JP3336089B2 - Signal processing device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a signal processing device,
Particularly, the present invention relates to a device for adding an effect to an input signal.

[0002]

2. Description of the Related Art There are various effects added to musical sound signals, such as distortion and overdrive. When this distortion or overdrive effect is obtained by digital signal processing,
A combination of non-linear processing based on clip or table conversion and high-pass filter or low-pass filter processing has been performed.

[0003]

By the way, in electric guitars, the tone of a vacuum tube type guitar amplifier tends to be preferred. In particular, a timbre change when an excessive input is given is characteristic. The timbre change obtained by the combination of the non-linear processing and the filter processing as described above is different from the timbre change when an excessive input is given to this vacuum tube type guitar amplifier, and is not always satisfied by the performer. Was. Similarly, if a clip circuit using a diode is used to obtain an effect similar to the distortion effect by a combination of the above-described processes, the performer will not be satisfied.

[0004]

SUMMARY OF THE INVENTION The present invention has solved the above-mentioned problems, and a first aspect of the present invention is to reduce the amplitude of an input signal.
Non-linear processing means in which the larger the distortion, the greater the distortion
Provided on the input side of the non-linear processing means
Adjustable high-pass filter means and
Depending on the output level, the frequency characteristics of this filter
Filter control means for controlling.

According to a second aspect, the amplitude value of the input signal is large.
Non-linear processing means that generates a large amount of distortion,
The output level of the pattern processing means is detected, and based on this detection result,
To change the bias applied to the input signal.
Ass changing means.

[0006]

Generally, a vacuum tube amplifier has a resistor connected between its grid and a reference potential point. When an input signal is supplied to the grid from the amplifier in the preceding stage, the input signal is supplied via a capacitor in order to prevent a DC component. The resistor and the capacitor constitute a high-pass filter. When an excessive input signal exceeding the bias voltage to the grid is supplied to the grid in such a vacuum tube amplifier, a grid current flows in a portion of the input signal larger than a predetermined level (bias voltage), and the grid resistance of the vacuum tube is reduced. From infinity to a small value, the output signal of the vacuum tube amplifier becomes distorted. Hereinafter, such an operation of the vacuum tube amplifier is referred to as overdrive. Since the grid resistance during such overdrive is connected in parallel to the resistor, the time constant of the high-pass filter is reduced,
The cutoff frequency of the high-pass filter increases.

[0007] In the first aspect of the present invention, for example, which has been made to this state enables the simulation, non-linear
High-pass filter means is provided on the input side of the processing means , and its cutoff frequency is variable. When the output signal of the high-pass filter means, that is, the input signal to the nonlinear processing means exceeds a predetermined value,
In accordance with the difference between the input signal level and the predetermined value, the cutoff frequency of the high-pass filter increases, and for example, it is possible to simulate signal amplification in a vacuum tube amplifier.

In a vacuum tube power amplifier, when overdriving due to excessive input, the bias point fluctuates due to a drop in power supply voltage, and crossover distortion occurs.

[0009] In the second invention, in order to simulate this condition, the greater the level change of the output signal of the non-linear processing means, the bias changing means for deep bias on the input side of the nonlinear processing means is provided is there.

In the first invention, the diode clip is provided.
The loop circuit can also be simulated.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention is shown in FIGS. In this embodiment, a triode vacuum tube type B as shown in FIG.
A class push-pull amplifier is simulated using a DSP (digital signal processor).

In the class B push-pull amplifier shown in FIG. 4, a phase inversion signal obtained by inverting the phase of the signal supplied to the input terminal 6 by the phase inversion stage 4 comprising the first-stage triode vacuum tube 2 is transmitted from the plate of the vacuum tube 2 , A non-inverted phase signal whose phase is not inverted is output from the cathode. Reference numerals 6 and 8 are a cathode resistor and a load resistor of the vacuum tube 2.

