JP3236869B2 - Electronic musical instrument and musical tone control method used for electronic musical instrument - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an electronic musical instrument capable of obtaining a delicate musical tone characteristic similar to that of a natural stringed musical instrument (for example, an acoustic guitar) by applying fuzzy inference.

[0002]

2. Description of the Related Art In recent years, with the rapid development of electronic technology, various techniques have been used to digitize stringed musical instruments using electronic technology. As a conventional stringed electronic instrument of this type, for example, a pitch extractor that picks up vibration of each string and extracts a string vibration period (pitch) from the picked-up pickup signal, and a pitch extractor extracted by the pitch extractor There is an electronic guitar that uses a pitch specifying device that specifies a corresponding pitch from a pitch.

Other electronic guitars include, for example, a switch that operates in response to a string operation and a pressure-responsive switch embedded in a matrix on a finger board. The former switch is a sound source. The latter switch is used to determine the pitch of the musical tone. In other electronic guitars, the envelope is also controlled.

[0004]

By the way, any conventional stringed electronic musical instrument has the characteristic of the musical tone alone based on the detection information of the fret operating position or the string operating position which is an element affecting the characteristic of the musical tone. Therefore, there is a problem in that it is not possible to obtain a musical tone characteristic close to a natural stringed musical instrument in detail according to the combination of these elements.

That is, in general, not only electronic guitars, but also natural stringed instruments such as acoustic guitars, an operation of pressing a string on a fingerboard with a left hand, that is, a fret operation and a play of playing one or more strings with a right hand. A string operation is performed, whereby the characteristics (for example, pitch) of the generated musical sound are determined.

In this case, the fret operation position includes a high position and a low position, and the string operation position includes a saddle side and a fret side. Therefore,
The characteristics of the musical tone (pitch, timbre, volume, volume, etc.) generated according to the combination of the fret operation position and the string operation position
Effects, envelopes, etc.) are slightly different.

More specifically, the characteristic of a musical tone changes as follows depending on each element. (I) Fret operation position When the fret operation position is in the high position, the envelope shape of the generated musical tone changes as shown in FIG. 27 (a), and its level rapidly attenuates as time (indicated by t) elapses. . On the other hand, when the fret operation position is at the low position, the envelope shape of the generated musical tone changes as shown in FIG. 27B, and the level gradually decreases with time.

(II) String operation position When the string operation position is on the saddle side (the side supporting the string), the envelope shape of the generated musical tone changes sharply as shown in FIG. Sounds harder, and the level decays quickly over time. On the other hand, when the string operation position is on the fret side (the side where the strings are bridged), the envelope shape of the generated musical tone changes completely as shown in FIG. 28 (b), and the tone becomes soft, Decays slowly over time.

Therefore, in order to obtain a tone characteristic close to that of a natural stringed instrument, it is not sufficient to simply change the characteristic of the tone based on the detection information of the fret operation position or the string operation position, as described above. It is necessary to control the characteristics of musical tones in consideration of a situation in which the pitch or the like changes subtly depending on the combination of the operation position and the string operation position.

[0010] Such a subtle change in pitch or the like depends on the performance sensation of a human, but in any case, in response to a demand for obtaining a musical tone characteristic that is fine and close to that of a natural stringed instrument, for example, a microcomputer is used to perform a normal operation. Although it is conceivable to cope with this by means of arithmetic processing, there is a drawback in that if this is attempted by ordinary arithmetic processing, an enormous amount of processing time is required.

For example, it is necessary to perform a complicated and large amount of calculation using a large number of input parameters which are considered to be related to the combination of each element, and the number of processes is dramatically increased, thereby realizing a realistic electronic musical instrument. It is difficult (for example, the cost is increased).

Therefore, the present invention applies fuzzy inference,
It is an object of the present invention to provide an electronic musical instrument capable of finely controlling the characteristics of musical tones generated at least in accordance with a combination of a fret operation position and a string operation position.

[0013]

According to an aspect of the present invention, there is provided an electronic musical instrument comprising: a fingerboard; a plurality of strings stretched along the fingerboard;
String operation detecting means for detecting that a string has been struck, string operation position detecting means for detecting a string position of a string whose string has been detected by the string operation detecting means, and fingering for the finger board Fingering operation position detection means for detecting the operation position of the finger, fuzzy inference means for performing a fuzzy inference operation in accordance with a predetermined fuzzy rule using the string position of the string and the fingering operation position as input parameters, and a calculation result of the fuzzy inference means And a musical sound generation instructing means for instructing generation of a musical sound to which an envelope that attenuates at a release rate corresponding to the musical tone is added .

In a preferred aspect, the fingering operation position detecting means detects a fret operation position of the finger board as the fingering operation position.

An electronic musical instrument according to a third aspect of the present invention.
Stretched along the fingerboard and the fingerboard
A set of strings and a string that detects when the string is struck
Operation detection means and detection of bullet strings by the string operation detection means
String operation position detection for detecting the string position for the selected string
Means and fingering for said fingerboard
Fingering operation position detecting means for detecting the operation position of a finger
Operating string detecting means for detecting the number of the string that has been struck
And the string position, fingering operation position and string number
Use the bar as an input parameter and follow the predetermined fuzzy rules.
Fuzzy inference means for performing a fuzzy inference operation
Reduced at the release rate according to the calculation result of the fuzzy inference means.
A music instructing the generation of musical sounds with a decaying envelope
Sound generation instructing means.

In a preferred embodiment, the fingering
The operation position detecting means is used as the fingering operation position.
Detecting the fingerboard operation position of the fingerboard
It is characterized by.

The electronic musical instrument according to claim 1 or 3 is
Music whose sound is instructed by the musical sound generation instructing means.
A tone generating means for generating a sound is provided.

An electronic musical instrument according to the present invention.
The tone control method is stretched along the fingerboard
Detected that one of multiple strings was hit
And the string position of the string with respect to the detected string
Detecting, and for the finger board
Detecting a fingering operation position; and detecting the fingering operation position.
The string position and fingering operation position
Force parameters according to the prescribed fuzzy rules.
Performing a fuzzy inference operation, and fuzzy inference operation
The envelope that attenuates at the release rate according to the result
Instructing the generation of the added musical tone.

An electronic musical instrument according to a sixth aspect of the present invention.
The tone control method is stretched along the fingerboard
Detected that one of multiple strings was hit
And the string position of the string with respect to the detected string
Detecting, and for the finger board
Detecting a fingering operation position;
Detecting the number of the detected string; and
String position, fingering operation position and string number
According to a predetermined fuzzy rule as an input parameter
Performing a fuzzy inference operation;
Envelope that attenuates at the release rate according to the calculation result
Instructing the generation of a musical tone to which a loop has been added.
You.

[0020]

According to the first or sixth aspect of the present invention, when one of the plurality of strings is struck, the string position (saddle side,
The release rate of the envelope added to the generated musical tone is controlled by fuzzy inference using the fret side and the fingering operation position (high position or low position) as input parameters. Therefore,
According to the delicate combination of the position of the string and the position of the fingering when the string is struck, the characteristics of the generated musical tone change delicately, and a musical sound characteristic very similar to a natural stringed instrument can be obtained.

[0021] In the invention of claim 3 or 7, wherein the string number is also one in addition et been in fuzzy inference of the input parameters is carried out. Therefore, it is possible to generate a musical tone close to a natural musical instrument by finely changing the musical tone characteristic more finely. Further, by applying the fuzzy inference, the number of processes required for the arithmetic processing is reduced, and the arithmetic load is reduced.

[0022]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an overall configuration of an embodiment when an electronic musical instrument according to the present invention is applied to an electronic stringed instrument (especially, an electric guitar).

In FIG. 1, the electric guitar 1 is roughly divided into a body 2, a neck 3 extending from the body 2, and a head 4 attached to the tip of the neck 3, and an upper surface of the neck 3 A finger board 6 on which the frets 5 are protruded and arranged is formed.

Six strings 7 are stretched on the finger board 6, one end of each string 7 is adjustably supported by a beg 8 provided on the head 4, and the other end is connected to a bridge 9 provided on the body 2. Supported. At the position on the body 2 opposite to each string 7, two types of pickups 10, 20 of a magnetic type or a piezoelectric type are provided, and the vibration of each string is detected. In addition, since the two systems of pickups 10 and 20 are provided, a bullet string position (saddle side or fret side) with respect to the string 7 is detected by signal processing as described later.

Then, the vibration period (pitch) of the string 7 and the string touch data indicating the strength of the plucked string are extracted from the picked-up string vibration signal. Further, a tremolo arm 21 is connected to the bridge 9, and the tension of the string 7 is adjusted by the tremolo arm 21 to change a string vibration, that is, a pitch.

Other switches and the like are also arranged on the body 2, and FIG. 1 shows a power switch 22, a tone selection switch 23, an effect mode selection switch 24, and a tone parameter setting panel 25. Power switch 2
Reference numeral 2 turns on / off the power, and the tone color selection switch 23 is used for selecting a tone color.

The effect mode selection switch 24 includes a chorus switch, a delay switch, a tremolo switch,
It consists of a reverb switch, and is used to select musical tone effects during performance. The musical tone parameter setting panel 25 is used to set effect parameters, set envelope parameters, and set each performance operator (string 7, tremolo arm 21, etc.) or a performance sensor (pickup 10,
20, tremolo arm sensor 26, etc.).

Two speakers 27a and 27b shown in the body 2 and the head 4 respectively convert an electronically generated tone signal into an acoustic signal and emit the sound to the outside.

The fingerboard 6 is made of an elastic rubber, and as shown in detail in FIG. 2, fret switches 30 are embedded at respective positions on the lower surface of the fingerboard 6 corresponding to the respective strings 7 between adjacent frets. ing. In FIG. 2, the second
The state where the fret switch 30 is buried at the third fret position is shown, but the fret switch 30 is buried for each fret, and FIG. 2 shows the buried state of the fret switch 30 corresponding to the other fret positions. Abbreviated.

Therefore, the plurality of fret switches 30
Thereby, the fret position with respect to the finger board 6 is detected. The fret switch 30 may be buried for each fret. For example, every other fret switch may be buried.
O. It is only necessary to be able to secure information.

FIG. 3 is a block diagram showing a configuration of an electronic circuit of the electric guitar 1. In FIG. 3, the fret switches 30 provided in the finger board 6 are connected so as to form a key matrix circuit 31, and the key scan circuit 3 connected to the key matrix circuit 31
2, the fret switches 30 are scanned and the state of each fret switch 30 is detected. The scan result of the key scan circuit 32 is taken into the microcomputer 40, and thus the operation fret position for each string 7 is detected.

The fret switch 30, the key matrix circuit 31, and the key scan circuit 32 constitute fingering operation position detecting means 33. Note that this embodiment is an example of application to an electric guitar.
When applied to a violin, a fingering operation position with respect to a finger board may be detected.

Numeral 34 comprehensively represents switches (for example, tone selection switch 23, effect mode selection switch 24) arranged on the body of the electric guitar 1.
The state of each switch of the switch section 34 is also monitored by the microcomputer 40.

The tone parameter setting panel 25 is also connected to the microcomputer 40, and the set tone parameters are stored in the memory of the microcomputer 40.

A signal from the first pickup 10 disposed on the saddle side for detecting the vibration of the string 7 is supplied to an amplifier 41.
The signal is input to an envelope detection circuit 43 via a DC blocking capacitor 42. Similarly, a signal from the second pickup 20 disposed on the fret side for detecting the vibration of the string 7 is input to an envelope detection circuit 53 via an amplifier 51 and a DC blocking capacitor 52.

The first pickup 10 has six pickups for detecting vibrations of six strings from the first string to the sixth string. The first pickup 10 includes first to sixth string pickups. Is shown. Similarly,
The second pickup 20 also has 6 strings from the first string to the sixth string.
There are six pickups for detecting the vibration of one string, each of which is illustrated as a first string pickup through a sixth string pickup.

The above pickups 10, 20 and amplifier 4
1, 51, capacitors 42 and 52, and envelope detection circuits 43 and 53, as a whole, string operation detecting means 60 for detecting that the string 7 has been struck, and a string operation position detection for detecting a string position with respect to the string 7. It comprises a means 61 and an operating string detecting means 62 for detecting the number of the string that has been struck.

The string operating position detecting means 61 determines the size of the output from the first pickup 10 on the saddle side and the output from the second pickup 20 on the fret side, thereby determining It is determined whether the string 7 has been picked up to detect the string operating position. Operation string detecting means 62
Is the output of the first pickup 10 (or the second pickup 2) arranged for each of the six strings 7
The string number of the struck string is detected from the output of 0.

The amplifiers 41 and 51 may be the same as those used in a pickup signal processing circuit of a pick extraction type guitar, for example. Envelope detection circuits 43, 5
Reference numeral 3 denotes an operational amplifier OP which receives a pickup signal grounded via a resistor R1 at a non-inverting input terminal, and an operational amplifier O
One end of the diode D is connected to the output terminal of the diode P, the other end is connected to the cathode output of the diode D, and the other end is grounded. Is fed back to the inverting input terminal of the operational amplifier OP.

The envelope detection circuits 43 and 53 detect the envelope of the string vibration. Envelope detection circuit 43
Is output through each gate 44 controlled by gate adjustment signals G1 to G6 from the microcomputer 40.
The signals are multiplexed on one common line and input to the A / D converter 45. The A / D converter 45 converts the analog envelope signal from the selection gate 44 into a digital signal in response to an A / D start command signal sent from the microcomputer 40 while one of the gates 44 is open. Convert to a signal.

When the conversion is completed, the A / D converter 45 sends an end command signal to the microcomputer 40. In response, the microcomputer 40 sets the A / D converter 4
5 is read and the converted envelope signal (vibration level data) is stored. Thereafter, the microcomputer 40 closes the selected gate 44, selects (opens) the next gate 44, and reads the string vibration envelope data for the next string.

