JPH10228276A - Device and method for correcting feedback path characteristic of automatically controlled musical instrument - Google Patents

Abstract

PROBLEM TO BE SOLVED: To correct characteristics of an element in a feedback loop by absorbing solid-body difference as to the automatically controlled musical instrument which acts feedback control for the driving of a movable part such as a key. SOLUTION: In the feedback loop consisting of a solenoid 4 which drives a key by operating a plunger at 2 speed corresponding to an exciting current, a speed sensor 7 which detects the plunger speed of the solenoid 4, an adder 14 which outputs the deviation of the plunger speed fed back from the speed sensor from 21 command, and a solenoid driving means 16 which excites the solenoid 4 with a current having a value corresponding to the deviation outputted from the adder 14, an amplifier 14 is provided which can have an offset and a gain set for an input signal; and the offset and gain of the amplifier 15 are so set that solid-body differences of an offset and a gain presumed to be already present for the plunger speed is absorbed.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus and a method for correcting a feedback path characteristic of an automatic musical instrument employing feedback control for driving a movable portion such as a key.

[0002]

2. Description of the Related Art In an automatic piano, which is an automatic control instrument, when a solenoid is excited to drive a corresponding key, a hammer is rotated and a string is struck accordingly. The strength of the hammer striking corresponds to the key driving speed, and the key driving speed corresponds to the power supply current to the solenoid, so by controlling the amount of power supplied to the solenoid, the strength of the string striking,
That is, the magnitude of the generated musical tone can be controlled. Note that the control of the power supply current to the solenoid is generally performed by pulse width modulation (PWM).

[0003]

By the way, the above-mentioned conventional automatic piano generally supplies a current corresponding to string striking strength data to a solenoid at a timing corresponding to string striking time data in performance information. The key drive corresponding to each sound is performed. That is, only the key press start timing and the key press force at the start of key press are controlled, and the trajectory of the key from the start of key press to the key release (the trajectory in this case is a change in position over time) is accurately determined Is not necessarily reproduced. In addition, since open loop control is adopted,
Even if the key trajectory does not match the expected trajectory, there is no way to know this, and there is also the disadvantage that it cannot be corrected, of course.

[0004] In view of the above circumstances, the applicant of the present application has
To solve the above drawbacks, we proposed an automatic piano that includes data corresponding to the key trajectory from the start of key depression to the key release in the performance information and adopts feedback control for driving the key. (See the specification and drawings attached to the application of Japanese Patent Application No. 5-344241). According to the automatic piano, accurate reproduction of the key trajectory from the start of key depression to the key release can be expected.

Further, according to the above-mentioned automatic piano, when the feedback control is properly performed, there is an advantage that the key can be driven in an appropriate trajectory by absorbing individual differences in characteristics of the automatic piano. is there. However, if there are individual differences in the characteristics of the elements in the feedback loop and the feedback control itself is not performed properly, it will be difficult to sufficiently absorb the individual differences in the characteristics of the automatic piano, and the key trajectory will not be accurate. Reproduction cannot be expected. The individual difference may already exist at the time of shipment from the factory, or may occur due to aging.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and in an automatic control instrument employing feedback control for driving a movable portion such as a key, the characteristics of elements in a feedback loop are absorbed by absorbing individual differences. It is an object of the present invention to provide an apparatus and a method for correcting a feedback path characteristic of an automatically controlled musical instrument, which can be corrected.

[0007]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, a feedback path characteristic correcting apparatus for an automatically controlled musical instrument according to the present invention comprises a movable part, a speed sensor for detecting an operating speed of the movable part, A driving means for inputting a feedback signal indicating an operation speed detected by the speed sensor, and driving the movable portion such that a target value based on the performance data matches a value represented by the input feedback signal. A feedback path characteristic correction device for an automatic control musical instrument, which is applied to a control instrument and corrects a characteristic of a path of the feedback signal including the speed sensor, wherein the characteristic is inserted into the path and has a characteristic according to an internal parameter. It is characterized by having a correction means.

