, and then turns down the volume with the expression pedal and ends the song as if it disappears. A sub notation symbol processing section is provided which stores such notation symbol bytes in the PIM and controls the program using the aforementioned UB. In this case, the configuration is simplified because it is only necessary to include a processing system that performs each function for each musical notation symbol. As explained above, according to the present invention, when recording performance information, one or more octaves to be generated at the same time formed from key press information of a keyboard switch,
For each set of music notation information consisting of note name bytes, the sound time information of one bar/beat byte consisting of the number of measures and the number of beats is set by operating the control switch, and the recorded key press information and sound time information are set. , along with performance order control information indicating various notation symbols, are transferred and stored in a performance information memory according to the musical score of accompaniment notes, and at the time of playback, the recorded key press information is used until the pronunciation time information is read from the performance information memory. After creating and retaining the key press information, the beat number of the beat number code, which is a component of the bar/beat byte of the read sound time information, is combined with the other components, such as the bar number code and the rhythm generator's beat. Based on the total number of beats obtained by adding the number of beats obtained from the product, determine the duration of sound generation and execute the sound generation of the stored key press information, and when the performance order control information is read out,
After searching the address of the performance information memory where the recorded key information to be generated next is stored, the next key depression information is generated. With this method, it is possible to faithfully reproduce the musical score, and as long as the accompaniment sounds are recorded in advance, the performer only has to play the melody part and the accompaniment will be automatic. In particular, in the present invention, measure/beat byte data is stored in the performance information memory independently from octave and note name byte data, so the same chord can be played over one measure or several measures, like an accompaniment note. This method is extremely effective as a recording method for songs in which the majority of data is recorded, as it reduces memory capacity and does not require the time and effort of inputting data.

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram showing the configuration of an embodiment of the present invention,
Fig. 2 is a waveform diagram showing the operation of the embodiment shown in Fig. 1, Fig. 3 is an example of an accompaniment piece, Figs. This is an explanatory diagram of the operation, and in the diagram, 10 is a keyboard circuit, 20 is an assigner, and 2
1 is the CPU, 22 is the CPU clock generator, 23 is the program, 24 is the assignment memory, 25
is event memory, 41 and 56 are OR circuits, 42
is a monostable multivibrator, 43,4
8, 49, 54, 55, 58, 58' are AND circuits, 44 is a rhythm selection controller, 45 is a rhythm pattern pulse generator, 46 is a rhythm sound source, 50 is a tempo display device, 51 is a D-type flip-flop, 5
2, 53 is an edge detection circuit, 57 is a NOT circuit,
Reference numeral 59 indicates a program, and 60 indicates a performance information memory.

Claims (1)

[Scope of Claims] 1. When recording performance information, a control switch is set for each set of recorded key press information consisting of one or more octaves and note name bytes that should occur at the same time formed from key press information of the keyboard switch. The sound time information of one measure/beat byte consisting of the number of measures and the number of beats is set by the operation, and the recorded key press information and sound time information are recorded according to the musical score of the accompaniment sound along with performance order control information indicating various notation symbols. It is transferred to and stored in the performance information memory, and during playback, key press information is created and held from the recorded key press information until the sound time information is read out from the performance information memory, and then the read sound time information is stored. Based on the total number of beats obtained by adding the number of beats of the beat number code, which is a component of the bar/beat bite, and the number of beats calculated from the product of the other component, the bar number code, and the beat of the rhythm generator, the sound is sustained. Determine the time and execute the sound generation of the stored recorded key press information, and when the performance order control information is read out,
An automatic accompaniment generation method characterized in that the address of a performance information memory storing recorded key press information to be sounded next is searched for, and then the next key press information is sounded. 2. The automatic accompaniment generation system according to claim 1, characterized in that in the case of a rest, sounding time information of only bar/beat bytes is provided.