JPH08146953A - Signal processor - Google Patents

Abstract

PURPOSE: To selectively obtain dummy exponential function value data or linear function value data by shifting multiplication data outputted from a multiplying means according to the contents of a supplied exponent data part. CONSTITUTION: A selector A11 selects and converts one of data from a counting register 10 and an internal bus and outputs the exponent data part of the converted data to a shifter 14 and the data part to a multiplier 12. The multiplier 12 multiplies data from the selector A11 by data from a selector B13 and outputs the result to the shifter 14. The shifter 14 shifts the output of the multiplier 12 with the exponent data supplied from multiplication data selector A11. An adder 15 adds the data outputted from the shifter 14 and data outputted from a selector C19 together and supplies the result to a full-wave rectifying means 19. Consequently, count data on dummy exponent characteristics or coefficient data on linear characteristics are obtained.

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 capable of selectively outputting pseudo exponential characteristic data or linear characteristic data, and is particularly suitable for application to a musical sound effect imparting device. It is something.

[0002]

2. Description of the Related Art A wah effect, a phasor effect, etc. have been conventionally known as effects for continuously changing the timbre of musical tones. The wah effect changes the timbre by continuously changing the harmonic characteristics of the musical sound.
As shown in (a), by continuously moving the frequency characteristics of a kind of bandpass filter in parallel on the frequency axis,
The wah effect is added to the musical sound. The phasor effect is to change the timbre by continuously changing the phase characteristic, and the phase characteristic of the all-pass filter is continuously moved in parallel on the frequency axis as shown in FIG. 13 (b). By doing so, the phasor effect is added to the musical sound.

Thus, the effects of wah, phasor, etc. are
It is the effect brought about by the characteristic change on the frequency axis, but for the listener, the sensitivity to the frequency change is
It is conventionally known that the characteristic is not linear but logarithmic. That is, the frequency is 10Hz → 20H
When changing to z and when changing from 100 Hz to 200 Hz, in the former case, the frequency change is 10 Hz, whereas in the latter case, the frequency change is 100 times the frequency, which is 10 times the frequency change. Both are changes of one octave, and will be heard at the same change rate. Therefore, when it is desired to keep the rate of change of the auditory effect constant, it is necessary to change the exponential characteristic.

However, the operation amount data of a general controller and the periodic waveform of an LFO (Low Frequency Oscillator) generally have a linear change when viewed on the time axis. Therefore, conventionally, an exponential function table is prepared to obtain the exponential function value, and the exponential function value is obtained by accessing this table. However, since the exponential function table is configured by using a memory, there is a drawback that a large memory is required and a configuration for accessing the memory is required.

Therefore, in order to solve this, Japanese Unexamined Patent Publication (Kokai) No.
As shown in Japanese Patent Laid-Open No. 149244, by connecting two squaring circuits in cascade to obtain a function value obtained by squaring,
It has been proposed to obtain a pseudo exponential value. The principle of this method is that the characteristic of X ^{4 obtained} by squaring data X is as shown in FIG. 14, and the exponential function 5 shown in FIG.
It aims to obtain a pseudo exponential function value by taking advantage of the characteristics that are close to 5.3 ^X-1 .

[0006]

As described above, the structure for obtaining the pseudo exponential value by multiplying the data X to the fourth power to obtain X ⁴ is generally two stages of multipliers by individual hardware. They are connected in cascade or are configured by a power calculation by DSP. However, when hardware is used, the configuration of the multiplier is complicated, and DSP is not good at power calculation in power calculation by DSP. ,
There is a problem that the number of steps of the microprogram is wasted. Further, in a signal processing device for processing voice or musical sound, a low pass filter having a very low cutoff frequency may be required. In such a case, a minute filter coefficient value is required, but in order to obtain a minute filter coefficient, the number of bits of the filter coefficient must be increased. There was a limit.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a signal processing device having a simple structure capable of selectively obtaining either pseudo exponential function value data or linear function value data. Another object of the present invention is to provide a signal processing device capable of obtaining a minute coefficient value.

[0008]

In order to achieve the above object, in the signal processing device of the present invention, the coefficient supplying means for supplying the coefficient and the coefficient supplied from the coefficient supplying means have a pseudo exponential characteristic. Output by converting the coefficient data supplied from the coefficient supply means into linear data, and a selector means for converting the coefficient data to be output as linear data and multiplying the data part of the coefficient data output from the selector means. Means and coefficient data output from the selector means,
The selector means is applied with a select signal for instructing which output is to be output, and the shift means applies the contents of the supplied exponent data section. Accordingly, the coefficient data having the pseudo exponential characteristic or the coefficient data having the linear characteristic is obtained by shifting the multiplication data output from the multiplying means.