The microcomputer 40 performs the same processing for the output of the other envelope detection circuit 53. That is, the output of the envelope detection circuit 53 is multiplexed on one common line via each gate 54 controlled by the gate adjustment signals K1 to K6 from the wait 40 and input to the A / D converter 55. The A / D converter 55 converts the analog envelope signal from the selection gate 54 into a digital signal in response to an A / D start command signal sent from the microcomputer 40 while one of the gates 54 is open. Convert to a signal.

When the conversion is completed, the A / D converter 55 sends an end command signal to the microcomputer 40. In response, the microcomputer 40 sets the A / D converter 5
5 is read and the converted envelope signal (vibration level data) is stored. Thereafter, the microcomputer 40 closes the selected gate 54, selects (opens) the next gate 54, and reads the string vibration envelope data for the next string.

The amplifiers 41 and 51, the capacitor 42,
The reference numeral 52 and the envelope detection circuits 43 and 53 constitute an analog processing circuit 63 for a string vibration pickup signal.

The microcomputer 40 is an ALU 70,
A ROM 71, a working RAM 72, and a timer 73 are used to process performance input data from each of the performance operators, and a sound source is generated based on the processed input data and information set by the tone parameter setting panel 25. 80, controls the envelope generation circuit 81; The microcomputer 40 has a membership function table R
OM74 is externally connected.

Also, as a feature of the present embodiment, the microcomputer 40 receives performance input data from each performance operator, and based on a predetermined program stored in the ROM 71, the string position, the fret operation position, and the string position. Fuzzy inference is performed in accordance with a predetermined fuzzy rule using the number as an input parameter, processing necessary for controlling the characteristics of musical sounds is performed in accordance with the inference output, and control data is supplied to one system of PCM sound source 80. The processing of the microcomputer 40 is shown by a flowchart described later.

The membership function table ROM 74 stores table data representing membership functions used in fuzzy inference. Working R
The AM 72 is used as a working area when the microcomputer 40 executes control, and temporarily stores information such as pitch data indicating a pitch, for example.

The PCM tone generator (tone generator) 80 is a polyphonic PCM tone generator that operates in a time-division manner, has a waveform ROM inside, and reads out based on designated tone color data and scale data supplied from the microcomputer 40. The address is supplied to the waveform ROM, and the selected waveform data is read from the waveform ROM storing the musical tone waveforms of a plurality of timbres and output to the multiplier 82. The multiplier 82
Envelope generating circuit 8 for the output of PCM sound source 80
Multiply the envelope output from 1 and output the multiplication result to the sound generator 83.

The tone generator 83 generates a corresponding musical tone according to the output of the multiplier 82. Thereafter, the output of the tone generator 83 is supplied to, for example, a D / A converter circuit, where the digital signal is converted into an analog signal, and an analog tone signal is output to the outside via an output unit including a low-pass filter, an output amplifier, and the like. Is emitted.

The microcomputer 40 independently constitutes a tone generation instruction means. The microcomputer 4
0 and the membership function table ROM 74 constitute the fuzzy inference means 100 as a whole.

Next, the fuzzy rules for controlling the tone characteristics will be described. 4 (a), (b) and (c) are membership functions of the antecedent part, of which FIG. 4 (b) is a membership function relating to the position of the string, and FIG. 4 (c) is a membership function relating to the string N.
O. Membership function.

4D and 4E show the fuzzy output in the consequent part. FIG. 4D shows the membership function related to the attack rate, and FIG. 4E shows the member related to the release (attenuation) rate. It is a ship function.

The meaning of the label in each membership function is as follows. FL: “The fret number is large” FS: “The fret number is small” PL: “The bullet string position is on the fret side” PS: “The bullet string position is on the saddle side” SL: “The string number is large SS: "String NO. Is small" AL: "Strengthen attack rate" ALL: "Slightly increase attack rate" ALS: "Slow attack rate" AS: "Extremely slow attack rate" RL: "Release release rate" RLL: "Release release rate slightly" RLS: "Release release rate slowly" RS: "Release release rate extremely slowly"

The fuzzy rule is a so-called IF, THE
N (if any). In the present embodiment, the following four rules are adopted for the attack rate, and the following four rules are adopted for the release rate. The rules for attack rate are shown in Table 1 below.
Is shown as

[0055]

[Table 1]

Also, the rules regarding the release rate are as follows:
This is shown in Table 2 below.

[0057]

[Table 2]

First, an attack rate rule will be described with reference to Table 1. Attack rate rule Rule A: IF fret NO. = FS string position =
PS AND String NO. = SS THEN fuzzy output = AL (strong attack rate)

The fuzzy rule A is described as “if fret N
O. Is small (low position), the string position is on the saddle side, and the string NO. If is small, increase the attack rate. It means.

Rule B: IF fret NO. = FS
String position = PS AND String NO. = SL OR fret NO. = FS string position = PL AND string NO. = SS OR fret NO. = FS string position = PL AND string NO. = SL THEN Fuzzy output = ALL (Attack rate somewhat strong)

The fuzzy rule B is described as “if fret N
O. Is small (low position), the string position is on the saddle side, and the string NO. Is large, or the fret NO. Is small (low position), the string position is on the fret side, and the string NO. Is small, or the fret NO. Is small (low position), the string position is on the fret side, and the string NO. Is larger, the attack rate is set to be slightly higher. It means.

Rule C: IF fret NO. = FL
String position = PS AND String NO. = SS OR fret NO. = FL string position = PS AND string NO. = SL THEN Fuzzy output = ALS (Slow attack rate)

The fuzzy rule C is described as “if fret N
O. Is large (high position), the string position is on the saddle side, and the string NO. Is small, or the fret NO. Is large (high position), the string position is on the saddle side, and the string NO. When is large, the attack rate is made slow. It means.

Rule D: IF fret NO. = FL
Bullet position = PL AND String NO. = SS OR fret NO. = FL string position = PS AND string NO. = SL THEN Fuzzy output = AS (Attack rate is extremely slow)

The fuzzy rule D is "if, fret N
O. Is large (high position), the string position is on the fret side, and the string NO. Is small, or the fret NO. Is large (high position), the string position is on the saddle side, and the string NO. Is large, the attack rate is made extremely slow. It means.

Next, the rules for the release rate will be described with reference to Table 2. Release rate rule Rule E: IF fret NO. = FL string position =
PS AND String NO. = SL THEN Fuzzy output = RL (Release rate is fast)

The fuzzy rule E is "if, fret N
O. Is large (high position), the string position is on the saddle side, and the string NO. If is large, increase the release rate. It means.

Rule F: IF fret NO. = FL
String position = PS AND String NO. = SS OR fret NO. = FL string position = PL AND string NO. = SL OR fret NO. = FS string position = PS AND string NO. = SS THEN Fuzzy output = RLL (Release rate is slightly faster)

The fuzzy rule F is described as “If fret N
O. Is large (high position), the string position is on the saddle side, and the string NO. Is small, or the fret NO. Is large (high position), the string position is on the fret side, and the string NO. Is large, or the fret NO. Is small (low position), the string position is on the saddle side, and the string NO. If is small, increase the release rate slightly. It means.

Rule G: IF fret NO. = FL
Bullet position = PL AND String NO. = SS OR fret NO. = FS string position = PL AND string NO. = SL OR fret NO. = FS string position = PS AND string NO. = SL THEN Fuzzy output = RLS (Release rate is slow)

The fuzzy rule G is defined as “if fret N
O. Is large (high position), the string position is on the fret side, and the string NO. Is small, or the fret NO. Is small (low position), the string position is on the fret side, and the string NO. Is large, or the fret NO. Is small (low position), the string position is on the saddle side, and the string NO. If is large, the release rate should be slow. It means.

Rule H: IF fret NO. = FS
Bullet position = PL AND String NO. = SS THEN Fuzzy output = RS (Release rate is very slow)

The fuzzy rule H is defined as “if fret N
O. Is small (low position), the string position is on the fret side, and the string NO. When is small, the release rate is made extremely slow. It means.

Next, the operation of the electric guitar of this embodiment will be described. Main Program FIG. 5 is a main program showing processing of musical tone generation control. When the power of the electronic guitar is turned on, first, step S
In step 1, an initialization process is performed. This allows, for example,
The reset of the register of the microcomputer 40, the clearing of the working RAM 72, and the initialization of a subroutine described later are performed.

Next, at step S2, the string NO. (Hereinafter referred to as a string pointer) n is set to [1]. The state where n = [1] is the string NO. Is [1], which corresponds to the first string. This is, from the first string,
This is to check the situation of the bullet. As a result, as described later, the string NO. Is detected. The value of the string pointer n is
0 is stored in the n register.

Next, in step S3, it is determined whether or not the string flag Ln for sound generation processing is [1]. The bullet string flag Ln is for instructing continuation of sound generation when the string 7 is pulled and its vibration level is equal to or higher than a certain level. When the string flag Ln = [0], it is determined that the string 7 has not been pulled, and the processing after step S4 is performed. on the other hand,
When the string flag Ln = [1], the string 7 is pulled, and it is determined that the vibration level is equal to or higher than a certain value, and the process proceeds to step S22 described later to continue the sound generation process.

When the routine proceeds to step S4 with the string flag Ln = [0], the gate adjustment signal Gn is turned on (that is, "H").
Level). For example, in this routine, n =
Since [1], Gn = G1, and the gate adjustment signal G1 becomes "H". As a result, the gate 44 turns on. Therefore, the output of the envelope detection circuit 43, which is the detection result of the vibration of the first string on the saddle side, is output from the multiplexed one common line via the gate 44 to the A / D converter 45.
Is input to the microcomputer and is A / D converted.

Next, similarly, in step S5, the gate adjustment signal Kn is turned on ("H" level). In this routine, since n = [1], Kn = K1, and the gate adjustment signal K1 becomes “H”. As a result, the gate 54 is turned on. Therefore, the output of the envelope detection circuit 53, which is the detection result of the vibration of the first string on the fret side, is output from one common line multiplexed through the gate 54 to A
After being input to the / D converter 55 and A / D converted, it is taken into the microcomputer 40.

Next, in step S6, the string vibration envelope data A / D converted by the first A / D converter 45 is stored in the A register inside the microcomputer 40. Thus, the vibration information of the first string on the saddle side is taken into the microcomputer 40.

In step S7, the string vibration envelope data A / D converted by the second A / D converter 45 is stored in the B register inside the microcomputer 40. Thus, the vibration information of the first string on the fret side is taken into the microcomputer 40.

Next, in step S8, the vibration information of the first string on the saddle side stored in the A register, that is, the string vibration envelope data of the first string (hereinafter, represented by A) is compared with a predetermined threshold value TH1. The threshold value TH1 is a threshold level for determining that a string has been struck when the vibration level of the string 7 is equal to or greater than a certain value.

When the string vibration envelope data A is equal to the threshold T
If it is equal to or greater than H1, the process proceeds to step S9, and the string flag L
Set n to [1]. As a result, it is determined that the string has been struck, and sound generation processing will be performed later. Since it is determined whether or not a string has been struck, the vibration information of the first string on the fret side is not compared with the threshold value TH1.

On the other hand, if the string vibration envelope data A is less than the threshold value TH1, the process proceeds to step S10, where the string pointer n is incremented by [1]. Then, in step S11, it is determined whether n = [7]. Is determined. n
If not = [7], the process returns to step S3 to repeat the same processing. As a result, the string discrimination processing for the second string is performed. If the string vibration envelope data A is equal to or larger than the threshold value TH1, the process proceeds to step S9, where the string string flag Ln is set to [1]. , The second string is sounded. Hereinafter, in the same manner, the third
A bullet string / string to sixth string discrimination process is performed.

When n = [7] in step S11, the process branches to YES and returns to step S2, where n = 7 again.
The same processing as described above is repeated from the state of [1]. In this way, the first to sixth strings on the saddle side are subjected to the string determination processing. At this time, the number of the string 7 played at the same time,
That is, the string NO. Is detected.

After branching from step S8 to YES and setting the string flag Ln to [1] in step S9, the process proceeds to step S12. In step S12, the position ratio C for determining the string position is calculated in accordance with the equation C = A / B. Here, B is vibration information of the string 7 on the fret side stored in the B register, that is, string vibration envelope data.

By calculating the position ratio C, it is possible to determine whether the string 7 has been hit on the saddle side or on the fret side. For example, when the value of C is large, the string vibration envelope data A is large.
It can be determined that the ball has been hit on the saddle side, and if the value of C is small, the string vibration envelope data B is large, so it can be determined that the ball has been hit on the fret side.

Next, in step S13, the n-string fret switch 30 is scanned. Since the fret switches 30 are connected to form a key matrix circuit 31, the fret switches 30 are scanned by a key scan circuit 32 connected to the key matrix circuit 31, and the state of each fret switch 30 is detected.

Next, in step S14, it is determined whether or not the fret switch 30 is turned on. If it is not turned on (if it is off), the value of the m register is set to [0] in step S15. m register is the fret NO. Is to store. Therefore, the state where m = [0] corresponds to the case where the string 7 is played with the open string. On the other hand, when the fret switch 30 is ON, the process proceeds to step S16, and the fret NO. Is stored in the m register. Thus, the operation fret position of the string 7 is detected.

Turning to FIG. 6, step S17, step S17
At 18, a fuzzy inference process is performed. Specifically, m
The fret number stored in the register , The string position and the string NO. Of whether it was hit on the saddle side or on the fret side. Is used as an input parameter to perform fuzzy inference, and calculate an attack rate and a release rate that vary the envelope of the musical sound characteristic.