Further, the parameters inside the characteristic correcting means may be changeable from the outside. In addition, a circuit comprising the movable part, the driving means and a path including the characteristic correcting means may be used as a model. Measuring means for actually measuring the values of the elements of the transfer function formula obtained by the conversion, and calculating and calculating the parameters based on the transfer function formula, the target value, the actually measured values by the measuring means, and the characteristics of the path. Adjustment means for setting a parameter to the characteristic correction means may be provided. Note that the element whose actual value is measured by the measuring means is a different element from the target value. Or,
When the target value is 0, a value of a feedback signal between the speed sensor and the characteristic correction unit may be measured, and a parameter inside the characteristic correction unit may be set based on the measured value.

In addition, according to the present invention, there is provided a method for correcting a feedback path characteristic of an automatically controlled musical instrument, comprising: a movable section; a speed sensor for detecting an operation speed of the movable section; A driving means for driving the movable portion so that a target value based on the performance data coincides with a value represented by the input feedback signal, wherein the feedback control includes the speed sensor. A method of correcting a feedback path characteristic of an automatic musical instrument for correcting a characteristic of a signal path, wherein a circuit including the movable section, the driving means, and the path is modeled, and a transfer function expression obtained by the modeling is obtained. A different value is set for at least one of the elements, the movable section is operated for each setting, and the values of the at least one element and another element are measured, and the transmission function is set. Obtains the value of the unknown elements based on the measured value by the measuring means with a set value of the Equation least one element is characterized by correcting the characteristics of the path based on the values obtained.

[0010]

Embodiments of the present invention will be described below with reference to the drawings. A. Configuration of Embodiment A-1. 1. Overall Configuration FIG. 1 is a diagram showing a configuration of a main part of an automatic piano equipped with an apparatus for correcting a feedback path characteristic of an automatic musical instrument according to an embodiment of the present invention. Although only one key (key 1) is shown in this drawing, 88 keys are actually arranged in the left-right direction of the player (front and back in the drawing). The key 1 is rotatably mounted on a shelf plate 2 via a balance pin (not shown). When the rear end of the key 1 is moved upward about the balance pin (when the key is pressed), the movement is not shown. The action is transmitted to the damper and the hammer via the action, whereby the damper separates from the string, and the hammer rotates to strike (strike) the string stretched corresponding to the key 1. When the force applied to the key 1 disappears, the key 1 rotates around the balance pin due to the weight of an action (not shown) and returns to the rest position.

A key driving unit 3 for driving the key 1 is provided in a portion of the shelf 2 below the rear end of the key 1 so as to face the lower surface of the rear end of the key 1. The key driving unit 3 includes a solenoid 4 including a substantially cylindrical bobbin and a coil wound around the body.
And a plunger 5 fitted in the bobbin portion,
Further, the plunger 5 has a plunger head 6 whose upper end is in contact with the lower rear end of the key 1, and pushes up the plunger 5 with a force corresponding to a current (excitation current) supplied to the coil of the solenoid 4.

A speed sensor (feedback sensor) 7 for detecting the operating speed of the plunger 5 is provided below the key driving unit 3.
Is provided. The speed sensor 7 has a small-diameter bobbin formed coaxially with the bobbin forming the solenoid 4, a coil 8 wound around the body of the bobbin, and a plunger whose upper end is inserted into the bobbin. 5 has a thin cylindrical magnet rod 9 coaxially fixed to the lower end thereof. When the magnet rod 9 moves upward integrally with the plunger 5, an induction generated in the coil 8 according to the speed of the operation. The voltage is output as the plunger speed (feedback signal). Note that when the player plays by himself, the plunger 5 and the magnet bar 9 do not operate.

Reference numeral 10 denotes a pre-reproduction processing unit which generates data indicating the trajectory of a key corresponding to each sound in performance data supplied from a recording medium or a real-time communication device. Reference numeral 11 denotes a motion controller that creates data indicating the movement of the plunger 5 based on the trajectory data created by the reproduction preprocessing unit 10. Reference numeral 12 denotes a servo controller, which excites the solenoid 4 in accordance with data supplied from the motion controller 11. Control the current.