[0009]

According to the present invention, the coefficient value of a linear change or the coefficient value of an exponential change can be selectively obtained, so that the hardware can be efficiently used. Further, even if the number of bits of the coefficient is small, it is possible to obtain a minute coefficient value.

[0010]

1 is a block diagram showing the configuration of an electronic musical instrument to which a signal processing device of the present invention is applied. In this figure,
Reference numeral 1 is a microprocessor (CPU) that detects an event of the performance operator 5 and supplies sound data and the like to a musical tone generating section 6 to generate a musical tone, and 2 is a preset tone color. ROM (Read Only Memory) in which data and control programs for the CPU 1 are stored,
Reference numeral 3 denotes a RAM (Rand) used for an area for storing tone color data set by the user, a work area for the CPU 1, and the like.
om Access Memory), 4 is a manipulator for imparting effects such as pedals and wheels, 5 is a keyboard (KB) as a performance manipulator, and 6 is a sound source section (tone generator) that generates a musical sound under the control of the CPU 1. TG) 6-1 and DSP (Digital Sign) that adds effects and the like to the musical sound synthesized by the sound source unit 6-1.
a musical tone generating unit including a signal processing unit 6-2 configured by an al processor), a speaker 7 for generating the generated musical tone,
Reference numeral 8 is an address / data bus.

In the electronic musical instrument thus constructed,
K. When B5 is operated, the event is detected by the CPU 1, and the K.K. The key data of B5 is transferred to the musical sound generating unit 6. Tone generator 6 of the tone generator 6-
1, the transferred key data and the musical tone signal corresponding to the tone color designation information transferred from the ROM 2 or the RAM 3 are synthesized, and then the musical tone synthesized by the signal processing unit 6-2 depends on the operation amount of the operator 4. The effect of K. A musical sound corresponding to the operation of B5 is produced from the speaker 7.

By the way, D including the signal processing device of the present invention
FIG. 2 shows a block diagram of the signal processing unit 6-2 including the SP. In the figure, 10 is a coefficient register in which coefficients transferred via a CPU bus corresponding to the address / data bus 8 shown in FIG. 1 are accumulated, and 11 is a coefficient register 1.
Selectors A and 12 which select either 0 or data from the internal bus to convert and output the exponent data part of the converted data to the shifter 14 and the data part to the multiplier 12 are the data from the selector A11. And selector B1
A multiplier that multiplies the data from 3 and outputs it to the shifter 14, 14 is a shifter that shifts the multiplication data output from the multiplier 12 by the exponent data supplied from the selector A 11, and 15 is output from the shifter 14. An adder for adding the data and the data output from the selector C19 to the full-wave rectification means 16, and a full-wave rectification of the data by folding back the negative data of the supplied data to the positive side. It is a full-wave rectifying means for outputting to the internal bus.

Reference numeral 17 is a temporary register (T-R) which selects either an input signal or data from the internal bus.
AM) 18 stores the selectors D and 18 temporarily store the data stored in the selector B13 and the selector C1.
9 is a temporary register to be supplied to 9, a selector C for selecting data from the internal bus, data from the temporary register 18 or data of "0" and supplying it to the adder 15 is executed by a selector C, 20 which performs signal processing. in order to,
For example, a micro program register that stores a micro program of 256 steps or 128 steps, 21 is an address register that stores address data that is read in response to the execution of the micro program, and 22 is read from the address register 21. Address controller for address, 23 is latch means for latching data from the internal bus and outputting to the adder 24, 2
Reference numeral 4 is an adder for adding the data latched in the latch means 23 to the address output from the address controller 22, 25 is a predetermined timing of the microprogram, and the internal bus is added according to the address added by the adder 24. It is a delay memory that controls writing of the above data or reading of stored data to the internal bus.