The phase inversion signal from the phase inversion stage 4 is passed through a DC blocking capacitor 10 to the triode vacuum tube 1 of the driving stage 11.
2 grids. Similarly, the non-inverted phase signal is
It is supplied to the grid of the triode vacuum tube 16 of the drive stage 11 via the DC blocking capacitor 14. Reference numerals 18 and 20 denote grid resistors of the vacuum tubes 12 and 16, 22, 24 and 23,
5 is the same cathode resistor and bypass capacitor, 2
6, 28 are the same load resistors. The values of the grid resistors 18 and 20 and the cathode resistors 22 and 24 are selected so that the driving stage operates as a class B push-pull amplifier.

In the vacuum tube 12 of the driving stage 11, the grid voltage changes according to the signal input through the capacitor 10, and the change in the grid voltage becomes the change in the plate current, that is, the output signal. This output signal is supplied to the grid of the triode vacuum tube 32 of the output stage 31 via the DC blocking capacitor 30. Similarly, an output signal from the vacuum tube 16 of the driving stage 11 is supplied to the grid of the triode vacuum tube 36 of the output stage 31 via the DC blocking capacitor 34. As a result, the vacuum tube 3
The plate currents 2 and 36 change and are taken out as output signals. 38 and 40 are grid resistors of the vacuum tubes 32 and 36, and 42, 44 and 43 and 45 are the cathode resistors and bypass capacitors. Grid resistor 3
8, 40 and the values of the cathode resistors 42, 44
1 is selected to operate as a class B push-pull amplifier.

The output signals of the vacuum tubes 32 and 36 are input to the primary side of an output transformer 46, and are supplied from the secondary side to a speaker (not shown). Each of the vacuum tubes 2, 12, 16 includes a load resistor 8, 26, 28 and a voltage drop resistor 57.
And a DC voltage of + B is applied through the vacuum tubes 32 and 36
, A + B DC voltage is applied through the primary side of the output transformer 46. The + B DC voltage is obtained by transforming an AC voltage from an AC power supply 48 by a transformer 50 and rectifying the AC voltage by rectifier diodes 52 and 54,
6. Smoothing is performed by a smoothing circuit 62 including capacitors 58 and 60.

In the driving stage 11, the DC blocking capacitor 10 and the grid resistor 18 constitute a high-pass filter. When the signal input from the phase inversion stage 4 is at a low level (normal amplification operation), the vacuum tube 12 Is a large value, the time constant of this high-pass filter is determined by the values of the capacitor 10 and the resistor 18, but when a large amplitude signal exceeding the normal amplification operation is input to the grid, When the grid current flows and the grid resistance is reduced, it is connected in parallel with the grid resistor 18, the time constant is reduced, and the cutoff frequency of the high-pass filter is reduced. Further, as described above, when the time constant changes asymmetrically with respect to the center of the amplitude of the input signal, a DC component is eventually accumulated in the capacitor 10, and the grid bias of the vacuum tube 12 due to the accumulated charge is deepened. appear.

Such a change in the cutoff frequency of the high-pass filter and a change in the bias are caused by the other vacuum tubes 16, 32,
36 also causes the same.

In the output stage 31, when the output signal is overdriven due to excessive input, the power supply voltage drops and the bias becomes deep, and as a result, crossover distortion occurs.

In order to simulate the operation of such a class B push-pull amplifier using a DSP, the DSP
FIG. 1 is a block diagram showing the means for realizing the above. That is, a phase inversion block 4a corresponding to the phase inversion stage 4, a drive block 11a corresponding to the drive stage 11, and an output block 31a corresponding to the output stage 31 are provided.

The phase inversion block 4a branches the input signal into two, supplies one of the branch outputs to the drive block 11a as it is, that is, in a state of non-phase inversion, and supplies the other of the branch outputs to the phase inversion circuit 64. Then, the phase is inverted and supplied to the drive block 11a.

The driving stage 11a includes an Ig block 66 for generating an output signal simulating a grid current of the vacuum tube 16 in response to the phase non-inverted signal, a grid voltage as an input signal, and a voltage of the vacuum tube 16 corresponding to the grid voltage. It has an Ip / Eg block 68 for generating an output signal simulating the plate current (actually, an output signal converted into a voltage via a load resistor). Similarly, for the phase inversion signal, an Ig block 70 for simulating the grid current of the vacuum tube 12 and an Ip / Eg block 72 for simulating the plate current are provided.