The servo controller 12 also detects a plunger speed detected by the speed sensor 9 and a hammer sensor (measuring means) (not shown) for detecting the operating speed (stringing speed) of the hammer near the stringing position. The entered string striking speed is input. Then, at the time of reproduction, the servo controller 12 controls the excitation current of the solenoid 4 so that the data supplied from the motion controller 11 and the plunger speed from the feedback sensor 13 match each other, while comparing them. In addition, since a well-known hammer sensor is applicable, detailed description is omitted here.

A-2. Configuration of Servo Controller 12 FIG. 2 is a block diagram for explaining the configuration of the servo controller 12. In this figure, parts corresponding to the respective parts in FIG. 1 are denoted by the same reference numerals, and description thereof will be omitted.

In the servo controller 12, reference numeral 13 denotes a target value generating means for generating a target value (voltage value) of the plunger speed according to data supplied from the motion controller 11, and 14 denotes a target value generating means 13. An adder that inputs a target value from a non-inverting input terminal, adds a value input from an inverting input terminal to the target value, and outputs the added value.
An amplifier (characteristic correcting means) 15 described later is connected to the inverting input terminal.
, The plunger speed from the speed sensor 7 is input. That is, the adder 14 outputs a deviation between the target value and the plunger speed. Reference numeral 16 denotes a solenoid driving unit that excites the solenoid 4 with a current having a value corresponding to the deviation output from the adder 14. As a driving method, for example,
A drive method based on pulse width modulation is adopted. As is clear from the above, the servo controller 12
It is configured to feedback control the plunger speed.

The above-mentioned amplifier 15 can set an offset and a gain (amplification degree) with respect to the input signal, add a preset offset to the input signal, and apply a predetermined gain to the result. Is amplified. The amplifier 15 needs to absorb and normalize the individual difference between the speed sensor 7 and the servo controller 12 in order to absorb individual differences in the characteristics of the automatic piano and perform appropriate feedback control. Elements of the difference include the ratio (gain) between the actual plunger speed (string striking speed) and the output value of the speed sensor 7, and the offset (voltage value) added to the output value of the speed sensor inside the servo controller 12. The gain and the offset of the amplifier 15 are hereinafter referred to as a sensor gain correction value and a sensor offset correction value, respectively.

Reference numeral 17 denotes an adjusting means, which controls various preset parameters (described later) and the target value generating means 1;
3, a sensor gain correction value and a sensor offset correction value of the amplifier 15 based on a plunger speed supplied via the amplifier 15 and a stringing speed detected by a hammer sensor (not shown). Is calculated,
Set to amplifier 15. Although the adjusting means 17 may be provided detachably from the automatic piano, it is assumed here that the adjusting means 17 is fixedly provided on the automatic piano. This is the same for the hammer sensor.

B. Next, a method of calculating the sensor gain correction value and the sensor offset correction value by the adjusting unit 17 will be described. The calculation and setting of each correction value is usually performed at the time of factory shipment or maintenance.

B-1. Modeling To calculate the sensor gain correction value and the sensor offset correction value, it is necessary to first model the target feedback control system. FIG. 3 is a block diagram showing a model of the feedback control system shown in FIG. 2, where G ₁ is a transfer function of the solenoid driving means 16,
G ₂ is the control target serving movable part transfer function (solenoid 4, keys, action, including hammer), f [N] to the external force (specifically applied to the solenoid 4, due to the weight of the key, action, etc. (Force applied to the solenoid 4). In the present model, the disturbance is only the external force f, and the time function f (t) of the external force f takes a steady value F. Also, the gain of the speed sensor 7 is a constant value k _3, the servo controller 12
The time function d (t) of the offset d is a constant value D, and the time function c (t) of the sensor offset correction value c is a constant value C
The sensor gain correction value takes a constant value k _4. Also, r
[M / s] is a target value generated by the target value generating means 13, y [m / s] is a control amount, and e is a deviation between the target value r and the control amount y. In the present embodiment, the control amount y
Corresponds to the stringing speed detected by a hammer sensor (not shown), and all variables are s unless otherwise specified.
Is a function of