Various kinds of signal processing are carried out in the signal processing section thus constructed. An example of signal processing executed by the signal processing apparatus according to the present invention will be described below. In the signal processing of this example, the coefficient signal processing for giving an effect to a musical sound is performed by the coefficient register 10 and the selector A1.
1, the multiplier 12, and the shifter 14. As described above, the effects of wah, phasor, etc. are the bandpass filter (BPF) and allpass filter (AP
The frequency characteristic of F) is given by moving in parallel on the frequency axis. Such BPF and APF
Is generally configured so that its frequency characteristic also changes in proportion to the filter coefficient.

Therefore, when the frequency characteristic of the filter is to be translated in parallel on the frequency axis in order to impart the effect of wah, phasor, etc., the filter coefficient is exponentially functioned in order to obtain a audible linear tone color change. Need to change to. On the other hand, the filter is a low pass filter (LPF)
For example, when forming a sine wave from a triangular wave,
It is convenient to adjust the filter coefficient that determines the cut-off frequency of the LPF according to the frequency of the fundamental wave and the filter coefficient that changes linearly. Therefore, in the signal processing device of the present invention, it is possible to obtain a filter coefficient value that is a pseudo-exponential function or a filter coefficient value that changes linearly as necessary.

Next, how the coefficient is converted into the pseudo exponential function value will be described with reference to the binary data table shown in FIG. The 4-bit binary data shown in FIG. 4A is a coefficient accumulated in the coefficient register 10, and the coefficient is preferably 10-bit binary data, but here, for ease of explanation, It has 4 bits. This 4 bit 2
The binary data is supplied to the selector A11, and when the data from the coefficient register 10 is selected by the selector A11, it is converted into the data shown in FIG.
It is output from 1. That is, binary data is separated into upper 2 bits and lower 2 bits, and a sign bit and a hidden bit are inserted between them to be converted into 6 bits. In this case, the sign bit is set to "0" indicating positive,
The hidden bit shall be "1".

Of the converted 6-bit data, the upper 2 bits are set as the data indicating the exponent and supplied to the shifter 14, and the lower 4 bits are set as the data indicating the value.
2 is supplied. In this case, in the multiplier 12, since "1" is output from the selector B13 and multiplication is performed, the input data is output as it is and the shifter 1
4 will be supplied. Then, the shifter 14 downshifts according to the supplied index data. In this case, since the exponent data is 2 bits, when the exponent data is "00", the shift is performed by 3 bits, when it is "01", the shift is performed by 2 bits, and when the exponent data is "10". Performs a 1-bit downshift, and if it is "11", the data is output as it is without downshifting.

In this way, the 7-bit converted coefficient shown in FIG. 7C can be obtained. The coefficient after conversion has a decimal point between the 4th bit and the 5th bit, and when the coefficient after conversion is graphed, it has the characteristic shown as "when hidden bit 1" in FIG. 4e. That is, it is understood that the coefficient is a pseudo exponential function by the line approximation. Note that, in the selector A11, conversion is performed so that 2 bits of "11" are added to the higher order of the coefficient, and if the added 2 bits are index data, if the index data is "11", it is not downshifted and is unchanged. Since it is output, the coefficient output from the coefficient register 10 is directly output from the shifter 14, and is output as a coefficient having a linear characteristic shown in FIG. 4f.

When the coefficients shown in FIG. 3A are converted as shown in FIG.
When converted as 0 ", the coefficient after conversion becomes as shown in FIG. 4D. When the coefficient after conversion is made into a graph, it becomes as shown in FIGS. 4A, 4B, 4C, and 4D. In particular, FIG. Paying attention to the characteristic of, it is possible to obtain a minute coefficient that linearly changes, and if this coefficient is set in the LPF as a filter coefficient, it becomes possible to obtain an LPF with an extremely low cutoff frequency.

Next, when the coefficient output from the coefficient register 10 is 10 bits, selectors for outputting a coefficient having a linear characteristic from the shifter 14 and for outputting a coefficient having a pseudo exponential characteristic from the shifter 14 The conversion of A11 will be described with reference to FIG. The coefficients output from the coefficient register 10 are “C9, C8, C
It is 10 bits of 7, C6, C4, C3, C2, C1, C0 ", and when outputting as a coefficient of linear characteristic,
3-bit "1" is added to the MSB "C9" to obtain data shown as a coefficient (linear). Since the added 3-bit "1" is used as exponential data, the shifter 1
4 does not shift down "C9, C8, C7, C
Since the 10-bit coefficient of 6, C4, C3, C2, C1, C0 ″ is output as it is, a coefficient having a linear characteristic is output from the shifter 14.

