JPH0764551A - Nonlinear processor - Google Patents

Abstract

PURPOSE:To prevent a foldover noise generated in nonlinear processing from being prominent by diffusing. CONSTITUTION:The nonlinear processing is applied to a digital input signal by a nonlinear processor 6 after being modulated by a phase modulator 2, and opposite phase conversion is performed on it by an opposite phase modulator 8. As for the phase modulator 2 and the opposite phase modulator 8, when an output signal from the phase modulator 2 is inputted to the opposite phase modulator 8 as it is, a signal with phase almost same as that of the input signal to the phase modulator 2 is outputted from the opposite phase modulator 8, therefore, the output signal almost same as the input signal to the phase modulator 2 can be outputted from the opposite phase modulator 8. Since the foldover noise generated by the nonlinear processor 6 is modulated only by the opposite phase modulator 8, it is possible to prevent the foldover noise from being prominent by diffusing.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a non-linear processing device which distorts an input signal and adds a harmonic component in order to add an effect to a tone signal in an electronic musical instrument or the like.

[0002]

2. Description of the Related Art Conventionally, as a nonlinear processing device, for example, as shown in FIG. 6A, when the level of an input signal a exceeds a predetermined level L, as shown in FIG.
There is a type in which the level of the output signal b is clipped to the level L and output. When such a non-linear processing is performed on a digital input signal, the harmonics f1, f2 ... As shown in the same figure (b) by the non-linear processing to the input signal of the frequency f0 as shown in FIG. . Occurs, but in these overtones, a harmonic component having a frequency higher than half the sampling frequency is generated, and when this is D / A converted,
May cause aliasing noise fa.

In order to reduce such aliasing noise, there has been a method in which a DSP is used to oversample an input signal, perform non-linear processing, and then undersample through a decimation filter.

[0004]

Although over-sampling and under-sampling can reduce aliasing noise, the amount of DSP processing increases due to the filter processing associated with over-sampling and under-sampling, and the processing time increases. There has been a problem that the DSP becomes longer and there is no room to allow the DSP to perform processing for adding an effect other than the effect by the non-linear processing.

[0005]

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and includes first modulating means for modulating a digital input signal and an output signal from the first modulating means. A non-linear processing means for performing a non-linear processing of the signal, and a second modulating means for modulating an output signal from the non-linear processing means,
Noise generating means for supplying a noise signal to the first and second modulating means. The first and second modulating means modulate the input signal with the noise signal, and the output signal from the first modulating means is directly used as the second signal.
This is a modulation method in which, when the signal is input to the second modulating means, a signal substantially the same as the signal input to the first modulating means is output from the second modulating means.

The modulation method of the first and second modulating means can be phase modulation, and this phase modulation can be performed by weighted averaging. Further, the noise signal generated by the noise generating means may be a binary signal.

[0007]

According to the present invention, the input signal is modulated by the first modulating means, then nonlinearly processed by the nonlinear processing means, and again modulated by the second modulating means. When the output signal from the first modulating means is directly input to the second modulating means, the first and second modulating means output from the second modulating means almost the same signal as the input signal to the first modulating means. Since the output modulation method is used, the first modulation from the second modulation means is performed.
An output signal that is substantially the same as the input signal to the modulation means of is output. However, since the aliasing noise generated by the non-linear processing means is only modulated by the second modulation means, it is diffused in the band and becomes inconspicuous.

[0008]

[Embodiment] The nonlinear processor of this embodiment is actually a DS
Although it is realized by P, DS is used to simplify the explanation.
The function realized by P will be described as a block as shown in FIG. This non-linear processing device has a modulation means, for example, a phase modulation device 2, and this phase modulation device 2 phase-modulates a digital input signal supplied from an electronic musical instrument based on a noise signal generated by a noise generation device 4. To do. The noise signal is a positive / negative binary digital signal.

The output signal from the phase modulator 2 is supplied to the nonlinear processor 6. This nonlinear processing device 6
Is constituted by, for example, a look-up table, and when the output signal of the phase modulation device 2 is at a level equal to or higher than a predetermined first threshold level, the output signal is set to the first threshold level and the predetermined second threshold level is set. If the level is less than or equal to the threshold level of, the output signal is set to the second threshold level.

The output signal of the non-linear processing device 6 is supplied to a modulating means, for example, an anti-phase modulating device 8. The anti-phase modulator 8 modulates the phase opposite to the phase modulator 2 based on the noise signal generated by the noise generator 4.
That is, when the output signal of the phase modulator 2 is directly input to the anti-phase modulator 8, the output signal of the anti-phase modulator 8 is almost the same as the input signal of the phase modulator 2.

In the modulation of the digital input signal by the noise signal in the phase modulator 2, when the noise signal has a positive value, a value delayed by α samples is obtained by linear interpolation (extrapolation), and the noise signal has a negative value. At this time, the value advanced by α samples is obtained by linear interpolation (interpolation). That is, the phase modulation is performed by the weighted average.