The Ig blocks 66 and 70 have the same configuration, and simulate a high-pass filter constituted by the DC blocking capacitors 10 and 14 and the grid resistors 18 and 20 in FIG. These Ig
FIG. 2 shows a detailed configuration of the block.

The phase non-inversion signal or the phase inversion signal x is
The output signal y of the adder 74 is supplied to the adder 74.
Are Ip / E as output signals of the Ig blocks 66 and 70.
g blocks 68 and 72 are supplied. Further, the output signal of the adder 74 is further multiplied by a coefficient K supplied from a control block 78 described later in a multiplier 76, and the multiplied output is supplied to an adder 80. The output signal of the adder 80 is delayed by one sample in a delay circuit 82, and the delayed output is supplied to adders 80 and 74.

The transfer function y / x in this circuit is given by
It is represented by

Y / x = (1−Z ⁻¹ ) / [1− (1−K) Z ⁻¹ ] From this transfer function, the adder 74, the multiplier 76, and the adder of the Ig blocks 66 and 70 are obtained. 80, it can be seen that the delay circuit 82 operates as the high-pass filter 83. And
The larger the value of K, the higher the cutoff frequency.

When the level of the output signal from the Ig blocks 66 and 70 increases (corresponding to an increase in the grid voltage of the vacuum tubes 12 and 16), a control block 78 is provided to increase the value of K. Have been. The control block 78 includes a comparator 90, and controls the selection device 92 by comparing the output signal y with the reference value V. The operation is such that the selector 92 selects 0 when the output signal y is smaller than the reference value V, and selects the output of the multiplier 86 when the output signal y is larger than the reference value V. It has been done. Therefore, when the output signal y is smaller than the reference value V, the result of addition of K0 + 0 in the adder 88 is supplied to the multiplier 76 as the coefficient K. When the output signal y is larger than the reference value V, the Ig block 66,
The difference between the output signal y of the adder 74 corresponding to the output signal 70 and the reference value v is obtained by the adder 84, the difference is multiplied by a coefficient K1 · C by a multiplier 86, and the multiplied value is multiplied by the reference value K0. Is added by an adder 88, and the added value is K
Is supplied to the multiplier 76. V is the vacuum tube 1
The grid voltage C at which the grid current starts to flow at 2 and 16 is an appropriate constant, for example, 16. K1 is a constant representing a grid resistance value when a grid current flows.

Since the control block 78 is configured as described above, when the output signal y is larger than the reference value V, K
Is represented by Equation 2.

K = K0 + K1 * C * (y-V) Therefore, when the output signal y of the Ig blocks 66 and 70 is V or less, K0 is supplied as K, and the cutoff frequency is K
However, when the output signal y becomes larger than V, K which is larger than K0 is supplied in proportion to (y−V), and the cutoff frequency of the Ig blocks 66 and 70 becomes higher. . Further, performing the above processing causes the value of the coefficient K of the high-pass filter 83 to change asymmetrically with respect to the center of the amplitude of the input signal, and as a result, the capacitor 10 described in FIG. It will also simulate the change in bias point due to the accumulation of charge on. In addition, every time the output signal y is supplied from the adder 74, the calculation may be performed in accordance with the above-described calculation process. However, the value of K corresponding to each value of y is stored in a table in advance, and Every time it is input, the value of K corresponding to the y may be output.

Output signals obtained by simulating the grid currents from the Ig blocks 66 and 70 are supplied to Ip / Eg blocks 68 and 72, and are output from the Ip / Eg blocks 6 and 72.
8, 72 are the vacuum tubes 1 based on the input grid voltage Eg.
An output signal corresponding to 2, 16 plate currents is generated. This output signal is represented by, for example, a cubic function as shown in Expression 3, and when the output signal from the Ig blocks 66 and 70 has a large value, a clipped output signal is generated.

F (X) = 2 * {([-1/7] * X + [1/14] * X + [4/7]} * X where X is an output signal from the Ig blocks 66 and 70. Each of the Ip / Eg blocks 68 and 72 is given by the following equation (3) every time an output signal is supplied from the Ig blocks 66 and 70.
It is also possible to use one that performs an operation in accordance with the following equation. Alternatively, a value of f (X) corresponding to each value of X is stored in advance as a table, and f (X) corresponding to the value of X is read out. You may.