When outputting as a coefficient of the pseudo exponential characteristic, the upper 3 bits of "C9, C8, C7" are used as exponent data, the sign bit of "0" is added to the upper part of "C6", and " 2 bits of "0" and "0" are added to the lower order of C0 "to form data shown as a coefficient (exponent), so that the shifter 14 can output the data shown in FIGS.
Coefficients having pseudo exponential characteristics as shown in d can be obtained. Furthermore, the 12-bit coefficient "Y11, Y10, Y9, Y8, Y obtained by the calculation on the internal bus.
7, Y6, Y5, Y4, Y3, Y2, Y1, Y0 "are selected by the selector A11, a coefficient having a linear characteristic is output from the shifter 14 and a coefficient having a pseudo exponential characteristic is output from the shifter 14. Selector A1 in case
The conversion state of No. 1 will be described with reference to FIG.

When outputting as a coefficient having a linear characteristic, a 3-bit "1" is added to the MSB "Y11", and the lower bits of "Y1, Y0" are cut off to obtain a coefficient (linear). The data shown as Since the added 3-bit "1" is used as exponential data, the shifter 14 does not shift down to "Y11,
Y10, Y9, Y8, Y7, Y6, Y5, Y4, Y3
Since a 10-bit coefficient of Y2 ″ is output, a linear characteristic coefficient is output from the shifter 14. When outputting as a pseudo exponential characteristic coefficient, the MSB “Y11” is truncated and “Y10, Y9” is output. , Y8 "upper 3 bits are used as exponential data.
By adding a sign bit of "0" and a hidden bit of "1", the lower 10 bits of the data shown as a coefficient (exponent) are "0, 1, Y7, Y6, Y5, Y4.
Y3, Y2, Y1, Y0 ". This allows the shifter 14 to obtain the coefficient of the characteristic as shown in FIG.

FIG. 6 shows the configuration of the selector A11 in which the signal processing is performed as described above. "C in this figure
“OEF” is a coefficient supplied from the coefficient register 10, and “12-bit” data indicates a coefficient obtained by calculation on the internal bus. The selector A11 is shown in FIG.
As described above, in the “COEF” input, two output modes of a coefficient (linear) and a coefficient (exponential), and “1
It has a total of four output modes of a coefficient (linear) and two output modes of a coefficient (exponential) in data input of "2 bits", and one of these output modes is selected by a select signal. Are configured to be output.

That is, four types of converted data are input to the selector A11 to obtain each output mode.
That is, in the "COEF" input, the upper 3 bits are "1" input and the lower 10 bits are "COEF".
Input converted to combine, the upper 3 bits of "COEF", the fourth bit "0", the remaining bits of "COEF" that are the fifth to eleventh bits, and 12 and 13 bits. It is an input converted so as to consist of "0" which is the eye. In the case of "12-bit" data input, the input converted to combine the input with the upper 3 bits being "1" and the 10-bit "12 bit" data with the lower 2 bits truncated, and the first bit truncated The upper 3 bits of the "12 bit" data, the 4th bit "0", the 5th bit "1", 6
The input is converted to have the remaining 8 bits of "12 bits", which are the 1st to 13th bits. Then, of the data output from the selected selector A11, the upper 3 bits are supplied to the shifter 14 as exponent data, and the lower 10 bits are supplied to the multiplier 12 as a data value.

Next, FIG. 7A shows the configuration of the shifter 14 in which the signal processing is performed as described above. Shifter 14
Is composed of a portion for generating eight kinds of input data sequentially downshifted by 1 bit from the data output from the multiplier 12, and a selector 30 for selecting one of the eight kinds of input data. There is. Eight types of data are used because the shift signal is 3 bits and there are eight downshift modes, and one of the eight types of input data is selected and output by the shift signal. As a result, the data down-shifted by the shift signal is obtained.

Therefore, in the input data generator, the upper bits are sequentially added to the eight types of input data so as to extend the sign bit. That is, in the second input data, the upper 1 bit (MSB) is the same as the MSBs up to that point, and in the 8th input data, the upper 7 bits are the same as the previous MSBs.
As a result, the downshifted coefficient can be obtained from the shifter 14 as described above. In addition, 20 parallel lines are respectively branched and supplied from the multiplier 12 to the portion for generating the eight types of input data, as shown in detail in FIG. Further, as shown in detail in FIG. 7C, the portion indicated by B indicates that the MSB line is branched by the number of numbers attached to the slash.