For example, assuming that the digital input signal to be phase-modulated is X0 and the digital input signal one sample before is X1, the digital output signal U of the phase modulator 2 is obtained.
To

## EQU1 ## U = (1 + α) * X0-α * X1 is calculated. However, when the noise signal is positive, α is also positive, and when the noise signal is negative, α is also negative.

If α is 1/4 sample, U is U when the noise signal is positive.

## EQU2 ## U = (5/4) * X0-1 / 4 * X1, and when the noise signal is negative, U is

## EQU3 ## U = (3/4) * X0 + 1/4 * X1.

For example, the digital input signals D1 to D7 shown in FIG. 2A and the noise signal shown in FIG.
If the signal is input to the phase modulator 2, when the digital input signal D1 is input, since the noise signal is positive, the digital output signal d1 obtained by delaying the 1/4 sample phase having the value calculated by Equation 2 is input. Output as. When the digital input signal D2 is input,
Since the noise signal is negative, it is output as a digital output signal d2 that is advanced by 1/4 sample phase having the value calculated by Equation 3.

Thereafter, in the same manner, the input signals D3 to D7 are input.
Are phase-modulated according to the positive / negative of the noise signal and produce output signals d3 to d7.

Output signals d1 and d2 from the phase modulator 2
The non-linear processing device 6 performs non-linear processing, such as clipping processing. That is, as shown in FIG. 3A, a signal having a level smaller than the threshold level L, for example, d
1, d2, and d7 are output as they are and become d1c, d2c, and d7c shown in FIG. 7B, but signals having a level higher than the threshold level L, for example, d3, d4, and d5 are clipped to the threshold level L. D3c, d4c, d5c
Is output as. A harmonic component is added by this clipping process. Although not shown in the figure, when the level of the output signal from the phase modulator 2 is smaller than the threshold level -L, it is clipped to -L.

The anti-phase modulator 8 anti-phase-modulates the output signals d1c, d2c, ... From the non-linear processor 6 by using the same noise signal as that supplied to the phase modulator 2. That is, a signal whose phase has been advanced by α samples in the phase modulator 2 is an output signal whose phase has been delayed by α samples, and a signal whose phase has been delayed by α samples in the phase modulator 2 has advanced its phase by α samples. Use as output signal. This is also performed by linear interpolation (extrapolation, interpolation), similar to the phase modulation in the phase modulator 2. That is, the weighted average is used.

Therefore, for example, the output signal from the non-linear processing device 6 to be subjected to anti-phase modulation is V, the output signal of the anti-phase modulation device 8 one sample before is Y1, and the current output signal of the anti-phase modulation device 8 is If Y0,

## EQU4 ## Y0 is calculated by Y0 = V / (1 + .alpha.)-Y1 * .alpha ./ (1 + .alpha.). However, when the noise signal is positive, α is also positive, and when the noise signal is negative, α is also negative.

Assuming that α is 1/4 sample as in the above, when the noise signal is positive, Y0 is

## EQU5 ## Y0 = (4/5) * V + 1/5 * Y1 is obtained. When the noise signal is negative, Y0 is

## EQU6 ## Y0 = (4/3) * V-1 / 3 * Y1.

Therefore, when the output signal d1c from the non-linear processing device 6 is input as shown in FIG. 4A, the noise signal is positive, and therefore the 1/4 sample advance having the value calculated according to the equation 5 is carried out. The output signal D11 is output from the anti-phase modulator 8. Further, when the output signal d2c from the non-linear processing device 6 is input, the noise signal is negative, so the output signal D21 of the phase delayed by ¼ sample having the value calculated according to Equation 6 is reverse phase modulated. It is output from the device 8. Hereinafter, similarly, d3c to d7c are
The signals are subjected to antiphase modulation in D31 to D71 and output.

Since the signal thus phase-modulated by the phase modulator 2 is reverse-phase modulated by the anti-phase modulator 8, the original waveform is restored, but the aliasing generated by the non-linear processing in the non-linear processing device 6 is returned. The noise only undergoes phase modulation by the anti-phase modulator 8. Therefore, each sideband Fa generated by this phase modulation is spread to various frequencies, as is clear from the signal spectrum diagram of the output signal of the anti-phase modulator shown in FIG.
It becomes inconspicuous. In FIG. 5, F0 represents a fundamental wave component, and F1 to F5 represent overtone components added by the nonlinear processing device 6.

In the above embodiment, the modulation by the modulator 2 and the inverse modulation by the inverse modulator 8 are 1 /
Although it is performed by delaying or advancing by 4 samples, it is not necessarily limited to 1/4 sample, and may be delayed or advanced by another sample value.

