JPH05197391A - Adaptive processor - Google Patents

Abstract

PURPOSE:To perform adaptive processing corresponding to the level of an audio signal component or that of a noise component. CONSTITUTION:Main input and reference input ((n*)-n) are formed from the output of microphones 1, 2. A comparator 21 compares a remainder component (epsilon) with the reference input ((n*)-n), then, a ratio Ri can be formed. A step gain (g) in accordance with a ratio R is found at a step gain setting circuit 22. An adaptive filter 9 updates a load coefficient Wnk based on the step gain (mu), and controls the ratio of reduction of a noise level.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an adaptive processing device, and more particularly to an adaptive processing device capable of reducing a noise component of a microphone output.

[0002]

2. Description of the Related Art Many microphones have a structure in which a change in sound pressure of sound waves is converted into mechanical vibration of a diaphragm and an electroacoustic conversion system is operated based on the vibration. Therefore, when the sound is picked up by the microphone, noise is generated if the diaphragm is affected by some factor.

In order to cancel the above noise, various techniques have been conventionally proposed. One of such techniques is so-called adaptive processing.

As one of the adaptive processes, for example, LM
Some use the S algorithm. In a general LMS algorithm, when updating the filter coefficient, a value obtained by multiplying the residual component after processing by a certain constant is used. in this case,
A technique is known in which the constant is first increased to increase the processing speed, and thereafter the constant is gradually decreased.

[0005]

In the above-mentioned prior art,
The constants and changes in the constants were uniformly determined without reference to the levels of signals, noises and the like. Therefore, the signal,
There is a problem in that it is difficult to perform optimal processing according to the level of noise or the like.

Therefore, an object of the present invention is to provide an adaptive processing apparatus capable of controlling the optimum processing and processing speed according to the signal and noise levels.

[0007]

According to a first aspect of the present invention, one of a pair of input signals is supplied to an adaptive filter, and the other signal of the pair of signals and the adaptive filter are supplied. In an adaptive processing device that performs subtraction with the output and outputs the subtraction output as a residual component, means for comparing the residual component and the reference input, and controlling the coefficient of the residual component based on the comparison result And a control unit for controlling the adaptive filter according to the output of the control unit.

According to the second aspect of the present invention, one of a pair of input signals is supplied to the adaptive filter, and the other signal of the pair of signals is subtracted from the output of the adaptive filter. In the adaptive processing device that outputs the subtraction output as a residual component, means for comparing the residual component and the reference input, and controls the level of the input signal supplied to the adaptive filter based on the comparison result. And a control means.

[0009]

The operation of the adaptive processing apparatus according to claim 1 will be described. By subtracting between the outputs of a pair of microphones provided close to each other, the output of one microphone includes a voice signal component and a noise component, and the output of the other microphone includes only a noise component.

The output containing the above-mentioned voice signal component and noise component is the main input, and the output containing only the noise component is the reference input. The main input described above is supplied to the subtraction means and the reference input is supplied to the adaptive filter.

Prior to the adaptive processing in the adaptive filter,
The above-mentioned reference input is compared with the residual component which is the output from the subtraction means. Based on the comparison result, the coefficient of the residual component is controlled, and thereby the adaptive filter is controlled.

In the adaptive filter, the reference input is adaptively processed so as to be equal to the noise component of the main input. Then, the adaptively processed reference input is subtracted from the main input, whereby the noise component of the main input is canceled and the audio signal component is output.

The operation of the adaptive processing apparatus according to claim 2 will be described. By subtracting between the outputs of the pair of microphones installed close to each other, the output of one microphone contains the audio signal component and the noise component, while the output of the other microphone contains only the noise component. It

The output containing the above-mentioned audio signal component and noise component is the main input, and the output containing only the noise component is the reference input. The above-mentioned main input is supplied to the subtracting means, and the reference input is supplied to the amplifier arranged in the preceding stage of the adaptive filter.

The above-mentioned reference input is compared with the residual component which is the output from the subtracting means. The level of the reference input is controlled by the amplifier based on the comparison result. The reference input is then provided to the adaptive filter.

In the adaptive filter, the reference input is adaptively processed so as to be equal to the noise component of the main input. Then, the adaptively processed reference input is subtracted from the main input, whereby the noise component of the main input is canceled and the audio signal component is output.

[0017]

Embodiments of the present invention will be described below with reference to FIGS. 1 to 10. 1 to 6 are views showing an embodiment of the present invention. In the pair of microphones 1 and 2 arranged close to each other, ambient sound is collected together with noise, converted into an electric signal, and output. Since the microphones 1 and 2 are arranged close to each other, substantially the same voice and noise are picked up, converted into an electric signal and output.

The microphones 1 and 2 are arranged not only in the same direction but also in mutually opposite directions if the distance between the microphones 1 and 2 is within a wavelength range defined by the frequency of the desired signal. May be Microphone 1
The electric signal output from the microphone 2 is supplied to the A / D conversion circuit 3, and the electric signal output from the microphone 2 is supplied to the A / D conversion circuit 4.