In the model shown in FIG. 3, a deviation e between the target value r and the control amount y is obtained at an addition point 18 and input to a block 19 of the transfer function G ₁ . The deviation between the output and the external force f of the block 19 is obtained by summing point 20 and input to block 21 of the transfer function G _2. Output control variable y next to block 21, are input to block 22, is where k ₃ times. Then, at block 22, k
The difference between ₃ times the value and the offset d is obtained at the summing point 23, the sum of the difference and the sensor offset correction value c is calculated by summing point 24. The sum obtained at the addition point 24 is multiplied by k _{4 in} a block 25 and supplied to the addition point 18 as a feedback value.

[0022] As apparent from FIG. 3, if appropriate sensor offset correction value c and sensor gain correction value k ₄ is can e = r-y, and the a _{e = e r + e f +} e d to. However,

(Equation 1) It is.

Further, since the steady-state error E of the time function e (t) of the error e is the limit value of e (t) at t → ∞, the final value theorem of the Laplace transform gives

(Equation 2) Thus, the steady-state error E can be obtained.

Here, the steady-state deviation E when the time function r (t) of the target value r is a constant value R, that is, in the case of step input (constant speed control) will be considered. If r (t) = R, then r = R / s. Further, as described above, since f = R / s dc = (D−C) / s, from the equation (1),

(Equation 3) Becomes

However, each term E _r , E in the right side of the equation (3)
_f, which corresponds to each term in the right side of E _d, respectively (2). In addition,

(Equation 4) It is. As apparent from the above description, as long as k ₁ is finite, steady-state error E _r with respect to the target value, the steady state error due to the external force E _f, and steady-state deviation E _d by the offset present.

B-2. Principle As is clear from the principle of feedback control, more accurate feedback control can be realized by making the steady-state error E close to zero. Therefore, in other words, the problem of the present invention is to “set the steady-state error E to 0”. Considering this problem, the control side (adjusting means 1
Since only k ₄ and C can be set in 7), E is set by setting only appropriate k ₄ and C.
Is important.

In an automatic piano, since the R differs for each tone, it is necessary to set the steady-state deviation E to 0 for different Rs. However, as is apparent from the equation (3), it is necessary to set E to 0. enough conditions, and _{E r = 0,, E f} -E d =
0. To the E _r = 0 is (2) and (3) As apparent from the _{equation, 1 + k 1 k 2 k} 3 k 4 -k 1 k 2 = 0 ... (4) must be established. Since the equation (4) there is a term including k _4, E by selecting the k ₄ properly
It can be seen that _r = 0 can be set. Also, E _f −
From Eqs. (2) and (3), it is clear that k ₂ F−k ₁ k ₂ k ₄ (DC) = 0 (5) must be satisfied in order to set E _d = 0. It is. Since there is a term including k ₄ and C in the equation (5), it is understood that E _f −E _d = 0 can be obtained by appropriately selecting k ₄ and C. Equation (5) indicates that k ₄ is _equal to E _r
It is shown that E _f −E _d = 0 can be established by appropriately selecting C even if it is fixed to establish = 0.

B-3. k ₄ and C calculated strategy will be explained calculation policies appropriate k ₄ and C. In the present embodiment, basically, to set the parameters within the expected range, the on the piano is played operate in, and calculates the appropriate k ₄ and C on the basis of the measurement value obtained at this time. At this time, it is desirable that the measured value is easy to measure, and in the present embodiment, the control amount (string striking speed) is measured by the above-mentioned hammer sensor (not shown) to obtain the measured value.

Here, the above policy will be described from another viewpoint. Using the target value r as a step input,

(Equation 5) Then, the following equation is obtained from the equation (2).

[0030]

(Equation 6) Since R is known in the equation (6), a steady-state error E is obtained by measuring Y. Once the steady-state error E is obtained, appropriate k ₄ and C can be calculated by calculating k ₄ and C so that the steady-state error E becomes zero.