Output signals from the Ip / Eg blocks 68 and 72 are supplied to Ig blocks 90 and 92 of the output block 31a, respectively. These Ig blocks 9
Numerals 0 and 92 simulate a high-pass filter including the DC blocking capacitors 30 and 34 and the grid resistors 38 and 40 in FIG. 4, and have the same configuration as the Ig blocks 66 and 70 of the drive block 11a, and have some parameters. Are different. The output signals of these Ig blocks 90 and 92 are subjected to a change in the bias by adders 94 and 96 as described later, and then are applied to the Ip / Eg block 9.
8, 100. These Ip / Eg blocks 9
8 and 100 have the same configuration as the Ip / Eg blocks 68 and 72 of the drive block 11a, but have slightly different parameters and grid voltage to plate current conversion characteristics. The output signals of these Ip / Eg blocks 98 and 100 are added by an adder 102 and output. The difference between the parameters of the Ig block and the Ig / Eg block and the conversion characteristics of the grid voltage to the plate current is as follows.
The values of the capacitors and resistors to be simulated,
This is caused by the difference between the characteristics of the vacuum tube and the vacuum tube.

The output signal of the adder 102 is Ip /
In order to change the bias of the Eg blocks 98 and 100, it is supplied to the level detection unit 104. The level detector 104 includes, for example, as shown in FIG.
6 to detect the absolute value of the output signal of the adder 102. This absolute value is supplied to the low-pass filter 108. This low-pass filter 108 has a very low cut-off frequency of, for example, 0.2 Hz, and suppresses the voltage drop of the power supply voltage + B generated when the output signal from the output stage 31 enters an overdrive state in FIG. It is provided in order to simulate gradual occurrence under the influence of 62. The output signal of the low-pass filter 108 is output as an output signal of the level detector 104. Note that a square circuit may be used instead of the absolute value circuit 106. Further, in order to convert a change (envelope) characteristic of the level detected by the level detection unit 104 into an appropriate (for example, a square characteristic), a conversion unit such as a table or a calculation may be provided.

The output signal of the level detector 104 is multiplied by an appropriate constant by a multiplier 110, and then supplied to an adder 112 where it is added to a predetermined standard bias. Supplied. Therefore, when the level detected by the level detection unit 104 increases, the bias supplied to the adders 94 and 96 also increases, simulating a state where the bias is deeply applied. Thus, the output signal of the class B push-pull amplifier can be simulated.

In the above embodiment, the drive block 11a was also simulated as a class B push-pull.
It can also be simulated as a class or AB class drive stage. However, in that case, it is necessary to change Equation 3 according to the class A or the class AB. Further, in the above embodiment, the drive block 11a does not simulate the fluctuation of the bias point due to the drop of the power supply voltage due to the overdrive state of the output signal. This is because the influence of the fluctuation of the bias is small in the drive block 11a such that the influence of the drop does not affect the drive stage 11 very much.
The variation of the bias point may be simulated similarly to the output block 31a.

In the above embodiment, the present invention is applied to a class B push-pull amplifier. However, the present invention can be applied to a class A single amplifier. In this case, for example, only the Ig block 90 and the Ip / Eg block 98 in FIG. 1 may be provided. However, in this case, Equation 3 used in the Ip / Eg block 82 needs to be changed to one according to the class A amplifier. Also in the class A, when simulating the voltage drop based on the clipped output signal, the Ig block 90 and the Ip / Eg block 9 are used.
In addition to 8, the level detector 104, the multiplier 110, the adder 112, and the adder 94 may be added.

FIG. 5 shows a second embodiment . This embodiment uses diodes 114 and 116 as shown in FIG.
Are simulated in a configuration in which a high-pass filter 124 composed of a DC blocking capacitor 120 and a resistor 122 is provided in the preceding stage of a diode clip circuit 118 connected in anti-parallel. In the circuit of FIG.
When the diode 114 or 116 is not conducting, the resistance value of the diode 114 or 116 is infinite. Therefore, the time constant of the high-pass filter 124 is determined by the value of the capacitor 120 and the resistor 122. However, when an input signal is provided, the diode 114 or 1
When 16 conducts, the resistance value of the diode 114 or 116 decreases from infinity, the time constant of the high-pass filter 124 decreases, and the cutoff frequency increases.