Next, FIG. 8 shows an example of a wah effect circuit for giving a wah effect. This wah effect circuit includes an adder 40,
42, 44, 47 and coefficient multipliers 41, 43, 4
6 and a band pass filter (BPF) including delay means 45 and 48. By changing the coefficient applied to the coefficient multiplier 41, the Q (Qual) of the BPF is changed.
(ity factor) can be changed. Further, when the same modulation signal is applied to the coefficient multiplier 43 and the coefficient multiplier 47, the cutoff frequency of the BPF changes according to the modulation signal, and as shown in FIG.
The frequency characteristic of the PF can be translated on the frequency axis according to the modulation signal. The details of this BPF are disclosed in Japanese Patent Application Laid-Open No. 61-18212.

FIG. 9 shows an example of a phasor effect circuit which gives a phasor effect. This phasor effect circuit
Allpass filters (APF) 50, 51, 52 are connected in cascade. And the input signal and the second
The output signal from the APF 51 of the third stage is added by the adder 57 to generate the L (left) signal, and the input signal and the APF 5 of the third stage are generated.
The output signal of 2 is added by the adder 56 to generate the R (right) signal, and the output signal of the APF 52 is input to the adder 59.
Have been fed back to. With such a configuration,
The monaural input signal is output as two-channel L and R stereo signals each having a different phase, so that the musical sound can be given a feeling of expanse.

Since the APFs 50, 51 and 52 have the same structure, the APF 50 will be described as an example. The adders 60 and 62, the delay means 61, the coefficient multiplier 63,
The APF 50 is composed of 64, and the coefficient multiplier 6
By applying the modulation signal to 3, 64 as a coefficient,
The phase characteristic with respect to the frequency can be changed. That is, by applying the modulation signal to the coefficient multipliers forming the APFs 50, 51, 52, the above-mentioned FIG.
As shown in, the phase characteristic can be translated in parallel on the frequency axis to impart the phasor effect.

The modulation signal applied to the wah circuit and phasor circuit shown in FIGS. 8 and 9 is generally L.
A low frequency sine wave or the like output from an FO (Low Frequency Oscillator) is used, and this sine wave is conventionally generated by reading a memory in which data values of a sine wave are stored. However, since the memory capacity of the memory for storing the data value of the sine wave becomes large and the control circuit of the memory is required, the configuration becomes complicated and the cost is increased.

Therefore, as shown in FIG. 10, the sawtooth shape shown in FIG. 10A generated by the adder 15 which repeatedly adds a constant value in the DSP forming the signal processing unit 6-2. The wave is rectified by the full-wave rectifying means 16 shown in FIG. 2 to form the triangular wave shown in FIG. 10B, and this triangular wave is L having a cutoff frequency as low as possible.
The fundamental wave component of the triangular wave is extracted by passing it through the PF to obtain a sine wave as shown in FIG. In this case, in order to extract only the fundamental wave and obtain a precise sine wave, it is necessary to make the cutoff frequency of the LPF as low as possible. This sets the LPF filter coefficient to 0.
Although it can be achieved by setting a coefficient value close to, the minimum value of the coefficient with a limited number of bits is limited, and in order to obtain a coefficient value close to 0, a coefficient with an increased number of bits is used. I needed to. However, when the number of bits of the coefficient is increased, it takes time to process the coefficient, which causes a problem that a high-performance and expensive DSP must be used.

The signal processing device of the present invention can also solve this point. That is, by setting the hidden bit to "0", an extremely small coefficient value can be output as shown in FIG.
By applying this small coefficient value to the LPF, a precise sine wave can be obtained from the LPF. In this case, since the triangular wave includes a direct current component as shown in FIG. 10 (b), the direct current component of the generated sine wave is shown in FIG. 10 (c).
As shown in, the value gradually converges to a predetermined value. In particular, when the cutoff frequency of the LPF is low, it takes a long time to converge the direct current component, which has an adverse effect on the musical tone given the effect.