The parameters required to set the steady-state error E to 0 are only C when k ₄ is already set and fixed, and only k ₄ when C is already set and fixed. both when it is not set and fixed is considered are three C and k _4, specific calculation processes are different. Therefore, the calculation method will be described below for each case. However, in the following description, E _r =
The optimum sensor gain correction value to be 0 is k ₄ ^′ , E _f −E _d
The optimal sensor offset correction value where = 0 is assumed to be C ^′ .

B-4. Case in which k ₄ is already set and fixed [1] Method of operating with two different sensor offset correction values First, appropriate different sensor offset correction values C ₀ and C ₁ within an expected range are determined. The reproduction operation is performed by setting, and the control amount (string striking speed) at that time is measured by a hammer sensor, and the measured values are defined as Y ₀ and Y ₁ . In addition, as C ₀ and C ₁ , arbitrary values can be selected within the expected range. For example, the highest value and the lowest value within the expected range may be used. It is assumed that parameters other than the sensor offset correction value are not changed in the two reproduction operations for C ₀ and C ₁ .

By the way, C and Y in the equation (6) are replaced with the above C
Substituting ₀ and Y ₀ , C ₁ and Y ₁ gives

(Equation 7) Is obtained.

Then, from these equations,

(Equation 8) Is obtained.

By the way, the optimum sensor offset correction value C ^', from the equation (5), ^C' becomes _{= D-F / k 1 k} 4. That is, even if the values of k ₁ , D, and F are individually unknown, if the value of DF / k ₁ k _{4 is known} , the values can be changed to C ^′.
It can be. Therefore, DF / k ₁ k ₄
Can be derived from the known value by modifying the equation so that is represented by a known value. Specifically, the following equation is obtained by substituting the equation (8) into the equation (7).

(Equation 9)

By modifying the above equation, DF / k ₁ k ₄ can be represented by a known value. That is, the following equation is obtained.

(Equation 10) This method is to obtain C ′ by substituting a known value into the above equation.

[2] Method of Operating with Two Different Target Values First, different suitable target values R ₀ , within the expected range.
R ₁ is set to perform a reproducing operation, and the control amount (string striking speed) at that time is measured by a hammer sensor, and the measured value is Y
₀ and Y ₁ . In addition, as R ₀ and R ₁ , arbitrary values can be selected within the expected range. For example, the highest value and the lowest value within the expected range may be used. It is assumed that parameters other than the target value are not changed in a total of two reproduction operations for R ₀ and R ₁ .

By the way, in C and Y in the above-mentioned equation (6),
R above₀ And Y₀ , R₁ And Y ₁ Substituting
An expression is obtained. Y₀ = A₁・ R₀+ A₀ ... (9a) Y₁ = A₁・ R₁+ A₀ … (9b) However,

[Equation 11] It is. Above (9a), from (9b) _{type, A 0 = (Y 0 R} 1 -Y 1 R 0) / (R 1 -R 0) A 1 = (Y 0 -Y 1) / (R 0 -R 1 ) Can be obtained, and A ₀ and A ₁ can be represented by known parameters.

On the other hand, from equations (10a) and (10b), 1
+ K ₁ k ₂ k ₃ k ₄

(Equation 12) Is obtained. Further, by dividing both sides of the equation (11) by A ₁ k ₄ , A ₀ / (A ₁ k ₄ ) = D−F / (k ₁ k ₄ ) −C is obtained. Therefore, C ^′ = D−F / (k ₁ k ₄ ) = A ₀ / (A ₁ k ₄ ) + C As described above, since A ₀ and A ₁ are represented by known parameters, C ′ Can be obtained. This method is to obtain C ′ by substituting a known value or a value obtained from the known value into the above equation.

B-5. Case where C is already set and fixed [1] Method of operating with two different sensor gain correction values First, different appropriate sensor gain correction values k ₄₀ and k ₄₁ within the expected range are set. To perform a reproducing operation, and the control amount (string striking speed) at that time is measured by a hammer sensor, and the measured values are defined as Y ₀ and Y ₁ . In addition, arbitrary values can be selected as k ₄₀ and k ₄₁ within an expected range.
The highest and lowest values within the expected range may be used. In addition, it is assumed that parameters other than the sensor gain correction value are not changed in a total of two reproduction operations for k ₄₀ and k ₄₁ .