As a result, the string position (saddle side,
Fret side), fret operation position (high position or low position), string NO. The envelope of the generated musical tone changes subtly according to each subtle combination of (large or small). Step S1
The detailed processing contents of 7 and step S18 will be described later in a subroutine.

Next, in step S19, the fret NO. And the pitch data based on the contents of the string pointer n stored in the n register.
As a result, the pitch of the musical tone is determined by the type of string played and the number of the fret pressed.

Next, at step S20, the PCM tone generator 8
The pitch data is sent to the n-channel of 0, and the sound generation process is performed.
There are six channels, for example, in accordance with the number of strings 7, and a tone of pitch data corresponding to the number of the played string (that is, n) is generated. Next, the data of the attack rate A and the attack level is sent to the envelope generation circuit 81 in step S21. As a result, the attack portion of the envelope of the generated musical tone is varied according to the attack rate A and the attack level. For example, the attack rate becomes strong, slightly strong, gentle, or extremely slow (almost variable). No).

In step S22, the envelope level of the n-channel (for example, the channel corresponding to the first string) in the PCM tone generator 80 (the envelope of the waveform actually sounding) is fetched. In step S23, the fetched envelope level is determined. With the threshold value TH2. The threshold value TH2 is a threshold level (ie, a silencing threshold level) for starting silencing processing from the envelope level in response to the vibration level of the string 7 falling below a certain level. Note that the n-channel has a value corresponding to the first to sixth strings.

If the envelope level of the n-channel exceeds the threshold value TH2, steps S24 and S25 are executed.
Jumps to step S26, where the target flag AR
It is determined whether or not Vn is [1]. Target flag ARV
n is up to ARV ₀ ~ARV _6, corresponds to the number of the n-channel. This target flag ARVn is set to the channel n
This is for determining whether or not the envelope has reached the maximum value every time. When the envelope reaches the maximum value in the attack section, ARVn is set to [1].

If it is determined in step S26 that ARVn = [1], the attack portion is terminated and the envelope shifts to the release portion. Therefore, in step S27, the release rate R and the release level are applied to the n channel of the envelope generation circuit 81. Send data. As a result, the release portion of the envelope of the generated musical tone is varied according to the release rate R and the release level, and for example, the release rate becomes faster, slightly faster, slower, or extremely slower.

On the other hand, ARVn is set to [1] in step S26.
If not, it is determined that the envelope is still in the process of changing the attack portion, and the process jumps from step S27 and returns to step S10.

Here, the processing of the release rate in step S27 will be specifically described. For example, when the release rate is increased, as shown in FIG. 28A, the release portion of the envelope of the generated musical tone attenuates as time elapses, as in the case where the string operation position is on the saddle side ( The envelope changes sharply). When the release rate is slightly increased, as shown in FIG. 27A, the release portion of the envelope of the generated musical tone attenuates as time elapses, similarly to the case where the fret operation position is at the high position.

When the release rate becomes gentle, the release portion of the envelope of the generated musical tone gradually attenuates as time elapses, as shown in FIG. 27B, as in the case where the fret operation position is the low position. I do.

Further, when the release rate becomes extremely slow, as shown in FIG. 28B, the release portion of the envelope of the generated musical tone is generated with time, similarly to the case where the string operation position is on the fret side. And decay very slowly (the envelope is rounded and hardly changes). That is, the sound is round and soft.

Further, such an envelope change corresponds to the string N
O. For example, FIG. This shows how the envelope changes depending on the element.

String NO. Is small (for example, the first string,
At around the second string), the envelope shape of the generated musical tone changes as shown in FIG. 7A, and the level gradually decreases with time. This is the string NO. It is a graph at the time of picking with the same strength regardless of. The reason why the envelope is gradually attenuated is that the string is thin and the vibration time is long.

On the other hand, when the string NO. Is large (for example, around the sixth and fifth strings), the envelope shape of the generated musical tone changes as shown in FIG. 7 (b), and the level attenuates rapidly with time. It should be noted that the envelope shapes of the attack and the attack part also slightly change. The reason why the envelope attenuates quickly is that the vibration time is short because the strings are thick and the air resistance is large.

When the processing in step S27 is completed, the flow returns to step S10, and the string pointer n is incremented. As a result, the same processing as described above for generating or silencing a musical tone for the next string is performed.

On the other hand, if the envelope level of the n-channel is equal to or smaller than the threshold value TH2 in step S23, the process proceeds to step S24, where the string flag Ln is reset to [0]. A mute process is performed on n channels (for example, a channel corresponding to the first string). Then, step S26
Proceed to.

As a result, when the envelope level of the currently sounding channel falls below a certain level, the sound is muted, and at the same time, the string flag Ln is reset, and the process proceeds to the tone control process of the next channel.

By applying the fuzzy inference as described above, the fret operation position, the bullet string operation position, and the string NO. The characteristics of the musical tone generated in accordance with the combination of are controlled delicately.

Attack Rate Calculation Processing FIG. 8 is a flowchart showing a subroutine of the attack rate calculation processing in step S17 of the main program. First, in step S31, the fret operation position,
String operation position and string NO. The above-described calculation of the fuzzy rule A is performed based on

Similarly, steps S32 and S3
3. In step S34, the fret operation position, the bullet string operation position, and the string NO. The fuzzy rule B, the fuzzy rule C, and the fuzzy rule D described above are respectively calculated based on. Detailed processing contents of these fuzzy rules A, B, C, and D will be described in a subroutine described later. As a result, the processing of the antecedent part in the fuzzy inference is performed, and the fitness for each membership function is obtained.

Then, in step S35, the maximum value A (A: A) that takes the maximum value of the data obtained by the above rule operation
(Corresponding to the meaning of attack)) processing (that is, MAX operation = OR processing), and the center of gravity calculation A of this data is performed in step S36 to perform defuzzification. Then, the process returns to the main program. Thereby, the processing of the consequent part in the fuzzy inference is performed, and the attack rate A level is obtained. The detailed processing contents of the maximum value A calculation and the center of gravity calculation A will be described in a subroutine described later.

Release Rate Calculation Processing FIG. 9 is a flowchart showing a subroutine of release rate calculation processing in step S18 of the main program. First, in step S41, the fret operation position,
String operation position and string NO. The above-described calculation of the fuzzy rule E is performed based on.

Similarly, steps S42 and S4
3. In step S44, the fret operation position, the bullet string operation position, and the string NO. The fuzzy rule F, the fuzzy rule G, and the fuzzy rule H described above are respectively calculated based on. The detailed processing contents of these fuzzy rules E, F, G, H will be described in a subroutine described later. As a result, the processing of the antecedent part in the fuzzy inference is performed, and the fitness for each membership function is obtained.

Next, at step S45, the maximum value R (R: R:
In addition to performing processing (corresponding to the meaning of release) (ie, MAX operation = OR processing), the center of gravity R of this data is calculated in step S46 to defuzzify. Then, the process returns to the main program. As a result, the processing of the consequent part in the fuzzy inference is performed, and the release rate R level is obtained. The detailed processing contents of the maximum value R calculation and the center of gravity calculation R will be described in a subroutine described later.

Fuzzy Rule A Calculation Process FIG. 10 is a flowchart showing a subroutine of the fuzzy rule A calculation process. First, in step S51, the fret NO. (Data of the antecedent part) Da is read from the FS table in the membership function table ROM 74 in which the membership function “the fret number is small (ie, low position)” is stored. The data in this case is shown in FIG.
And is a value in the range of 0 to 1.

That is, the data Da is the fret NO. Is the degree of conformity indicated by the rule A with respect to the input value. The data Da is stored in a register Da of the same name in the microcomputer 40, which stores data Db and Dc to be described later.
The same applies to.

Similarly, in step S52, a membership function "the string position is on the saddle side" in the membership function table ROM 74 is stored using the string operation position (hereinafter simply referred to as the string position) C as an address. From the existing PS table (antecedent data) D
Read b. The data in this case corresponds to the fitness (grade) of the membership function PS shown in FIG. 4B and is a value in the range of 0 to 1. That is, the data Db is the degree of conformity indicated by the rule A with respect to the input value of the string position C.

In step S53, the string NO. (Data of the antecedent part) Dc is read from the SS table in the membership function table ROM 74 in which the membership function “small string number is small (about the first string and second string)” is stored. The data in this case is the membership function SS shown in FIG.
And is a value in the range of 0 to 1. That is, the data Dc is the string NO. Is the degree of conformity indicated by the rule A with respect to the input value.

Next, at step S54, each data Da,
The minimum value of Db and Dc is determined, and the minimum data D1 is obtained. This process calculates the logical product of the respective data, and corresponds to the so-called “MIN”. This is to obtain a state in which at least all three input values are satisfied by “taking MIN”.

If the data Da is the minimum, step S
Going to 55, D1 = Da is adopted, and when the data Db is the minimum, proceeding to step S56, D1 = Db is adopted, and when the data Dc is the minimum, step S is adopted.
Proceeding to 57, D1 = Dc is adopted. As a result, the smallest one of the data Da, Db, and Dc is adopted as the determination data D1 indicated by the rule A in this routine.

As described above, three input values (fret No., string position C and string No.) at that time are determined for rule A, and the degree of conformity of rule A to the antecedent is obtained. Can be

Next, at step S58, the address pointer i is reset to [0], and at step S59, it is determined whether or not the address pointer i is equal to the end address. If it is not the end address, the data (consequent part data) E is read from the AL table in the membership function table ROM 74 in which the membership function "strengthen the attack rate" is stored, using i as the address in the subsequent step S60. The data in this case is shown in FIG.
In the membership function AL shown in FIG. Data E is stored in register E of the same name.

Next, in step S61, data D1 and data E are compared. If D1 ≧ E, the flow advances to step S62 to write data E to the P1 memory in the working RAM 72 using i as an address. On the other hand, D1 <E
In step S63, the process proceeds to step S63 to write the data D1 into the P1 memory in the working RAM 72 using i as the address. As described above, the process of cutting the membership function AL by using the smaller value is performed.

In this processing, a so-called head truncation method for cutting the membership function AL, which is the consequent part, according to the degree of conformity of the antecedent part is performed.

Next, in step S64, the address pointer i is incremented by [1], and thereafter, the flow returns to step S59. Then, the same processing is repeated until the address pointer i becomes equal to the end address in step S59. When the address pointer i becomes equal to the end address, the process returns to the main program.

By incrementing the address pointer i to the end address, the membership function AL
Will be searched. Thus, the “head truncation method” is executed for all the membership functions AL according to the degree of conformity of the antecedent part.

Thus, the fret NO. , String position C and string NO. The fuzzy rule A is calculated based on the fuzzy rule A. When the fret number is small (low position), the string position is on the saddle side, and the string number is small, the attack rate is increased. The degree of conformity to the processing is calculated. That is, the fuzzy operation amount for increasing the attack rate is obtained as the size of the truncated membership function AL.

Fuzzy Rule B Calculation Process FIGS. 11 and 12 are flowcharts showing a subroutine of the fuzzy rule B calculation process. First, step S71
Fret NO. (Data of the antecedent part) Da is read from the FS table in the membership function table ROM 74 in which the membership function “the fret number is small (ie, low position)” is stored. The data in this case is shown in FIG.
It corresponds to the degree of conformity (grade) of the membership function FS shown in (a), and is a value in the range of 0 to 1.

That is, the data Da is the fret NO. Is the degree of conformity indicated by the rule B with respect to the input value. The data Da is stored in a register Da of the same name in the microcomputer 40.
The same applies to c, Dd, and De.

Similarly, in step S72, the membership function table ROM 74 is set using the string position C as an address.
The data (antecedent part data) Db is read from the PS table in which the membership function "the bullet position is on the saddle side" is stored. The data in this case corresponds to the fitness (grade) of the membership function PS shown in FIG. 4B and is a value in the range of 0 to 1. That is, the data Db is the degree of conformity indicated by the rule B with respect to the input value of the string position C.

Also, in step S73, the string NO. (Data of the antecedent part) Dc is read from the SL table in the membership function table ROM 74 in which the membership function “the string number is large (about the sixth and fifth strings)” is stored. The data in this case is a membership function SL shown in FIG.
And is a value in the range of 0 to 1. That is, the data Dc is the string NO. Is the degree of conformity indicated by the rule B with respect to the input value.

Next, in step S74, each data Da,
The minimum value of Db and Dc is determined, and the minimum data F1 is obtained. This process calculates the logical product of the respective data, and corresponds to the so-called “MIN”. This is to obtain a state in which at least all three input values are satisfied by “taking MIN”.

If the data Da is the minimum, step S
Going to 75, F1 = Da is adopted, and when the data Db is the minimum, proceeding to step S76, F1 = Db is adopted, and when the data Dc is the minimum, step S76 is adopted.
Proceeding to 77, F1 = Dc is adopted. As a result, the smallest one of the data Da, Db, and Dc is adopted as the determination data F1 indicated by the rule B in this routine.

As described above, three input values (fret No., string position C, and string No.) at that point in time are determined for rule B, and the matching of rule B with the first antecedent is performed. Degree is required. Specifically, “Fret NO.
Is small (low position), the string position is on the saddle side, and the string NO. Is larger, the attack rate is set to be slightly higher. Is determined for the first antecedent part.

Next, the process proceeds to a process for obtaining the degree of conformity to the second antecedent part. First, in step S78, the membership function table ROM 7 stores the string position C as an address.
4. Data (antecedent part data) Dd is read from the PL table in which the membership function "the bullet position is on the fret side" is stored. The data in this case corresponds to the fitness (grade) of the membership function PL shown in FIG. 4B, and is a value in the range of 0 to 1. That is, the data Db is the degree of conformity indicated by the rule B with respect to the input value of the string position C.