Therefore, the triangular wave shown in FIG.
As shown in FIG. 1 (a), the level is shifted to remove the DC component, and the triangular wave is input to the LPF and passed through the LPF to obtain a stable sine wave without the DC component as shown in FIG. To get. Further, the obtained sine wave is level-shifted in the opposite direction to the above, so that the target modulation waveform of the sine wave shown in FIG. In this way, L for shaping the triangular wave into a sine wave
An example of the structure of the PF is shown in FIG. This LPF
Are coefficient multipliers 70 and 72 by which the coefficient C is multiplied, delay means 73, coefficient multiplier 70 and delay means 7.
The output from the coefficient multiplier 72 is added and the output from the coefficient multiplier 72 is subtracted. Then, by the coefficient C applied to the coefficient multipliers 70 and 72, L
The cutoff frequency of the PF has been determined.

Further, as shown in FIG. 11, the configuration of a waveform shaping circuit for level shifting to obtain a sine wave is shown in FIG.
It shows in (b). In this figure, the value A, which is half the amplitude of the input triangular wave, is subtracted by the subtractor 74 to remove the DC component. This triangular wave is converted into a sine wave by passing through the LPF 75, and the value A is added in the adder 76, whereby a target sine wave as shown in FIG. 11C can be obtained.

The wah effect circuit, the phasor circuit, and the LPF described above are realized by executing a microprogram in the above-mentioned DSP, but they can be configured by hardware. In the above description, the LFO is used for modulation, but real-time modulation may be performed by using the operation displacement of the controller as a modulation signal. Further, the coefficient input from the coefficient register 10 may be converted by adding a hidden bit set to "1", or the hidden bit may be set to "0" in the case of using the operation result as a coefficient. May be. The present invention is not limited to the case of processing the coefficient that gives the wah effect or the phasor effect, but can be applied to the case of selectively obtaining either the exponential characteristic characteristic data or the linear characteristic data. It is a thing.

[0036]

Since the present invention is configured as described above, it is possible to selectively obtain the coefficient value of linear change or the coefficient value of exponential change, so that the hardware can be efficiently used. Can be used. Further, by controlling the hidden bit, it is possible to obtain a minute coefficient value even if the number of bits of the coefficient is small.

[Brief description of drawings]

FIG. 1 is a block diagram of an electronic musical instrument to which a signal processing device of the present invention is applied.

FIG. 2 is a block diagram of a signal processing unit including the signal processing device of the present invention.

FIG. 3 is a table showing how a coefficient is converted into a pseudo exponential function value.

FIG. 4 is a graph showing input / output characteristics of the signal processing device of the present invention.

FIG. 5 is a diagram for explaining how a 10-bit coefficient is converted into a linear characteristic and a pseudo exponential characteristic.

6 is a diagram showing a configuration of a selector A. FIG.

FIG. 7 is a diagram showing a configuration of a shifter.

FIG. 8 is a diagram showing an example of the configuration of a wah effect circuit.

FIG. 9 is a diagram showing an example of a configuration of a phasor effect circuit.

FIG. 10 is a diagram illustrating a process of obtaining a sinusoidal LFO signal.

FIG. 11 is a diagram illustrating a process of obtaining a stable sinusoidal LFO signal.

FIG. 12 shows an example of an LPF configuration and a sinusoidal LFO.
It is a figure which shows the structure of the waveform shaping circuit for obtaining a signal.

FIG. 13 is a diagram for explaining a wah effect and a phasor effect.

FIG. 14 is a diagram for explaining a conventional pseudo exponential function.

[Explanation of symbols] 1 CPU, 2 ROM, 3 RAM, 4 operators, 5
Keyboard, 6 tone generator, 6-1 tone generator, 6-
2 signal processor, 7 speaker, 10 coefficient register,
11 selector A, 12 multiplier, 13 selector B,
14 shifters, 15 adders, 16 full-wave rectifying means, 1
7 selector D, 18 temporary register, 19 selector C, 20 micro program register, 41,
43, 46, 63, 64, 70, 72 coefficient multiplier, 4
5,48,61,73 delay means, 74,76 adder

Claims (1)

[Claims] 1. A coefficient supply means for supplying a coefficient, and a coefficient supplied from the coefficient supply means is converted and output so as to have a pseudo exponential characteristic, or a coefficient supplied from the coefficient supply means is linear data. Selector means for converting and outputting so as to output, multiplication means for supplying a data part of coefficient data output from the selector means, and exponent data part of coefficient data output from the selector means for supplying A select signal for instructing which output is to be output is applied to the selector means, and the shift means is configured to perform the multiplication in accordance with the content of the supplied exponent data section. By shifting the multiplication data output from the means, pseudo exponential characteristic coefficient data or linear characteristic coefficient data is obtained. Processing apparatus.