By the way, that C has already been set and fixed means that E _f + E _d = 0 has already been established. Therefore, from equation (6), Y is represented by the following equation.

(Equation 13)

Therefore, Y ₀ and Y ₁ are represented by the following equations.

[Equation 14]

From the above equation, if k ₁ k ₂ of the numerator on the right side is eliminated, k ₁ k ₂ k ₃ = (Y ₀ −Y ₁ ) / (Y ₁ k ₄₁ −Y ₀ k ₄₀ ) is obtained. When the equation is substituted into the equation (12), k ₁ k ₂ is represented by the following equation.

(Equation 15)

From this equation and equation (4), the following equation is obtained.

(Equation 16) This method is to obtain k ₄ ^′ by substituting a known value into this equation.

[2] Method of operating with two or more different target values In this method, similar to the method of deriving C ^′ in B-4 [2] above, different methods are used within the expected range. The reproducing operation is performed by setting the desired target values R ₀ and R _1, and the control amount (string striking speed) at that time is measured by a hammer sensor, and the measured values are defined as Y ₀ and Y ₁ . Also, (9a), (9b), (10
a), (10b) and (11) are also used in B-4.
This is the same as the method for deriving C ^′ in [2]. However,
Regarding A ₀ in the equations (9a) and (9b), C may be regarded as almost “0” on the assumption that C is already an optimum value. Meanwhile, A _1, as described above, because that corresponds to the slope of the approximate line can be obtained the actual value by the following equation. A ₁ = (Y ₀ −Y ₁ ) / (R ₀ −R ₁ )

From equation (10a), 1 + k ₁ k ₂ k ₃ k ₄ = k ₁ k ₂ / A _1. Therefore, k ₁ k ₂ = A ₁ + k ₁ k ₂ k ₃ k ₄ A ₁ is there. Therefore, the optimum sensor gain correction value k ₄ ^′ is expressed by the following equation.

[Equation 17] It is. At this time, if A ₁ = (Y ₀ −Y ₁ ) / (R ₀ −R ₁ ) ≒ 1, k ₄ ^′ ≒ k ₄ A ₁ (14) may be used. In this method, k ₄ is converged by repeating the above operation until A ₁ ≒ 1, and k ₄ ^′ is obtained.

[3] For each of two different k ₄ values, 2
Method of operating each time In this method, appropriate values k ₄₀ and k ₄₁ are set within the expected range, and the keyboard is operated twice for each of them, and A ₁ at that time is changed to B-5 [2 determined by the method described in, and a _10, a _11, respectively. Note that parameters other than the target value are not changed during operation with the same sensor gain correction value, and parameters other than the sensor gain correction value are not changed during operation with different sensor gain correction values.

From the equation (13), k ₄ ^′ = (A ₁₀ −1) / (k ₁ k ₂ k ₃ ) + k ₄₀ A ₁₀ k ₄ ^′ = (A ₁₁ −1) / (k ₁ k ₂₎ Since k ₃ ) + k ₄₁ A ₁₁ , k ₁ k ₂ k ₃ = (A ₁₀ −A ₁₁ ) / (k ₄₁ A ₁₁ −k ₄₀ A ₁₀ ). Therefore, the following equation is obtained.

(Equation 18) This method is to obtain k ₄ ^′ by substituting a known value into this equation.

B-6. Case in which k ₄ and C are not determined In this method, appropriate target values and sensor gain correction values (R ₀₀ and k ₄₀ , R ₀₁ and k ₄₀ ,
R ₁₀ and k ₄₁ , R ₁₁ and k ₄₁ ) are set and a reproducing operation is performed, the control amount (string striking speed) at that time is observed, and the values are referred to as Y ₀₀ , Y ₀₁ , Y ₁₀ , and Y ₁₁ . I do. In each reproduction operation, parameters other than the target value and the sensor gain correction value are not changed. Also, R ₀₀ , R ₀₁ , R ₁₀ , R ₁₁
And k ₄₀ , k ₄₀ , k ₄₁ , and k ₄₁ are different from each other.