In step S79, the string NO. Is read from the SS table in the membership function table ROM 74 in which the membership function “small string number is small (about the first and second strings)” is stored. The data in this case is the membership function SS shown in FIG.
And is a value in the range of 0 to 1. That is, the data De contains the string NO. Is the degree of conformity indicated by the rule B with respect to the input value.

Next, in step S80, each data Da,
The minimum value of Dd and De is determined, and the minimum data F2 is obtained. This process calculates the logical product of the respective data, and corresponds to the so-called “MIN”. This is to obtain a state in which at least all three input values are satisfied by “taking MIN”.

If the data Da is the minimum, step S
Proceeding to step 81, F2 = Da is adopted, and when the data Dd is the minimum, proceeding to step S82, F2 = Dd is adopted, and when the data De is the minimum, step S82 is adopted.
Proceeding to 83, F2 = De is adopted. As a result, the smallest one of the data Da, Dd, and De is adopted as the determination data F2 indicated by the rule B in this routine.

As described above, three input values (fret NO., Bullet string position C, and string NO.) At that time are determined for rule B, and the matching of rule B with the second antecedent is performed. Degree is required. Specifically, “Fret NO.
Is small (low position), the string position is on the fret side, and the string NO. When is small, the attack rate is made slightly stronger. Is determined for the second antecedent part.

Next, the process proceeds to a process of obtaining the degree of conformity to the third antecedent part. In step S84, each of the data Da,
The minimum value of Dd and Dc is determined, and the minimum data F3 is obtained. This process calculates the logical product of the respective data, and corresponds to the so-called “MIN”. This is to obtain a state in which at least all three input values are satisfied by “taking MIN”.

If the data Da is the minimum, step S
Proceeding to S85, F3 = Da is adopted, and when the data Dd is the minimum, proceeding to step S86 to adopt F3 = Dd, and when the data Dc is the minimum, step S86 is adopted.
Proceeding to 87, adopt F3 = Dc. As a result, the smallest one of the data Da, Dd, and Dc is adopted as the determination data F3 indicated by the rule B in this routine.

As described above, three input values (fret No., bullet string position C, and string No.) at that point in time are determined for rule B, and the matching of rule B with the third antecedent is performed. Degree is required. Specifically, “Fret NO.
Is small (low position), the string position is on the fret side, and the string NO. Is larger, the attack rate is set to be slightly higher. Is determined for the third antecedent part.

Referring to FIG. 12, the flow advances to step S88 to determine the maximum value of the above-mentioned degree-of-fit data F1-F3 for the antecedent part. Then, the process proceeds to any one of steps S89 to S91 according to the maximum value determination result of the data F1 to F3.

The reason why the maximum value is determined is as follows.
Since it is necessary to take a logical sum (OR logic) in order to integrate the inference results of the three antecedent parts described above, this corresponds to a so-called "MAX" processing.

If the data F1 is the maximum, step S
At step 89, F1 is adopted as the antecedent part data D2 obtained by integrating the inference results of the three antecedent parts described above. When the data F2 is the maximum, F2 is adopted at step S90 as the integrated antecedent part data D2. When the data F3 is the maximum, F3 is adopted as the integrated antecedent data D2 in step S91.

As described above, by obtaining the integrated antecedent data D2, three input values (fret NO., Bullet string position C, and string NO.) At that time are determined for rule B. The degree of conformity to the antecedent of Rule B is determined.

After going through any of the steps S89 to S91, the address pointer i is reset to [0] in a step S92, and it is determined in a step S93 whether or not the address pointer i is equal to the end address. If it is not the end address, i follows in step S94.
Function table ROM7 with address as address
4. Data (consequent part data) E is read from the ALL table in which the membership function "Attack rate is slightly increased" is stored. The data in this case is
This corresponds to the fitness of the membership function ALL shown in FIG. The data E is stored in the microcomputer 4
It is stored in a register of the same name in 0.

Next, in step S95, the integrated antecedent data D2 and data E are compared. If D2 ≧ E, the flow advances to step S96 to write data E to the P2 memory in the working RAM 74 using i as an address. . On the other hand, if D2 <E, the process proceeds to step S97 to write data D2 into the P2 memory in the working RAM 74 using i as an address. in this way,
A process of cutting the membership function ALL by adopting the smaller value is performed.

In this processing, a so-called head truncation method for cutting the membership function ALL, which is the consequent part, according to the fitness of the antecedent part is performed.

Next, in step S98, the address pointer i is incremented by [1], and thereafter, the flow returns to step S93. Then, the same processing is repeated until the address pointer i becomes equal to the end address in step S93. When the address pointer i becomes equal to the end address, the process returns to the main program.

By incrementing the address pointer i to the end address, the membership function AL
All L will be searched. Thus, the “head truncation method” is executed for all the membership functions ALL according to the degree of conformity of the antecedent part.

Thus, the fret NO. , String position C and string NO. The calculation of the fuzzy rule B is performed on the basis of "the fret number is small (low position), the string position is on the saddle side and the string number is large, or the fret number is small (low position). When the string position is on the fret side and the string NO. Is small, or when the fret number is small (low position), the string position is on the fret side and the string N
O. Is larger, the attack rate is set to be slightly higher. Is calculated. That is, a fuzzy operation amount for increasing the attack rate or a little is obtained as the size of the truncated membership function ALL.

Fuzzy Rule C Calculation Process FIGS. 13 and 14 are flowcharts showing a subroutine of the fuzzy rule C calculation process. First, step S10
1 and fret NO. (Address data) is read from the FL table in the membership function table ROM 74 in which the membership function “the fret number is large (ie, high position)” is stored. The data in this case corresponds to the fitness (grade) of the membership function FL shown in FIG. 4A and is a value in the range of 0 to 1.

That is, the data Da is the fret NO. Is the degree of conformity indicated by the rule C with respect to the input value. The data Da is stored in a register Da of the same name in the microcomputer 40.
The same applies to c and Dd.

Similarly, in step S102, the membership function table ROM 7
4. Data (antecedent part data) Db is read from the PS table in which the membership function "the bullet position is on the saddle side" is stored. The data in this case corresponds to the fitness (grade) of the membership function PS shown in FIG. 4B and is a value in the range of 0 to 1. That is, the data Db is the degree of conformity indicated by the rule C with respect to the input value of the string position C.

In step S103, the string NO. "String NO. Is small (about the first string and the second string)" in the membership function table ROM 74 with the address as the address.
Is read from the SS table in which the membership function is stored. The data in this case is the membership function S shown in FIG.
It corresponds to the degree of conformity (grade) of S, and is a value in the range of 0 to 1. That is, the data Dc is the string NO. Is the degree of conformity indicated by the rule C with respect to the input value.

Next, at step S104, each data D
The minimum values of a, Db, and Dc are determined, and the minimum data F1 is obtained. This process calculates the logical product of the respective data, and corresponds to the so-called “MIN”. This is to obtain a state in which at least all three input values are satisfied by “taking MIN”.

If the data Da is the minimum, step S
Proceeding to 105, F1 = Da is adopted, and when the data Db is the minimum, the process proceeds to step S106, where F1 = Db is adopted. Further, when the data Dc is the minimum, the process proceeds to step S107 to set F1 = Dc. adopt. This allows
Among the data Da, Db, and Dc, the smallest one is adopted as the determination data F1 indicated by the rule C in this routine.

As described above, three input values (fret No., string position C, and string No.) at that time are determined for rule C, and the matching of rule C with the first antecedent is performed. Degree is required. Specifically, “Fret NO.
Is large (high position), the string position is on the saddle side, and the string NO. When is small, the attack rate is made slow. Is determined for the first antecedent part.

Next, the processing shifts to the processing for obtaining the degree of conformity to the second antecedent part. First, in step S108, the string NO. Is used as an address, the membership function table ROM7
SL in which the membership function of “the string number is large (around the sixth and fifth strings)” is stored.
Data (antecedent part data) Dd is read from the table.
The data in this case corresponds to the fitness (grade) of the membership function SL shown in FIG. 4C and is a value in the range of 0 to 1. That is, the data Dd is the string NO. Is the degree of conformity indicated by the rule C with respect to the input value.

Next, in step S109, the minimum value of each of the data Da, Db, Dd is determined, and the minimum data F2
Get. This process takes the logical product of each data.
This is equivalent to the so-called “MIN”.
This is to obtain a state in which at least all three input values are satisfied by “taking MIN”.

If the data Da is the minimum, step S
Proceeding to 110, F2 = Da is adopted, if the data Db is the minimum, proceed to step S111 to adopt F2 = Db, and if data Dd is the minimum, proceed to step S112 and F2 = Dd. adopt. This allows
Among the data Da, Db, and Dd, the smallest one is adopted as the determination data F2 indicated by the rule C in this routine.

As described above, three input values (fret No., string position C, and string No.) at that time are determined for rule C, and the matching of rule B with the second antecedent is performed. Degree is required. Specifically, “Fret NO.
Is large (high position), the string position is on the saddle side, and the string NO. When is large, the attack rate is made slow. Is determined for the second antecedent part.

Next, in step S113, the data F1 and the data F2 are compared with each other.
If <F2, the process proceeds to step S114, and the larger value is adopted as data D3. In this case D3 =
F2.

This processing is to calculate the logical sum of each data, which is equivalent to the so-called "MAX". This is to obtain a state of adopting the one having a greater influence on two input values by “taking MAX”. Thereby, of the data F1 and F1,
The larger one is adopted as the determination indicated by rule C in this routine. On the other hand, when F1 ≧ F2, the process proceeds to step S115, and D3 = F1 is adopted.

As described above, three input values (fret NO., Bullet string position C, and string NO.) At that time are determined for rule C, and the degree of conformity of rule C to two antecedent parts is determined. Is required.

Step S114 or step S11
After passing through any of the steps of step 5, step S11
At 6, the address pointer i is reset to [0].

Referring to FIG. 14, step S117 follows.
It is determined whether or not the address pointer i is equal to the end address. If it is not the end address, the following step S11
8 with i as the address and membership function table R
The data (consequent part data) E is read from the ALS table in which the membership function “Make the attack rate slow” of the OM 74 is stored. The data in this case corresponds to the degree of fitness of the membership function ALS shown in FIG. The data E is stored in a register of the same name in the microcomputer 40, for example.

Next, in step S119, the integrated antecedent part data D3 and data E are compared. If D3 ≧ E, the flow advances to step S120 to write data E to the P3 memory in the working RAM 74 using i as an address. . On the other hand, if D3 <E, the process proceeds to step S121 to write data D3 into the P3 memory in the working RAM 74 using i as an address. Thus, the smaller value is adopted and the membership function AL
A process of cutting S is performed.

In this process, a so-called head truncation method for cutting the membership function ALS, which is the consequent part, according to the fitness of the antecedent part is performed.

Next, in step S122, the address pointer i is incremented by [1], and thereafter, the flow returns to step S117. Then, the same processing is repeated until the address pointer i becomes equal to the end address in step S117. When the address pointer i becomes equal to the end address, the process returns to the main program.

The membership function AL is incremented by incrementing the address pointer i to the end address.
All S will be searched. As a result, the “head truncation method” is executed for all of the membership functions ALS according to the degree of conformity of the antecedent part.

Thus, the fret NO. , String position C and string NO. The calculation of the fuzzy rule C is performed on the basis of "the fret number is large (high position), the string position is on the saddle side and the string number is small, or the fret number is large (high position). If the string position is on the saddle side and the string number is large, the attack rate is reduced. " In other words, the fuzzy operation amount that makes the attack rate slow is the membership function AL that has been truncated.
It is obtained as the size of S.

Fuzzy Rule D Calculation Process FIGS. 15 and 16 are flowcharts showing a subroutine of the fuzzy rule D calculation process. First, step S13
1 and fret NO. (Address data) is read from the FL table in the membership function table ROM 74 in which the membership function “the fret number is large (ie, high position)” is stored. The data in this case corresponds to the fitness (grade) of the membership function FL shown in FIG. 4A and is a value in the range of 0 to 1.

That is, the data Da is the fret NO. Is the degree of conformity indicated by the rule D with respect to the input value. The data Da is stored in a register Da of the same name in the microcomputer 40.
The same applies to c, Dd, and De.

Similarly, in step S132, the membership function table ROM 7
4. Data (antecedent part data) Db is read from the PL table in which the membership function "the bullet position is on the fret side" is stored. The data in this case corresponds to the fitness (grade) of the membership function PL shown in FIG. 4B, and is a value in the range of 0 to 1. That is, the data Db is the degree of conformity indicated by the rule D with respect to the input value of the string position C.

In step S133, the string NO. "String NO. Is small (about the first string and the second string)" in the membership function table ROM 74 with the address as the address.
Is read from the SS table in which the membership function is stored. The data in this case is the membership function S shown in FIG.
It corresponds to the degree of conformity (grade) of S, and is a value in the range of 0 to 1. That is, the data Dc is the string NO. Is the degree of conformity indicated by the rule D with respect to the input value.

Next, at step S134, each data D
The minimum values of a, Db, and Dc are determined, and the minimum data F1 is obtained. This process calculates the logical product of the respective data, and corresponds to the so-called “MIN”. This is to obtain a state in which at least all three input values are satisfied by “taking MIN”.

If the data Da is the minimum, step S
The process proceeds to step 135 to adopt F1 = Da, and if the data Db is the minimum, the process proceeds to step S136 to employ F1 = Db. If the data Dc is the minimum, the process proceeds to step S137 to set F1 = Dc. adopt. This allows
The smallest of the data Da, Db, and Dc is adopted as the determination data F1 indicated by the rule D in this routine.

Thus, three input values (fret NO., Bullet string position C, and string NO.) At that time are determined for rule D, and the matching of rule D with the first antecedent part is performed. Degree is required. Specifically, “Fret NO.
Is large (high position), the string position is on the fret side, and the string NO. Is small, the attack rate is made extremely slow. Is determined for the first antecedent part.