To simulate this state, FIG.
As shown in (a), a high-pass filter block 126 and a diode clip block 128 are provided. The output signal of the diode clip block 128 is supplied to the control block 130, and the control block 130 changes the time constant of the high-pass filter 126.

The high-pass filter 126 may have the same configuration as the high-pass filter 83 shown in FIG. 2, and the control block 130 may have the same configuration as the control block 78. However, the constants K0, C, K1, etc. in Equation 3 need to be changed as appropriate. Further, the diode clip block 128 outputs the input signal as it is while the input signal having a level between the predetermined positive and negative threshold levels + L and -L is supplied, and the input signal having a level larger than + L is supplied. This can be realized by a table configured to output + L when supplied and to output -L when an input signal having a level smaller than -L is supplied. According to FIG. 2, in the case of the embodiment of FIG.
Input the input signal of the high-pass filter 126, but may input the output signal of the diode clip block 128 as shown by the broken line. In some cases, the high-pass filter 124
May be simulated in which the resistor 122 is omitted, in which case the diode clipping circuit 118
May be simulated instead of the resistor 122 although the input impedance is connected to the subsequent stage. Further, in the above description, the simulation in which two diodes are connected in parallel in opposite directions is simulated, but the table of the diode clip block is changed or the control block 13 is changed.
By changing the threshold level of 0 to only positive, one diode can be easily simulated.

[0036]

As described above, according to the first aspect of the present invention,
On the input side of the non-linear processing means that generates a larger distortion as the input signal increases, high-pass filter means capable of varying the cutoff frequency is provided. For example, the output signal level of the high-pass filter means increases. The cut-off frequency of the high-pass filter formed by a DC blocking capacitor or a resistor on the input side of a vacuum tube amplifier or a diode clipping circuit is controlled by the grid current or the diode current. Can be simulated, and the operation of the vacuum tube amplifier and the diode clipping circuit can be accurately simulated.

Further, according to the second invention, as the level of the output signal of the nonlinear processing means large distortion as the input signal is large is generated increases, the non-linear processing hand
Since the bias of the stage is made deeper, it is possible to simulate the fluctuation of the bias due to the fluctuation of the power supply voltage when the output signal is overdriven in the vacuum tube amplifier, and it is possible to accurately simulate the operation of the vacuum tube amplifier.

[0038]

[Brief description of the drawings]

FIG. 1 is a block diagram of a first embodiment of a signal processing device according to the present invention.

FIG. 2 is an Ig block 66 or 70 of the first embodiment.
It is a detailed block diagram of.

FIG. 3 is a detailed block diagram of a level detection unit 104 according to the first embodiment.

FIG. 4 is a circuit diagram of a class B push-pull amplifier simulated according to the first embodiment.

FIG. 5 is a block diagram of a second embodiment of the first invention and a circuit diagram of a diode clip circuit simulated in the second embodiment.

[Explanation of symbols]

68 72 98 100 Ip / Eg block (simulation means for vacuum tube amplifier) 83 126 High-pass filter (high-pass filter means) 78 130 Control block (filter control means) 104 Level detector (bias changing means) 110 Multiplier (bias change) Means) 112 adder (bias changing means)

Continuation of the front page (51) Int.Cl. ⁷ identification code FI H03G 5/16 H03G 5/16 D (58) Field surveyed (Int.Cl. ⁷ , DB name) H03F 1/00 G10H 1/02 H03F 3 / 28 H03G 3/20 H03G 5/16

Claims (2)

(57) [Claims] 1. Non-linear processing means which generates a larger distortion as the amplitude value of an input signal increases, and high-pass filter means provided on the input side of the non-linear processing means and whose frequency characteristics can be changed; A signal processing device comprising: a filter control unit that controls a frequency characteristic of the filter unit according to an output level of the filter unit. 2. A non-linear processing means in which a larger distortion is generated as the amplitude value of the input signal increases, detecting an output level of the non-linear processing means, and applying a bias to be added to the input signal based on the detection result. And a bias changing means for changing the bias.