Further, in the above embodiment, the modulation by the modulator 2 is performed by the digital input signal X to be phase-modulated.
Although linear interpolation (first-order interpolation) is performed by the digital input signal X1 0, 1 sample before, other interpolation methods can be used. For example, the digital input signal X0,
In addition to X1, a digital input signal X2 that is two samples before the digital input signal X0 to be phase-modulated is used, and a value a [= X1 + (X0−X1) (advanced by α samples from X0 by the values of X0 and X1. 1-α)] and a value b [= X obtained by advancing α samples from X1 by the values of X1 and X2.
2+ (X1−X2) (1−α)], the slope between a and b is calculated in advance, and when the noise signal is positive, a and b are extrapolated and the value c advanced by α samples is When the signal has a negative value, a and b may be interpolated to obtain a value d delayed by α samples. In this case, c and d are represented by Equations 7 and 8.

[0024]

## EQU00007 ## c = a + .alpha. (A-b) = X1 + (X0-X1) (1-.alpha.) +. Alpha. (X1 + (X0-X1) (1-.alpha.)-X2- (X1-X2) (1-.alpha.) ^{= (1-α 2) X0} + 2α 2 X1-α 2 X2

D = a-α (a-b) = X1 + (X0-X1) (1-α) -α (X1 + (X0-X1) (1-α) -X2- (X1-X2) (1 -Α) = (1-α) ² X0 + 2α (1-α) X1 + α ² X2

When α is 1/4 sample, the equations 7 and 8 are represented by the equations 9 and 10, respectively.

[0026]

## EQU00009 ## c = (15/16) * X0 + (1/8) * X1- (1/16) * X2

D = (9/16) * x0 + (3/8) * X1 + (1/16) * X2

Inverse modulation in the case of performing the modulation based on the equations 9 and 10 is performed by the equation 11 when the noise signal is positive and by the equation 12 when the noise signal is negative. As a result of the experiment, the number of noises was smaller in the modulation and the reverse modulation according to the numbers 9, 10, 11 and 12.

[0028]

[Equation 11] Y0 = 0.6 * V + 0.53 * Y1-0.13 * Y2

[Mathematical formula-see original document] Y0 = V where Y0 is the output signal of the anti-phase modulator 8 this time, Y0
1 is the output signal of the anti-phase modulator 8 one sample before, Y2
Is the output signal of the anti-phase modulator 8 two samples before, and V is the input signal of the anti-phase modulator 8 this time.

In the above embodiment, the nonlinear processing device 6 is
When the level of the input signal is equal to or higher than the threshold level L, the output level is clipped to L, and when the level of the input signal is equal to or lower than the threshold level -L, the output level is clipped to -L, and the level of the input signal is the threshold level. Although the input signal is output as it is between L and -L, even if the level of the input signal is between the threshold levels L and -L, a signal level different from the input signal level may be output. Good.

In the above embodiment, the binary noise signal is shown. However, for example, positive, zero, and negative ternary signals are used. In the case of zero, modulation or inverse modulation is performed. You may not. Furthermore, the first and second positive and the first and second negative four-valued noise signals are used, and in the case of the first positive or negative, 1/4 sample delay or advanced modulation and inverse modulation is performed, and the second In the case of positive or negative, 1/8 sample delay or advanced modulation and inverse modulation may be used.

[0031]

As described above, according to the present invention, the aliasing noise generated by the non-linear processing means is diffused and inconspicuous because it is modulated only by the second modulation means. Therefore, there is no need to perform oversampling or undersampling, and even if the present invention is carried out by a DSP, the processing amount of the DSP does not increase much, the burden on the DSP is small, and there is room to add other effects to the input signal. Occurs.

[Brief description of drawings]

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

FIG. 2 is an explanatory diagram of an operation of the modulator of the embodiment.

FIG. 3 is an explanatory diagram of an operation of the nonlinear processing device of the same embodiment.

FIG. 4 is an explanatory diagram of an operation of the inverse modulation device according to the embodiment.

FIG. 5 is a spectrum diagram of an output signal of the inverse modulator according to the embodiment.

FIG. 6 is an explanatory diagram of an operation of a conventional nonlinear processing device.

FIG. 7 is a frequency spectrum diagram of an input signal and an output signal of a conventional nonlinear processing device.

[Explanation of symbols]

 2 phase modulator (first modulator) 4 noise generator (noise generator) 6 non-linear processor (non-linear processor) 8 anti-phase modulator (second modulator)

Claims (4)

[Claims] 1. A first modulating means for modulating a digital input signal, a non-linear processing means for performing non-linear processing on an output signal from the first modulating means, and a second for modulating an output signal from the non-linear processing means. And a noise generating means for supplying a noise signal to the first and second modulating means, wherein the first and second modulating means modulate the input signal with the noise signal. In addition, when the output signal from the first modulating means is input to the second modulating means as it is, the same signal as the input signal to the first modulating means is the second signal.
Non-linear processing device, characterized in that it is a modulation system output from the modulation means of. 2. The non-linear processing device according to claim 1, wherein the modulation methods of the first and second modulating means are phase modulation. 3. The nonlinear processing device according to claim 2, wherein the phase modulation is performed by weighted averaging. 4. The non-linear processing device according to claim 1, wherein the noise signal is a binary signal.