Next, the processing shifts to the processing for obtaining the degree of conformity to the second antecedent part. First, in step S138, the string position C
With membership function table ROM as address
The data (antecedent part data) Dd is read from the PL table in which the membership function “the string position is on the fret side” in 74 is stored. The data in this case corresponds to the fitness (grade) of the membership function PL shown in FIG. 4B, and is a value in the range of 0 to 1. That is, the data Db is the degree of conformity indicated by the rule B with respect to the input value of the string position D.

In step S139, the string NO. Is used as an address, and “the string number is large (around the sixth and fifth strings)” in the membership function table ROM 74.
Is read from the SL table in which the membership function is stored. The data in this case is the membership function S shown in FIG.
It corresponds to the degree of conformity (grade) of L and is a value in the range of 0 to 1. That is, the data De contains the string NO. Is the degree of conformity indicated by the rule D with respect to the input value.

Next, at step S140, each data D
The minimum values of a, Dd, and De are determined, and the minimum data F2 is obtained. This process calculates the logical product of the respective data, and corresponds to the so-called “MIN”. This is to obtain a state in which at least all three input values are satisfied by “taking MIN”.

If the data Da is the minimum, step S
In step 141, F2 = Da is adopted. When the data Dd is the minimum, the process proceeds to step S142, and F2 = Dd is employed. When the data De is minimum, the process proceeds to step S143, in which F2 = De is determined. adopt. This allows
Among the data Da, Dd, De, the smallest one is adopted as the determination data F2 indicated by the rule D in this routine.

As described above, three input values (fret No., bullet string position C, and string No.) at that time are determined for rule D, and the matching of rule D with the second antecedent part is performed. Degree is required. Specifically, “Fret NO.
Is large (high position), the string position is on the saddle side, and the string NO. Is large, the attack rate is made extremely slow. Is determined for the second antecedent part.

Then, in step S144, the data F1 and the data F2 are compared with each other.
If ≤F2, the process proceeds to step S145, and the larger value is adopted as data D4. In this case D4 =
F2.

This processing is to calculate the logical sum of each data, which is equivalent to the so-called “MAX”. This is to obtain a state of adopting the one having a greater influence on two input values by “taking MAX”. Thereby, of the data F1 and F2,
The larger one is adopted as the determination indicated by rule D in this routine. On the other hand, when F1> F2, the process proceeds to step S146, and D4 = F1 is adopted.

As described above, three input values (fret NO., String position C, and string NO.) At that time are determined for rule D, and the degree of conformity of rule D to two antecedents is determined. Is required.

Step S145 or step S14
After going through any one of the steps of FIG.
In step S147, the address pointer i is reset to [0]. Next, in a step S148, it is determined whether or not the address pointer i is equal to the end address. If it is not the end address, data (consequent part data) E is stored in the membership function table ROM 74 from the AS table in the membership function table ROM 74 in which the membership function "Reduce the attack rate is extremely slow" is stored in step S149. read out. The data in this case is shown in FIG.
This corresponds to the fitness of the membership function AS shown in FIG. The data E is stored in a register of the same name in the microcomputer 40, for example.

Then, in step S150, the integrated antecedent part data D4 described above is compared with data E. If D4 ≧ E, the flow advances to step S151 to write data E into the P4 memory in the working RAM 74 using i as an address. . On the other hand, if D4 <E, the process proceeds to step S152 to write data D4 into the P4 memory in the working RAM 74 using i as an address. Thus, the smaller value is adopted and the membership function AS
Is performed.

In this processing, a so-called head truncation method is performed in which the membership function AS, which is the consequent part, is cut in accordance with the fitness of the antecedent part.

Next, in step S153, the address pointer i is incremented by [1], and thereafter, the flow returns to step S148. Then, the same processing is repeated until the address pointer i becomes equal to the end address in step S148. When the address pointer i becomes equal to the end address, the process returns to the main program.

The membership function AS is incremented by incrementing the address pointer i to the end address.
Will be searched. As a result, the “head truncation method” is executed for all of the membership functions AS in accordance with the fitness of the antecedent part.

Thus, the fret NO. , String position C and string NO. The calculation of the fuzzy rule D is performed based on the following equation: "If the fret number is large (high position), the string position is on the fret side and the string number is small, or the fret number is large (high position). If the string position is on the saddle side and the string number is large, the attack rate is made extremely slow. " That is, a fuzzy operation amount that makes the attack rate extremely slow is obtained as the magnitude of the truncated membership function AS.

[0193] The maximum value A processing 17 is a flow chart showing a subroutine of a maximum value A calculation process. First, in step S161, the address pointer i is reset to [0], and in step S162, the working RAM 7 is set using the address pointer i as an address.
The data D1 is read from the P1 memory contained in No. 2.

The data D1 is the result of the calculation for the fuzzy rule A. Specifically, "the fret number is small (low position), the string position is on the saddle side,
And the string NO. If is small, increase the attack rate. "Corresponds to the membership function AL calculated by calculating the degree of conformity to the processing.

In step S163, the data D2 is read from the P2 memory in the working RAM 72 using the address pointer i as an address.

The data D2 is a calculation result for the fuzzy rule B. Specifically, the data D2 indicates that the fret number is small (low position), the string position is on the saddle side,
And the string NO. Is large, or the fret NO. Is small (low position), the string position is on the fret side, and the string NO. Is small, or fret N
O. Is small (low position), the string position is on the fret side, and the string NO. Is larger, the attack rate is set to be slightly higher. "Corresponds to the membership function ALL truncated by calculating how much the process fits.

In step S164, the data D3 is read from the P3 memory in the working RAM 72 using the address pointer i as an address.

The data D3 is a calculation result of the fuzzy rule C. Specifically, the data D3 indicates that the fret number is large (high position), the string position is on the saddle side,
And the string NO. Is small, or the fret NO. Is large (high position), the string position is on the saddle side, and the string NO. When is large, the attack rate is made slow. Membership function ALS calculated by calculating the degree of conformity to the process "
Corresponding to

Next, in a step S165, similarly, the data D4 is read from the P4 memory in the working RAM 72 using the address pointer i as an address.

The data D4 is the result of the calculation for the fuzzy rule D. Specifically, when the fret number is large (high position), the string position is on the fret side, and the string number is small, Or fret N
O. Is large (high position), the string position is on the saddle side, and the string NO. Is large, the attack rate is made extremely slow. "Corresponds to the membership function AS which is calculated to what extent it is suitable for the processing and truncated.

Next, in step S166, the maximum value among the data D1 to D4 is determined. Then, the process proceeds to step S167, step S168, step S169, and step S170, respectively, according to the maximum value determination result of the data D1 to D4.

The reason why the maximum value is determined is as follows.
Since it is necessary to take a logical sum (OR logic) in order to integrate the inference results for each of the fuzzy rules A to D, it corresponds to a so-called "MAX" process.

Specifically, for example, if any of the rules from the fuzzy rule A to the fuzzy rule D affects the blowing performance at a certain moment,
It answers the need to reflect it in the determination of pitch change, and this is the basic concept of the "OR logic" relationship. In the set theory, this concept is “sum”, and in the present embodiment, as long as any one of them has an effect, the above processing is executed as an object of consideration.

In step S167, data D1 is written to the P9 memory in the working RAM 72 using the address pointer i as an address. This allows
The current address pointer i (in the first routine, i =
Part of the truncated membership function AL that is the result of the fuzzy rule A specified by the value of 0) is P
9 is written to the memory.

As will be described later, the address pointer i
Is incremented to the end address, all of the truncated membership functions AL are searched and written to the P9 memory.

In step S168, the address pointer i
P in working RAM 72 with
9 Write data D2 to memory. As a result, a part of the truncated membership function ALL that is the result of the calculation of the fuzzy rule B specified by the value of the current address pointer i is written to the P9 memory.

Similarly, in step S169, data D3 is written to the P9 memory in the working RAM 72 using the address pointer i as an address. As a result, a part of the truncated membership function ALS, which is the calculation result of the fuzzy rule C specified by the value of the current address pointer i, is written to the P9 memory.

Similarly, in step S170, data D4 is written to the 95 memory in the working RAM 72 using the address pointer i as an address. As a result, a part of the truncated membership function AS, which is the result of the calculation of the fuzzy rule D specified by the value of the current address pointer i, is written to the P9 memory.

Next, in step S171, it is determined whether or not the address pointer i is equal to the end address. If it is not the end address, the address pointer i is incremented by [1] in the following step S172, and thereafter, the process returns to step S162. The above process is repeated until the address pointer i becomes equal to the end address in step S171. When the address pointer i becomes equal to the end address, the process returns to the main program.

By incrementing the address pointer i to the end address, each of the truncated membership functions AL, ALL, ALS, and AS are all searched and written to the P9 memory as data D1 to D4.

Thus, each of the fuzzy rules A to D
The logical sum is taken to integrate the inference results for each (“MAX
Is performed), and the membership functions AL, ALL, ALS, and AS that have been truncated are OR-combined. That is, the MAX synthesis process superimposes the inference results for each of the fuzzy rules A to D to generate a synthesized output.

[0212] centroid calculation A process 18 is a flowchart showing a subroutine of the center of gravity calculation A process. First, in step S181, the address pointer i is reset to [0], and the area register area and the center of gravity register haf are similarly reset to [0]. The area register area stores data of the area required for calculating the center of gravity, and the center of gravity register haf stores a center of gravity value haf required for calculating the center of gravity.

Next, in step S182, the data D9 is read from the P9 memory in the working RAM 72 using the address pointer i as an address. Data D9
Is a part of the OR combination of the membership functions AL, ALL, ALS, and AS of the consequent part of the consequent part truncated by the above-described maximum value calculation processing (a part specified by i as an address). In other words, This is a part of synthesized output data (hereinafter, referred to as an OR synthesis function) in which the inference results for each of the fuzzy rules A to D are superimposed by the MAX synthesis processing.

This time, i (i = 0 in the first routine)
Means that a part of the combined output data has been read from the P9 memory by using As will be described later, by incrementing the address pointer i to the end address, all the combined output data is searched and read.

That is, in the following step S183, the integrated value of the area area is calculated by adding the value of the area register area by the composite output data D9 read this time.

Next, in step S184, it is determined whether or not the address pointer i is equal to the end address. If it is not the end address, the address pointer i is incremented by [1] in the subsequent step S185, and thereafter, the process returns to step S182. The above process is repeated until the address pointer i becomes equal to the end address in step S184. When the address pointer i becomes equal to the end address, the process goes to step S186.

By incrementing the address pointer i to the end address, the OR of each of the membership functions AL, ALL, ALS, and AS of the consequent part of the truncated part is obtained.
All the combined output data is searched and read. In other words, the data set obtained by the maximum value calculation processing is made to correspond to a two-dimensional figure, and the area of this data set portion is searched, read, and calculated by integral calculation. Thereafter, a value at which the integral value corresponding to the area of the data set portion becomes １ ／ is calculated as the center of gravity value.

In order to calculate such a barycenter value, the process first proceeds to step S186, and the value of １ ／ of the integrated area area obtained in step S183 is stored as a barycenter value haf in the barycenter register haf. Next, step S187
Resets both j and the area integration area bal corresponding to the horizontal axis of the OR composite function of the membership functions AL, ALL, ALS, and AS to [0].

The OR synthesis function is a function when the data set obtained by the maximum value calculation processing is made to correspond to a two-dimensional figure. The horizontal axis corresponds to j, and the integrated area of the area is b.
It corresponds to al. The values of j and bal are stored in, for example, a register having the same name.

Then, in step S188, data D9 corresponding to the OR combination function is read from the P9 memory in the working RAM 72 using the horizontal axis j as an address.
In a succeeding step S189, the value of the area integration area bal is added by the data D9 read this time, and the area integration area bal is integrated.

Next, the area integration area bal integrated in step S190 is compared with the center of gravity value haf, and if the area integration area bal is smaller than the center of gravity value haf, the process proceeds to step S190.
After proceeding to 192 to increment the horizontal axis j, the process returns to step S188. Then, in step S190, the above processing is repeated until the area integration area bal becomes equal to or larger than the barycenter value haf. When bal ≧ haf, the value of the horizontal axis j at that time is determined to be the barycenter point, and the process goes to step S191.

In step S191, the bit on the horizontal axis j of the OR combination function is obtained as the attack rate in this routine (that is, the above-mentioned attack rate A level).

In this way, the center of gravity of the OR synthesis function corresponding to the two-dimensional figure is taken, and the defuzzification process (defuzzifier process) whose value is representative of the logical sum is performed. Is calculated as a single definite value, that is, in this case, an attack rate.

Next, fuzzy inference processing for calculating a release rate will be described. Fuzzy Rule E Calculation Process FIG. 19 is a flowchart showing a subroutine of the fuzzy rule E calculation process. First, in step S201, the fret NO. (Address data) is read from the FL table in the membership function table ROM 74 in which the membership function “the fret number is large (ie, high position)” is stored. The data in this case is shown in FIG.
It corresponds to the degree of conformity (grade) of the membership function FL shown in FIG.

That is, the data Da is the fret NO. Is the degree of conformity indicated by the rule E with respect to the input value. The data Da is stored in a register Da of the same name in the microcomputer 40, which stores data Db and Dc to be described later.
The same applies to.

Similarly, in step S202, the membership function table ROM 7
4. Data (antecedent part data) Db is read from the PS table in which the membership function "the bullet position is on the saddle side" is stored. The data in this case corresponds to the fitness (grade) of the membership function PS shown in FIG. 4B and is a value in the range of 0 to 1. That is, the data Db is the degree of conformity indicated by the rule E with respect to the input value of the string position C.