By substituting the above-mentioned Y ₀₀ , Y ₀₁ , Y ₁₀ , and Y ₁₁ into the equation (6) and taking the difference between them, the following equation is obtained.

[Equation 19]

From the above equations (15) and (16), k
If you clear the ₁ k _2,

(Equation 20) Is obtained, and when this is substituted into the equation (15), k ₁ k ₂ is represented by the following equation.

[0052]

(Equation 21) By the way, _assuming that k ₄ for setting _Er to 0 is k ₄ ^′ , (4)
From the equation, k ₄ ^′ is represented by the following equation. k ₄ ^′ = (k ₁ k ₂ −1) / (k ₁ k ₂ k ₃ ) That is, if (k ₁ k ₂ −1) / (k ₁ k ₂ k ₃ ) is known, k ₄ ^′ is obtained. be able to.

As described above, k ₁ k ₂ and k ₁ k ₂ k ₃ can be represented by known values.
₄ ^' is obtained from known values.

(Equation 22) This method is to obtain k ₄ ^′ by substituting a known value into this equation.

Even if R ₀₀ = R ₁₀ and R ₀₁ = R ₁₁ , there is no practical problem. In this case, R ₀₀ = R
₁₀ ≡R ₀ , R ₀₁ = R ₁₁ ≡R ₁
Equation (17) can be simplified as follows.

(Equation 23)

C: Summary The adjusting means 17 can calculate the optimum sensor offset correction value and sensor gain correction value by any one or a combination of the above-mentioned methods, and can set them in the amplifier 15. That is, each element (actually, a parameter inside the servo controller 12) in the feedback loop can be appropriately set, and more accurate feedback control can be realized. Therefore, the automatic performance piano can accurately reproduce the key trajectory during the reproducing operation.

In an automatic performance piano, a performance operation by a player is detected, the performance operation is recorded as performance data, and the recorded performance data is read out and an automatic performance is performed by driving a solenoid. Will be
In this case, a hammer sensor for detecting the operation speed of the hammer, a key sensor for detecting key operation, and the like are provided to detect a performance operation by the player. Then, when calculating the sensor offset correction value and the sensor gain correction value, the sensor provided for recording the performance operation can be used. Therefore, it is possible to calculate the sensor offset correction value and the sensor gain correction value without adding a new configuration to the existing automatic performance piano.

In the above-described embodiment, an example is shown in which the amplifier 15 capable of changing the sensor offset correction value and the sensor gain correction value is used.
An amplifier that cannot change the sensor offset correction value and / or the sensor gain correction value may be used.
However, in this case, an accurate sensor offset and / or sensor gain is obtained in a configuration excluding the amplifier, and an amplifier having a sensor offset correction value and / or a sensor gain correction value corresponding to the sensor offset and / or the sensor gain is used. It is the procedure of inserting.
In addition, the sensor offset correction value and / or the sensor gain correction value of the amplifier 15 may be directly set by an operator or the like.

Further, the optimum sensor offset correction value and / or sensor gain correction value may be obtained by a method other than the method shown in this embodiment. For example, first, r is set to 0, and the process waits until the key (plunger) becomes stationary, that is, y = 0.
Next, k ₄ = 0, the feedback input to the adding point 18 is cut off, and at this point, the value between the adding points 23 and 24 is measured. Since the measured value here can be regarded as "-d", the sensor offset d can be surely and easily canceled if C = d during the reproduction driving.

In the present embodiment, the case of step input (constant speed control) has been described, but the present invention may be applied to acceleration control. More specifically, a sensor offset correction value and a sensor gain correction value for each expected target value are obtained by any of the above-described methods, and are set in accordance with the target value generated by the target generation unit 13 during the reproducing operation. The sensor offset correction value and the sensor gain correction value are switched.

Further, the number of reproduction operations for obtaining the measured value is not limited to two or four, but may be three or five or more. In this case, for example, after calculating a plurality of C ^′ and / or k ₄ ^′ candidates, C ^′ and / or k ₄ ^′ can be obtained by taking an average.