In step S203, the string NO. Is used as an address, and “the string number is large (around the sixth and fifth strings)” in the membership function table ROM 74.
Is read from the SL table in which the membership function is stored. The data in this case is the membership function S shown in FIG.
It corresponds to the degree of conformity (grade) of L and is a value in the range of 0 to 1. That is, the data Dc is the string NO. Is the degree of conformity indicated by the rule E with respect to the input value.

Next, at step S204, each data D
The minimum values of a, Db, and Dc are determined, and the minimum data D5 is obtained. This process calculates the logical product of the respective data, and corresponds to the so-called “MIN”. This is to obtain a state in which at least all three input values are satisfied by “taking MIN”.

If the data Da is the minimum, step S
Proceeding to 205, D5 = Da is adopted, and when the data Db is the minimum, the procedure proceeds to step S206, where D5 = Db is adopted. Further, when the data Dc is the minimum, the procedure proceeds to step S207, and D5 = Dc is adopted. adopt. This allows
Among the data Da, Db, and Dc, the smallest one is adopted as the determination data D5 indicated by the rule E in this routine.

As described above, three input values (fret No., bullet string position C, and string No.) at that time are determined for rule E, and the degree of conformity of rule E to the antecedent is obtained. Can be

Next, in step S208, the address pointer i is reset to [0], and in step S209, it is determined whether or not the address pointer i is equal to the end address. If it is not the end address, in the following step S210, the membership function table ROM is set with i as the address.
The data (consequent part data) E is read from the RL table in which the membership function of “Raise the release rate” is stored at 74. The data in this case is shown in FIG.
This corresponds to the fitness of the membership function RL shown in FIG. Data E is stored in register E of the same name.

Next, data D5 and data E are compared in step S211. If D5 ≧ E, step S212 is performed.
To write data E into the P5 memory in the working RAM 72 using i as an address. On the other hand, D5
If <E, the process proceeds to step S213 to write the data D5 into the P5 memory in the working RAM 72 using i as the address. As described above, the process of cutting the membership function RL by using the smaller value is performed.

In this processing, a so-called head truncation method for cutting the membership function RL, which is the consequent part, according to the fitness of the antecedent part is performed.

Next, in step S214, the address pointer i is incremented by [1], and thereafter, the flow returns to step S209. Then, the same processing is repeated until the address pointer i becomes equal to the end address in step S209, and when the address pointer i becomes equal to the end address, the process returns to the main program.

By incrementing the address pointer i to the end address, the membership function RL
Will be searched. As a result, the “head truncation method” is executed for all the membership functions RL according to the degree of matching of the antecedent part.

Thus, the fret NO. , String position C and string NO. The fuzzy rule E is calculated on the basis of the following formula: "If the fret number is large (high position), the string position is on the saddle side, and the string number is large, the release rate is increased." The degree of conformity to the processing is calculated. That is, the fuzzy operation amount for increasing the release rate is obtained as the size of the truncated membership function RL.

Fuzzy Rule F Calculation Process FIGS. 20 and 21 are flowcharts showing a subroutine of the fuzzy rule F calculation process. First, step S22
1 and fret NO. (Address data) is read from the FL table in the membership function table ROM 74 in which the membership function “the fret number is large (ie, high position)” is stored. The data in this case corresponds to the fitness (grade) of the membership function FL shown in FIG. 4A and is a value in the range of 0 to 1.

That is, the data Da is the fret NO. Is the degree of conformity indicated by the rule F with respect to the input value. The data Da is stored in a register Da of the same name in the microcomputer 40.
The same applies to c, Dd, De, and Df.

Similarly, in step S222, the membership function table ROM 7
4. Data (antecedent part data) Db is read from the PS table in which the membership function “the string position is on the saddle side” is stored. The data in this case corresponds to the fitness (grade) of the membership function PS shown in FIG. 4B and is a value in the range of 0 to 1. That is, the data Db is the degree of conformity indicated by the rule F with respect to the input value of the string position C.

In step S223, the string NO. "String NO. Is small (about the first string and the second string)" in the membership function table ROM 74 with the address as the address.
Is read from the SS table in which the membership function is stored. The data in this case is the membership function S shown in FIG.
It corresponds to the degree of conformity (grade) of S, and is a value in the range of 0 to 1. That is, the data Dc is the string NO. Is the degree of conformity indicated by the rule F with respect to the input value.

Next, at step S224, each data D
The minimum values of a, Db, and Dc are determined, and the minimum data F1 is obtained. This process calculates the logical product of the respective data, and corresponds to the so-called “MIN”. This is to obtain a state in which at least all three input values are satisfied by “taking MIN”.

If the data Da is the minimum, step S
Proceeding to 225, F1 = Da is adopted, and when the data Db is the minimum, the procedure proceeds to step S226, and F1 = Db is employed. When the data Dc is the minimum, the procedure proceeds to step S227, and F1 = Dc is adopted. adopt. This allows
Among the data Da, Db, and Dc, the smallest one is adopted as the determination data F1 indicated by the rule F in this routine.

As described above, three input values (fret No., bullet string position C, and string No.) at that time are determined for rule F, and the rule F is applied to the first antecedent part. Degree is required. Specifically, “Fret NO.
Is large (high position), the string position is on the saddle side, and the string NO. If is small, increase the release rate slightly. Is determined for the first antecedent part.

Next, the processing shifts to the processing for obtaining the degree of conformity to the second antecedent part. First, in step S228, the string position C
With membership function table ROM as address
The data (antecedent part data) Dd is read from the PL table in which the membership function “the string position is on the fret side” in 74 is stored. The data in this case corresponds to the fitness (grade) of the membership function PL shown in FIG. 4B, and is a value in the range of 0 to 1. That is, the data Db is the degree of conformity indicated by the rule F with respect to the input value of the string position C.

In step S229, the string NO. Is used as an address, and “the string number is large (around the sixth and fifth strings)” in the membership function table ROM 74.
Is read from the SL table in which the membership function is stored. The data in this case is the membership function S shown in FIG.
It corresponds to the degree of conformity (grade) of L and is a value in the range of 0 to 1. That is, the data De contains the string NO. Is the degree of conformity indicated by the rule F with respect to the input value.

Next, at step S230, each data D
The minimum values of a, Dd, and De are determined, and the minimum data F2 is obtained. This process calculates the logical product of the respective data, and corresponds to the so-called “MIN”. This is to obtain a state in which at least all three input values are satisfied by “taking MIN”.

If the data Da is the minimum, step S
In step S231, F2 = Da is adopted. When the data Dd is the minimum, the process proceeds to step S232, and F2 = Dd is employed. When the data De is minimum, the process proceeds to step S233 in which F2 = De is determined. adopt. This allows
Among the data Da, Dd, and De, the smallest one is adopted as the determination data F2 indicated by the rule F in this routine.

As described above, three input values (fret No., string position C, and string No.) at that time are determined for rule F, and the rule F is applied to the second antecedent part. Degree is required. Specifically, “Fret NO.
Is large (high position), the string position is on the fret side, and the string NO. If is large, make the release rate a little faster. Is determined for the second antecedent part.

Next, the processing shifts to the processing for obtaining the degree of conformity to the third antecedent part. In step S234, the fret NO. Is used as an address, the membership function table ROM7
4 from the FS table in which the membership function “Fret No. is small (ie, low-high position)” is stored.
Is read. The data in this case corresponds to the fitness (grade) of the membership function FS shown in FIG.
It is a value in the range of 0 to 1.

That is, the data Df is the fret NO. Is the degree of conformity indicated by the rule F with respect to the input value. The data Df is stored in a register Da of the same name in the microcomputer 40.

Moving to FIG. 21, step S235 follows.
To determine the minimum value of each of the data Df, Db, and Dc to obtain the minimum data F3. This process calculates the logical product of the respective data, and corresponds to the so-called “MIN”. This is done by "taking MIN"
This is to obtain a state in which at least all three input values are satisfied.

If the data Df is the minimum, step S
Proceeding to 236, F3 = Df is adopted, and when the data Db is the minimum, the process proceeds to step S237 to adopt F3 = Db, and when the data Dc is the minimum, the process proceeds to step S238 to set F3 = Dc. adopt. This allows
Among the data Df, Db, Dc, the smallest one is adopted as the determination data F3 indicated by the rule F in this routine.

As described above, three input values (fret No., string position C, and string No.) at that time are determined for rule F, and the rule F is applied to the third antecedent part. Degree is required. Specifically, “Fret NO.
Is small (low position), the string position is on the saddle side, and the string NO. If is small, increase the release rate slightly. Is determined for the third antecedent part.

Then, the flow advances to step S239 to determine the maximum value of the above-mentioned conformity data F1 to F3 for the antecedent part. Then, according to the maximum value discrimination result of the data F1 to F3, steps S240 to S242 are respectively performed.
Go to any of

The reason why the maximum value is determined is as follows.
Since it is necessary to take a logical sum (OR logic) in order to integrate the inference results of the three antecedent parts described above, this corresponds to a so-called "MAX" processing.

If the data F1 is the maximum, step S
At step 240, F1 is adopted as the antecedent part data D6 obtained by integrating the inference results of the three antecedent parts described above, and when the data F2 is the maximum, at step S241, the integrated antecedent part data D6
When the data F3 is the maximum, F2 is used as the integrated antecedent part data D6 in step S242.
3 is adopted.

As described above, by obtaining the integrated antecedent part data D6, three input values (fret NO., Bullet string position C, and string NO.) At that time are determined for rule F. The degree of conformity to the antecedent part of rule F is determined.

After going through any one of steps S240 to S242, the address pointer i is reset to [0] in step S243, and the process proceeds to step S2.
At 44, it is determined whether or not the address pointer i is equal to the end address. If it is not the end address, the following step S
In step 245, data (consequent part data) E is read from the RLL table in the membership function table ROM 74 in which the membership function of "slightly increase the release rate" is stored, using i as an address.

The data in this case corresponds to the fitness of the membership function RLL shown in FIG. Data E
Is stored in a register of the same name in the microcomputer 40, for example.

Next, in step S246, the integrated antecedent part data D6 and data E are compared. If D6 ≧ E, the flow advances to step S247 to store the data E in the P6 memory in the working RAM 74 using i as an address. Write. On the other hand, if D6 <E, the process proceeds to step S248 to write data D6 into the P6 memory in the working RAM 74 using i as an address. Thus, the smaller value is adopted and the membership function R
Processing for cutting LL is performed.

In this processing, a so-called head truncation method for cutting the membership function RLL, which is the consequent part, according to the fitness of the antecedent part is performed.

Next, in step S249, the address pointer i is incremented by [1], and thereafter, the flow returns to step S244. Then, the same processing is repeated until the address pointer i becomes equal to the end address in step S244. When the address pointer i becomes equal to the end address, the process returns to the main program.

By incrementing the address pointer i to the end address, the membership function RL
All L will be searched. As a result, a “head truncation method” is executed for all the membership functions RLL according to the degree of matching of the antecedent part.

Thus, the fret NO. , String position C and string NO. The calculation of the fuzzy rule F is performed based on the following equation: "The fret number is large (high position), the string position is on the saddle side and the string number is small, or the fret number is large (high position). , The string position is on the fret side and the string NO. Is large, or the fret number is small (low position), the string position is on the saddle side, and the string No.
If is small, increase the release rate slightly. "
Is calculated to what degree the process is suitable. That is, the fuzzy operation amount of slightly increasing the release rate is reduced by the truncated membership function RLL.
It is calculated as the size of

Fuzzy Rule G Calculation Process FIGS. 22 and 23 are flowcharts showing a subroutine of the fuzzy rule G calculation process. First, step S25
1 and fret NO. (Address data) is read from the FL table in the membership function table ROM 74 in which the membership function “the fret number is large (ie, high position)” is stored. The data in this case corresponds to the fitness (grade) of the membership function FL shown in FIG. 4A and is a value in the range of 0 to 1.

That is, the data Da is the fret NO. Is the degree of conformity indicated by the rule G with respect to the input value. The data Da is stored in a register Da of the same name in the microcomputer 40.
The same applies to c, Dd, De, and Df.

Similarly, in step S252, the membership function table ROM 7
4. Data (antecedent part data) Db is read from the PL table in which the membership function "the bullet position is on the fret side" is stored. The data in this case corresponds to the fitness (grade) of the membership function PL shown in FIG. 4B, and is a value in the range of 0 to 1. That is, the data Db is the degree of conformity indicated by the rule G with respect to the input value of the string position C.

In step S253, the string NO. "String NO. Is small (about the first string and the second string)" in the membership function table ROM 74 with the address as the address.
Is read from the SS table in which the membership function is stored. The data in this case is the membership function S shown in FIG.
It corresponds to the degree of conformity (grade) of S, and is a value in the range of 0 to 1. That is, the data Dc is the string NO. Is the degree of conformity indicated by the rule G with respect to the input value.

Next, at step S254, each data D
The minimum values of a, Db, and Dc are determined, and the minimum data F1 is obtained. This process calculates the logical product of the respective data, and corresponds to the so-called “MIN”. This is to obtain a state in which at least all three input values are satisfied by “taking MIN”.

If data Da is the minimum, step S
In step S256, the process proceeds to step S256 to adopt F1 = Db. When the data Dc is minimum, the process proceeds to step S257 in which F1 = Dc. adopt. This allows
Among the data Da, Db, and Dc, the smallest one is adopted as the determination data F1 indicated by the rule G in this routine.

As described above, three input values (fret NO., String position C, and string NO.) At that time are determined for rule F, and the matching of rule G with the first antecedent is performed. Degree is required. Specifically, “Fret NO.
Is large (high position), the string position is on the fret side, and the string NO. If is small, the release rate should be slow. Is determined for the first antecedent part.

Next, the processing shifts to the processing for obtaining the degree of conformity to the second antecedent part. First, in step S258, fret N
O. , The membership function table R
Data (antecedent part data) from the FS table in which the membership function “Fret No. is small (ie, low position)” in OM74 is stored.
Read Dd. The data in this case corresponds to the fitness (grade) of the membership function FS shown in FIG.

That is, the data Dd is the fret NO. Is the degree of conformity indicated by the rule G with respect to the input value. The data Dd is stored in a register Da of the same name in the microcomputer 40.

Next, in step S259, the string NO. Is used as an address, the membership function table ROM 74
, The membership function that “the string number is large (around the sixth and fifth strings)” is stored.
Data (antecedent part data) De is read from the table.
The data in this case corresponds to the fitness (grade) of the membership function SL shown in FIG. 4C and is a value in the range of 0 to 1. That is, the data De contains the string NO. Is the degree of conformity indicated by the rule G with respect to the input value.

Next, in step S260, the minimum value of each of the data Db, Dd, De is determined, and the minimum data F2
Get. This process takes the logical product of each data.
This is equivalent to the so-called “MIN”.
This is to obtain a state in which at least all three input values are satisfied by “taking MIN”.

If the data Db is at the minimum, step S
The process proceeds to step 261 to adopt F2 = Db. If the data Dd is minimum, the process proceeds to step S262 to employ F2 = Dd. If the data De is minimum, the process proceeds to step S263 to determine F2 = De. adopt. This allows
The smallest one of the data Db, Dd, and De is adopted as the determination data F2 indicated by the rule G in this routine.

As described above, three input values (fret NO., Bullet string position C, and string NO.) At that time are determined for rule G, and the matching of rule G with the second antecedent is performed. Degree is required. Specifically, “Fret NO.
Is small (low position), the string position is on the fret side, and the string NO. If is large, the release rate should be slow. Is determined for the second antecedent part.

Next, the processing shifts to the processing for obtaining the degree of conformity to the third antecedent part. In step S264, using the string position C as an address, data (antecedent part data) Df is read from the PS table in the membership function table ROM 74 in which the membership function "the string position is on the saddle side" is stored. The data in this case corresponds to the fitness (grade) of the membership function PS shown in FIG. 4B and is a value in the range of 0 to 1. That is,
The data Df corresponds to the rule G for the input value of the string position C.
Is the degree of fit shown.

Referring to FIG. 23, in step S265, the minimum value of each of the data Df, Dd, De is determined, and the minimum data F3 is obtained. This process calculates the logical product of the respective data, and corresponds to the so-called “MIN”. This is to obtain a state in which at least all three input values are satisfied by “taking MIN”.

If the data Df is the minimum, step S
Proceeding to 266, F3 = Df is adopted, and when the data Dd is the minimum, the flow proceeds to step S267, and F3 = Dd is adopted. Further, when the data De is the minimum, the flow proceeds to step S268 to set F3 = De. adopt. This allows
Among the data Df, Dd and De, the smallest one is adopted as the determination data F3 indicated by the rule G in this routine.

As described above, three input values (fret NO., Bullet string position C, and string NO.) At that point in time are determined for rule G, and the matching of rule G to the third antecedent is performed. Degree is required. Specifically, “Fret NO.
Is small (low position), the string position is on the saddle side, and the string NO. If is large, the release rate should be slow. Is determined for the third antecedent part.

Then, the flow advances to step S269 to determine the maximum value of the above-mentioned antecedence data F1 to F3. Then, according to the maximum value discrimination results of the data F1 to F3, steps S270 to S272 are respectively performed.
Go to any of

The reason why the maximum value is determined is as follows.
Since it is necessary to take a logical sum (OR logic) in order to integrate the inference results of the three antecedent parts described above, this corresponds to a so-called "MAX" processing.

If the data F1 is the maximum, step S
In step 270, F1 is adopted as the antecedent part data D7 obtained by integrating the inference results of the three antecedent parts described above. If the data F2 is the maximum, the integrated antecedent part data D7 is obtained in step S271.
When the data F3 is the maximum, F2 is set as the integrated antecedent part data D7 in step S272.
3 is adopted.

As described above, by obtaining the integrated antecedent data D7, three input values (fret NO., Bullet position C, and string NO.) At that time are determined for the rule G, The degree of conformity of the rule G to the antecedent is obtained.

After going through any one of steps S270 to S272, the address pointer i is reset to [0] in step S273, and
At 74, it is determined whether or not the address pointer i is equal to the end address. If it is not the end address, the following step S
In step 275, data (consequent part data) E is read from the RLS table in the membership function table ROM 74 in which the membership function "Release the release rate" is stored, using i as the address.

The data in this case corresponds to the fitness of the membership function RLS shown in FIG. Data E
Is stored in a register of the same name in the microcomputer 40, for example.

Then, in step S276, the integrated antecedent part data D7 and data E are compared. If D7 ≧ E, the flow advances to step S277 to store the data E in the P7 memory in the working RAM 74 using i as an address. Write. On the other hand, if D7 <E, the process proceeds to step S278, and data D7 is written to the P7 memory in the working RAM 74 using i as an address. Thus, the smaller value is adopted and the membership function R
A process for cutting the LS is performed.

In this processing, a so-called head truncation method for cutting the membership function RLS, which is the consequent part, according to the fitness of the antecedent part is performed.

Next, in step S279, the address pointer i is incremented by [1], and thereafter, the flow returns to step S274. Then, the same processing is repeated until the address pointer i becomes equal to the end address in step S274. When the address pointer i becomes equal to the end address, the process returns to the main program.

By incrementing the address pointer i to the end address, the membership function RL
All S will be searched. As a result, the “head truncation method” is executed for all of the membership functions RLS according to the degree of matching of the antecedent part.

Thus, the fret NO. , String position C and string NO. The fuzzy rule F is calculated on the basis of the following equation: "If the fret number is large (high position) and the string position is on the fret side and the string number is small, or if the fret number is small (low position) , The string position is on the fret side, and the string NO.
Is large, or the fret NO. Is small (low position), the string position is on the saddle side, and the string N
O. If is large, the release rate should be slow. Is calculated. That is, a fuzzy operation amount for making the release rate gentle is obtained as the size of the truncated membership function RLS.

Fuzzy Rule H Calculation Process FIG. 24 is a flowchart showing a subroutine of the fuzzy rule H calculation process. First, in step S281, the fret NO. (Data of the antecedent part) Da is read from the FS table in the membership function table ROM 74 in which the membership function “the fret number is small (ie, low position)” is stored. The data in this case is shown in FIG.
It corresponds to the degree of conformity (grade) of the membership function FS shown in (a), and is a value in the range of 0 to 1.

That is, the data Da is the fret NO. Is the degree of conformity indicated by the rule H with respect to the input value of. The data Da is stored in a register Da of the same name in the microcomputer 40, which stores data Db and Dc to be described later.
The same applies to.

Similarly, in step S 282, the membership function table ROM 7
4. Data (antecedent part data) Db is read from the PL table in which the membership function "the bullet position is on the fret side" is stored. The data in this case corresponds to the fitness (grade) of the membership function PL shown in FIG. 4B, and is a value in the range of 0 to 1. That is, the data Db is the degree of conformity indicated by the rule H with respect to the input value of the string position C.

In step S283, the string NO. "String NO. Is small (about the first string and the second string)" in the membership function table ROM 74 with the address as the address.
Is read from the SS table in which the membership function is stored. The data in this case is the membership function S shown in FIG.
It corresponds to the degree of conformity (grade) of S, and is a value in the range of 0 to 1. That is, the data Dc is the string NO. Is the degree of conformity indicated by the rule H with respect to the input value of.

Next, at step S284, each data D
The minimum values of a, Db, and Dc are determined, and the minimum data D8 is obtained. This process calculates the logical product of the respective data, and corresponds to the so-called “MIN”. This is to obtain a state in which at least all three input values are satisfied by “taking MIN”.

If the data Da is the minimum, step S
Proceeding to 285, adopt D8 = Da, if the data Db is the minimum, proceed to step S286 and employ D8 = Db, and if the data Dc is the minimum, proceed to step S287 to advance D8 = Dc. adopt. This allows
Among the data Da, Db, and Dc, the smallest one is adopted as the determination data D8 indicated by the rule H in this routine.

As described above, three input values (fret NO., Bullet string position C, and string NO.) At that time are determined for rule H, and the degree of conformity of rule H to the antecedent is obtained. Can be

Then, in step S288, the address pointer i is reset to [0], and in step S289, it is determined whether or not the address pointer i is equal to the end address. If it is not the end address, in the following step S290, the membership function table ROM is set with i as the address.
The data (consequent part data) E is read from the RS table 74 in which a membership function of “make release rate extremely slow” is stored. The data in this case corresponds to the fitness of the membership function RS shown in FIG. Data E is stored in register E of the same name.

Next, data D8 and data E are compared in step S291, and if D8 ≧ E, step S292
And writes the data E into the P8 memory in the working RAM 72 using i as an address. On the other hand, D8
If <E, the process proceeds to step S293 to write data D8 into the P8 memory in the working RAM 72 using i as the address. As described above, the process of cutting the membership function RS by adopting the smaller value is performed.

In this processing, a so-called head truncation method for cutting the membership function RS, which is the consequent part, according to the fitness of the antecedent part is performed.

Then, in step S294, the address pointer i is incremented by [1], and thereafter, the flow returns to step S289. Then, the same processing is repeated until the address pointer i becomes equal to the end address in step S289. When the address pointer i becomes equal to the end address, the process returns to the main program.

By incrementing the address pointer i to the end address, the membership function RS
Will be searched. Thus, the “head truncation method” is executed for all of the membership functions RS in accordance with the fitness of the antecedent part.

Thus, the fret NO. , String position C and string NO. The fuzzy rule H is calculated based on the following equation: "If the fret number is small (low position), the bullet string position is on the fret side, and the string number is small, the release rate is made extremely slow. Is calculated. That is, the fuzzy operation amount that makes the release rate extremely slow is obtained as the magnitude of the truncated membership function RS.

Maximum Value R Calculation Processing FIG. 25 is a flowchart showing a subroutine of maximum value R calculation processing. First, in step S301, the address pointer i is reset to [0], and in step S302, the working RAM 7 is set using the address pointer i as an address.
The data D5 is read from the P5 memory contained in P2.

The data D5 is a calculation result of the fuzzy rule E. Specifically, the data D5 indicates that the fret number is large (high position), the string position is on the saddle side,
And the string NO. If is large, increase the release rate. "Corresponds to the membership function RL which is calculated to what extent it is suitable for the processing and truncated.

In step S303, the data D6 is read from the P6 memory in the working RAM 72 using the address pointer i as an address.

The data D6 is a calculation result for the fuzzy rule F. Specifically, the data D6 indicates that the fret number is large (high position), the string position is on the saddle side,
And the string NO. Is small, or the fret NO. Is large (high position), the string position is on the fret side, and the string NO. Is large, or fret N
O. Is small (low position), the string position is on the saddle side, and the string NO. If is small, increase the release rate slightly. The membership function R is calculated by calculating the degree of conformity to the processing
LL.

In step S304, data D7 is read from the P7 memory in the working RAM 72 using the address pointer i as an address.

[0311] Data D7 is the result of calculation for the fuzzy rule G. Specifically, when the fret number is large (high position), the string position is on the fret side, and the string number is small, Or fret N
O. Is small (low position), the string position is on the fret side, and the string NO. Is large, or the fret NO. Is small (low position), the string position is on the saddle side, and the string NO. If is large, the release rate should be slow. "Corresponds to the membership function RLS which is calculated to what extent it is suitable for processing and truncated.

Next, in a step S305, similarly, the data D8 is read from the P8 memory in the working RAM 72 using the address pointer i as an address.

The data D8 is the result of the calculation for the fuzzy rule H. More specifically, the data D8 is obtained when the fret number is small (low position), the string position is on the fret side, and the string number is small. , Makes the release rate extremely slow. "

Next, in step S306, the maximum value among the data D5 to D8 is determined. Then, the process proceeds to step S307, step S308, step S309, and step S310, respectively, according to the maximum value determination result of the data D5 to D8.

The reason why the maximum value is determined is as follows.
Since it is necessary to take a logical sum (OR logic) in order to integrate the inference results for each of the fuzzy rules E to H, it corresponds to a so-called "MAX" process.

Specifically, for example, if any of the rules from the fuzzy rule E to the fuzzy rule H affects the blowing performance at a certain moment,
It answers the need to reflect it in the determination of pitch change, and this is the basic concept of the "OR logic" relationship. In the set theory, this concept is “sum”, and in the present embodiment, as long as any one of them has an effect, the above processing is executed as an object of consideration.

In step S307, data D5 is written to the P10 memory in the working RAM 72 using the address pointer i as an address. As a result, the current address pointer i (in the first routine, i
= 0), a part of the truncated membership function RL, which is the calculation result of the fuzzy rule E, is written to the P10 memory.

Note that, as described later, the address pointer i
Is incremented to the end address, all the truncated membership functions RL are searched and written to the P10 memory.

In step S308, the address pointer i
P in working RAM 72 with
10 Write data D6 to memory. As a result, a part of the truncated membership function RLL, which is the calculation result of the fuzzy rule F specified by the value of the current address pointer i, is written to the P10 memory.

Similarly, in step S309, data D7 is written to the P10 memory in the working RAM 72 using the address pointer i as an address. As a result, a part of the truncated membership function RLS, which is the calculation result of the fuzzy rule G specified by the value of the current address pointer i, is written to the P10 memory.

Similarly, in step S310, the data D8 is written to the P10 memory in the working RAM 72 using the address pointer i as an address. As a result, a part of the truncated membership function RS, which is the calculation result of the fuzzy rule H specified by the value of the current address pointer i, is written to the P10 memory.