In this embodiment, an example in which a hammer sensor is used has been described. However, the present invention is not limited to this as long as the control amount can be accurately detected. Alternatively, a sensor for detecting the key operation or a sensor for detecting the operation of the key may be used. However, you should not use a sensor that becomes a component of the feedback loop. Further, although the key is shown as an example of the movable portion, it is needless to say that the pedal may be used.

Further, in the above-described embodiment, an example is shown in which the present invention is applied to an automatic piano. However, any instrument that automatically drives operators such as keys and pedals can be used. The present invention is applicable to natural musical instruments such as harpsichords and pipe organs, and electronic musical instruments that generate musical sounds electronically.

[0063]

As described above, according to the present invention,
Since the characteristic of the path of the feedback signal indicating the operation speed of the movable part such as a key can be set, the characteristic of the path can be set to an optimum characteristic. That is, the characteristics of the feedback loop can be corrected so as to absorb the individual differences between the automatically controlled musical instruments. Further, since the characteristic of the route can be set from the outside, the process of setting the optimum characteristic can be simplified as compared with the case where the setting is not possible. Further, since the characteristic calculation procedure is formulated, semi-automatic or fully automatic processing becomes possible, and the burden on the operator can be reduced.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a main part of an automatic piano equipped with an apparatus for correcting a feedback path characteristic of an automatic musical instrument according to an embodiment of the present invention.

FIG. 2 is a block diagram for explaining a configuration of a servo controller 12 in the same device.

FIG. 3 is a block diagram showing a model of a feedback control system in the apparatus.

[Explanation of symbols]

4 ... solenoid, 7 ... speed sensor, 12 ... servo controller, 13 ... target value generation means, 14 ... adder, 15 ...
Amplifier, 16: Solenoid driving means, 17: Adjusting means

Claims (5)

[Claims] 1. A movable unit, a speed sensor for detecting an operation speed of the movable unit, and a feedback signal indicating an operation speed detected by the speed sensor, and a target value based on performance data and a feedback signal input And a driving unit for driving the movable portion so that the value represented by the following formula (1) is equal to the value of the return path of the automatic control instrument for correcting the characteristic of the path of the feedback signal including the speed sensor. A characteristic correction device, comprising: a characteristic correction unit inserted into the path and having a characteristic according to an internal parameter. 2. A feedback path characteristic correcting device for an automatically controlled musical instrument according to claim 1, wherein parameters inside said characteristic correcting means can be changed from outside. 3. The path includes the characteristic correction unit, a measurement unit that measures a value of an element of a transfer function equation obtained by modeling a circuit including the movable unit, the driving unit, and the path, Adjusting means for calculating the parameter based on a transfer function formula, the target value, an actually measured value by the measuring means, and characteristics of the route, and setting the calculated parameter in the characteristic correcting means; 3. The feedback path characteristic correction device for an automatically controlled musical instrument according to claim 2, wherein the element whose value is actually measured is a different element from the target value. 4. When the target value is 0, a value of a feedback signal between the speed sensor and the characteristic correction unit is measured, and a parameter inside the characteristic correction unit is set based on the measured value. The feedback path characteristic correction device for an automatically controlled musical instrument according to claim 1, wherein: 5. A movable unit, a speed sensor for detecting an operation speed of the movable unit, and a feedback signal representing an operation speed detected by the speed sensor, and a target value based on performance data and a feedback signal input thereto. And a drive unit for driving the movable unit so that the value is equal to the value represented by the formula (1), wherein the feedback path of the automatic control instrument for correcting the characteristic of the path of the feedback signal including the speed sensor A characteristic correction method, comprising modeling a circuit including the movable section, the driving unit, and the path, and among the elements of a transfer function equation obtained by the modeling,
Different values are set for at least one element, the movable section is operated for each setting, and the at least one element is set.
One element and another element are actually measured, and the value of the unknown element is obtained based on the transfer function formula, the set value of the at least one element, and the actually measured value of the measuring means, and based on the obtained value. A method for correcting a feedback path characteristic of an automatically controlled musical instrument, wherein the characteristic of the path is corrected.