Next, in step S311, it is determined whether or not the address pointer i is equal to the end address. If it is not the end address, the address pointer i is incremented by [1] in the following step S312, and thereafter, the process returns to step S302. Then, the above processing is repeated until the address pointer i becomes equal to the end address in step S311, and when the address pointer i becomes equal to the end address, the process returns to the main program.

By incrementing the address pointer i to the end address, the truncated membership functions RL, RLL, RLS, and RS are all searched and written to the P10 memory as data D5 to D8.

Thus, each of the fuzzy rules E to H
The logical sum is taken to integrate the inference results for each (“MAX
Is performed), and the membership functions RL, RLL, RLS, and RS that have been truncated are OR-combined. That is, the MAX synthesis process superimposes the inference results for each of the fuzzy rules E to H to generate a synthesized output.

[0325] centroid calculation R Process FIG 26 is a flowchart showing a subroutine of centroid calculation R process. First, in step S321, the address pointer i is reset to [0], and the area register area and the center of gravity register haf are similarly reset to [0]. The area register area stores data of the area required for calculating the center of gravity, and the center of gravity register haf stores a center of gravity value haf required for calculating the center of gravity.

Next, in step S322, the data D10 is read from the P10 memory in the working RAM 72 using the address pointer i as an address. The data D10 includes the membership functions RL, RLL, RLS, and RLS of the consequent part truncated by the above-described maximum value calculation processing.
Composed output data obtained by superimposing the inference results for each of the fuzzy rules E to H by the MAX composing process (part of the OR composing output data) (Expressed as a function).

In this case, i (i = 0 in the first routine)
Means that a part of the combined output data is read from the P10 memory using the address as an address. As will be described later, by incrementing the address pointer i to the end address, all the combined output data is searched and read.

That is, in the following step S323, the integrated value of the area area is calculated by adding the value of the area register area by the composite output data D10 read this time.

Next, in a step S324, it is determined whether or not the address pointer i is equal to the end address. If it is not the end address, the address pointer i is incremented by [1] in the following step S325, and thereafter, the process returns to step S322. Then, the above processing is repeated until the address pointer i becomes equal to the end address in step S324. When the address pointer i becomes equal to the end address, the process goes to step S326.

By incrementing the address pointer i to the end address, the OR of each of the membership functions RL, RLL, RLS and RS of the truncated consequent part is obtained.
All the combined output data is searched and read. In other words, the data set obtained by the maximum value calculation processing is made to correspond to a two-dimensional figure, and the area of this data set portion is searched, read, and calculated by integral calculation. Thereafter, a value at which the integral value corresponding to the area of the data set portion becomes １ ／ is calculated as the center of gravity value.

In order to calculate such a barycenter value, the process first proceeds to step S326, and the value of １ ／ of the integrated area area obtained in step S323 is stored in barycenter register haf as barycenter value haf. Next, step S327
Resets both j and the area integration area bal corresponding to the horizontal axis of the OR function of the membership functions RL, RLL, RLS, and RS to [0].

The OR synthesis function is a function when the data set obtained by the maximum value calculation processing is made to correspond to a two-dimensional figure. The horizontal axis corresponds to j, and the integrated area of the area is b.
It corresponds to al. The values of j and bal are stored in, for example, a register having the same name.

Then, in step S328, data D10 corresponding to the OR combination function is read from the P10 memory in the working RAM 72 using the horizontal axis j as an address. In a succeeding step S329, the value of the area integration area bal is added by the data D10 read this time, and the area integration area bal is integrated.

Next, the area integrated area bal integrated in step S330 is compared with the center of gravity value haf, and if the area integrated area bal is smaller than the center of gravity value haf, step S330 is performed.
Proceeding to 332, the horizontal axis j is incremented and then returning to step S328. Then, in step S330, the above processing is repeated until the area integration area bal becomes equal to or greater than the barycenter value haf. When bal ≧ haf, the value of the horizontal axis j at that time is determined to be the barycenter point, and the process exits to step S331.

In step S331, the bit of the horizontal axis j of the OR combination function is obtained as the release rate in this routine (that is, the release rate R level described above).

In this way, the center of gravity of the OR synthesis function corresponding to the two-dimensional figure is obtained, and the defuzzification process (difuzzifire process) whose value represents OR is performed. Is calculated as one fixed value, that is, a release rate in this case.

As described above, in this embodiment, when at least one of the plurality of strings is struck, the string position (saddle side or fret side), fingering operation position (high position, low position) is determined. The characteristics (in this case, the envelope shape) of the musical tone to be generated are controlled by fuzzy inference using the position or the string number (one of the first to sixth strings) as input parameters. Therefore, according to a delicate combination of a string position, a fingering operation position, and a string number at the time of stringing, the characteristics of the generated musical sound change delicately, and a musical sound characteristic very close to a natural stringed instrument can be obtained.

Further, since the fuzzy inference is used, a huge amount of processing is not required for the arithmetic processing of the envelope control, and the calculation load can be reduced. As a result, the calculation can be easily performed by the existing CPU,
A high-performance electronic musical instrument can be realized at low cost.

In the above embodiment, the envelope shape is changed as the characteristic of the musical tone. However, the musical tone characteristic changed by the fuzzy inference is not limited to the envelope shape, and other elements such as pitch, tone, volume, or the like. The effect (effect) may be changed. Alternatively, these elements may be changed in combination. By doing so, it is possible to electronically obtain a more natural tone characteristic close to a natural stringed instrument.

In the above embodiment, the string position, the fingering operation position, and the string number are used as input parameters for fuzzy inference. By using them as parameters, fuzzy inference may be performed according to a combination of these. Alternatively, fuzzy inference may be performed by appropriately combining the three parameters. In short, the inference input may be determined according to the type of the musical instrument to be used.

In the above embodiment, the fuzzy inference is realized by software using the microcomputer 40 and the membership function table ROM 74.
For example, it may be realized by hardware using a fuzzy chip.

The value of the membership function used in the fuzzy inference may be appropriately set according to the type of the target musical instrument. In this way, various flavors of control for not only the envelope but also other tone characteristics can be achieved. Further, the shape of the membership function is not limited to a trapezoid or a triangle, but may be another shape (for example, a bell shape).

Further, in the above embodiment, the present invention is applied to an electronic musical instrument which generates a musical tone of a natural string instrument called a guitar. However, the present invention is not limited to such an application to an electric guitar.
The present invention can be widely applied to an electronic musical instrument that generates a musical tone of a type in which a string is picked up.

[0344]

According to the first or sixth aspect of the present invention, a string position (a saddle side or a fret side) and a fingering operation position (a high position or a low position).
Since the release rate of the envelope added to the generated musical tone is controlled by the fuzzy inference using the input parameter as the input parameter, the characteristic of the generated musical tone is determined according to a delicate combination of the string position and the fingering operation position at the time of stringing. Can be delicately changed, and a tone characteristic very similar to a natural stringed instrument can be obtained.

Further, since the fuzzy inference is used, enormous processing is not required for the arithmetic processing of the musical tone characteristic control, and the arithmetic load can be reduced. As a result, calculations can be easily performed by the existing CPU, and a high-performance electronic musical instrument can be realized at low cost.

According to the third or sixth aspect of the present invention, the string number (for example, any of the first to sixth strings) is one of the input parameters in addition to the string position and the fingering operation position. In addition to fuzzy inference,
By finely changing the tone characteristics more finely, it is possible to obtain a tone similar to a natural stringed instrument.

[Brief description of the drawings]

FIG. 1 is a diagram showing an overall configuration of an embodiment of an electronic musical instrument according to the present invention.

FIG. 2 is a diagram showing an example of embedding the fret switch of the embodiment.

FIG. 3 is a block diagram showing a configuration of an electronic circuit of the electric guitar of the embodiment.

FIG. 4 is a diagram showing a membership function of tone characteristic control according to the embodiment.

FIG. 5 is a main program showing a part of a process of tone generation control of the embodiment.

FIG. 6 is a main program showing a part of processing of tone generation control of the embodiment.

FIG. 7 is a waveform chart showing an example of envelope control of the embodiment.

FIG. 8 is a flowchart showing a subroutine of an attack rate calculation process of the embodiment.

FIG. 9 is a flowchart showing a subroutine of a release rate calculation process of the embodiment.

FIG. 10 is a flowchart showing a subroutine of a fuzzy rule A calculation process of the embodiment.

FIG. 11 is a flowchart showing a part of a subroutine of a fuzzy rule B calculation process of the embodiment.

FIG. 12 is a flowchart illustrating a part of a subroutine of a fuzzy rule B calculation process of the embodiment.

FIG. 13 is a flowchart showing a part of a subroutine of a fuzzy rule C calculation process of the embodiment.

FIG. 14 is a flowchart showing a part of a subroutine of a fuzzy rule C calculation process of the embodiment.

FIG. 15 is a flowchart showing a part of a subroutine of a fuzzy rule D calculation process of the embodiment.

FIG. 16 is a flowchart illustrating a part of a subroutine of a fuzzy rule D calculation process of the embodiment.

FIG. 17 is a flowchart showing a subroutine of a maximum value A calculation process of the embodiment.

FIG. 18 is a flowchart showing a subroutine of a center-of-gravity calculation A process of the embodiment.

FIG. 19 is a flowchart showing a subroutine of a fuzzy rule E calculation process of the embodiment.

FIG. 20 is a flowchart showing a part of a subroutine of a fuzzy rule F calculation process of the embodiment.

FIG. 21 is a flowchart showing a part of a subroutine of a fuzzy rule F calculation process of the embodiment.

FIG. 22 is a flowchart illustrating a part of a subroutine of a fuzzy rule G calculation process of the embodiment.

FIG. 23 is a flowchart showing a part of a subroutine of a fuzzy rule G calculation process of the embodiment.

FIG. 24 is a flowchart showing a part of a subroutine of a fuzzy rule H calculation process of the embodiment.

FIG. 25 is a flowchart showing a subroutine of a maximum value R calculation process of the embodiment.

FIG. 26 is a flowchart showing a subroutine of a center-of-gravity calculation R process of the embodiment.

FIG. 27 is a waveform chart showing an example of the envelope control of the embodiment.

FIG. 28 is a waveform chart showing an example of envelope control of the embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Electric guitar 2 Body 3 Neck 4 Head 5 Fret 6 Finger board 7 Strings 10, 20 Pickup 30 Fret switch 31 Key matrix circuit 32 Key scan circuit 33 Fingering operation position detecting means 40 Microcomputer (musical sound generation instruction means) 43 , 53 Envelope detecting circuit 60 String operation detecting means 61 String operating position detecting means 62 Operating string detecting means 63 Analog processing circuit 71 ROM 72 Working RAM 74 Membership function table ROM 80 PCM sound source (musical sound generating means) 81 Envelope generating circuit 83 Sounding circuit 100 Fuzzy inference means

Continuation of the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G10H 1/053 G10H 1/00

Claims (7)

(57) [Claims] 1. A finger board, a plurality of strings stretched along the finger board, string operation detecting means for detecting that a string has been struck, and a string detected by the string operation detecting means. String operating position detecting means for detecting the position of the string with respect to the finger string, fingering operating position detecting means for detecting the operating position of the fingering with respect to the finger board, and input parameters for the string position and the fingering operating position of the string. Fuzzy inference means for performing a fuzzy inference operation according to a predetermined fuzzy rule, and a release rate according to the operation result of the fuzzy inference means
An electronic musical instrument comprising: a musical sound generation instructing means for instructing the generation of a musical sound to which an envelope attenuated by the above is added . 2. The electronic musical instrument according to claim 1, wherein the fingering operation position detecting means detects a fret operation position of the finger board as the fingering operation position. (3)Fingerboard, With multiple strings stretched along the fingerboard, String operation detecting means for detecting that a string has been struck; For the string whose bullet is detected by the string operation detecting means.
A string operating position detecting means for detecting a string position; Operating position of fingering on the fingerboard
Fingering operation position detecting means for detecting the position, Operating string detecting means for detecting the number of the string that has been struck, String position, fingering operation position and string number
According to a predetermined fuzzy rule as an input parameter
A fuzzy inference means for performing a fuzzy inference operation; Release rate according to the operation result of the fuzzy inference means
Indicates the occurrence of a musical tone with an envelope attenuated by
Music sound generation instruction means, Characterized by having Electronic musical instruments. 4. The fingering operation position detecting means.
Is the fingerboard as the fingering operation position.
Detecting the fret operation position of the door.
Item 3. The electronic musical instrument according to Item 3. 5. The electronic musical instrument further comprises the musical tone generating finger.
Tone generation for generating a tone that is instructed by the indicating means
4. An electronic device according to claim 1, further comprising:
Child instrument. 6.Duplexed along the fingerboard
Detect if any of a number of strings is hit
Steps and A step for detecting the position of the string with respect to the detected string
And Operating position of fingering on the fingerboard
Detecting the location; Detected string position and fingering operation position
According to a predetermined fuzzy rule as an input parameter
Performing a fuzzy inference operation; Decrease at the release rate according to the calculation result of the fuzzy inference
A switch that instructs the generation of musical tones with a decaying envelope
Tep, A musical tone control method used for an electronic musical instrument. 7.Duplexed along the fingerboard
Detect if any of a number of strings is hit
Steps and A step for detecting the position of the string with respect to the detected string
And Operating position of fingering on the fingerboard
Detecting the location; Detecting the number of the string that was struck; The detected string position of string, fingering operation position and
Using the string number as an input parameter,
Performing a fuzzy inference operation according to Decrease at the release rate according to the calculation result of the fuzzy inference
A switch that instructs the generation of musical tones with a decaying envelope
Tep, A musical tone control method used for an electronic musical instrument.