In the A / D conversion circuits 3 and 4, electric signals supplied from the microphones 1 and 2 are converted into digital signals. The digital signal converted by the A / D conversion circuit 3 is the main input represented by (S + n). Also, A /
The digital signal converted by the D conversion circuit 4 is ((S *)
+ (N *)).

In the above digital signal, S and (S
*) Indicates almost in-phase voice signal components. Further, n and (n *) represent noise components having low correlation with additivity.

The above-mentioned main input (S + n) is the adder 5,
The signal is supplied to the delay circuit 7 provided in the adaptive noise canceller 6. Then, the output of the A / D conversion circuit 4 is supplied to the adder 5.

In the adder 5, the main input [-(S + n)] with a negative sign is added to the output [(S *) + (n *)] of the A / D conversion circuit 4. As a result of this addition,
Since the audio signal components S and (S *) are almost in phase, they are removed and only the noise component represented by ((n *)-n) is formed. The noise component ((n *)-n) is used as a reference input. The reference input ((n *)-n) is supplied to the comparison circuit 21 and the adaptive filter 9 provided in the adaptive noise canceller 6.

In the delay circuit 7 of the adaptive noise canceller 6, the main input (S + n) is delayed for a predetermined time and then output. This delay amount corresponds to the time delay required for the calculation for the adaptive processing or the time delay in the adaptive filter 9, and can be set appropriately according to the system configuration. The main input (S + n) passed through the delay circuit 7 is supplied to the adder 8.

In the adaptive filter 9, the main input (S + n)
The signal Y as a component similar to the noise component n of is formed. That is, the filter characteristics are sequentially self-adjusted so that the output of the adaptive noise canceller 6 resembles the audio signal component S of the main input (S + n).

As the adaptive filter 9, an FIR filter type adaptive linear combiner having the configuration shown in FIG. 2 is used. In the configuration of FIG. 2, DL1 to DLL represent delay circuits, and MP1 to MPL represent coefficient multipliers. Further, 16 is an adder, and 15 and 17 are terminals, respectively.

[Z ^-1 ] in the above-mentioned delay circuits DL1 to DLL represents the delay of the unit sampling time, and W _nk supplied to the coefficient multipliers MP1 to MPL represent the weighting coefficients, respectively. The weighting coefficient W _nk is determined based on the step gain μ as described later. Further, if the weighting coefficient W _nk is fixed, it is a normal FIR digital filter.

In the adder 8, the output from the delay circuit 7 and
The signal Y output from the adaptive filter 9 and added with a negative sign is added. This signal Y is the main input (S +
It is assumed to be a component similar to the noise component n in n). Therefore, in the adder 8, the noise component n is subtracted from the main input (S + n), and the residual component ε that is substantially equal to the audio signal component S is formed. In other words, the noise component n of the main input (S + n) is minimized.

The audio signal component S as the residual component ε is supplied to the step gain setting circuit 22, the comparison circuit 21, and the D / A conversion circuit 10. In the D / A conversion circuit 10, the audio signal component S represented by a digital signal is converted into an analog signal, and the analog signal is taken out from the terminal 11.

In the comparison circuit 21, the voice signal component S as the residual component ε and the reference input ((n
*)-N) and the ratio R [R = (ε / ((n
*)-N))] is required. Then, a control signal SR for selecting the step gain μ corresponding to the ratio R is formed, and the control signal SR is set to the step gain setting circuit 2
2 is supplied.

In the step gain setting circuit 22, the step gain μi corresponding to the ratio Ri is based on the control signal SR.
Is required. An example of the correspondence relationship between the above-mentioned ratio R and step gain μ is shown in FIG. According to FIG. 3, the value of the step gain μi on the vertical axis is obtained based on the ratio Ri on the horizontal axis. Note that the relationship between the step gain μ and the ratio R is not limited to that shown by the solid line in FIG. 3, and for example, as shown by the alternate long and short dash line in FIG. You may make it set the characteristic which has the convex shape. Then, the step gain μi is supplied to the adaptive filter 9.

In the adaptive filter 9, the step gain μi
The weighting coefficient W _nk is updated based on the above, and adaptive processing is performed. The relationship between the adaptive processing performed by the adaptive filter 9 and the step gain μ is shown in FIG. As shown in the figure, if the step gain μ is large, the speed of the adaptive processing is accelerated, and the noise level sharply drops in a short time.If the step gain μ is small, the speed of the adaptive processing is slowed down. The noise level drops slowly.

An example of the effect of noise reduction according to this embodiment is shown in FIG. FIG. 5 shows a frequency spectrum when this embodiment is applied and when it is not applied. In FIG. 5, the frequency spectrum shown by the solid line is the case where one embodiment is applied, and in FIG.
This is the case where the frequency spectrum indicated by the broken line is not applied to the embodiment. Then, a sine wave of 500 Hz is added as a pseudo representation of the audio signal component S. As is clear from FIG. 5, when the embodiment is applied, the noise level in the frequency region on both sides of the 500 Hz sine wave is significantly reduced.

Here, an algorithm for adaptively operating the adaptive filter 9 will be described. Various algorithms can be used as the calculation algorithm in the adaptive filter 9, but a practical and frequently used LMS (least mean square) algorithm, which requires a relatively small amount of calculation, will be described below.

If the input vector Xk is expressed as X _k = [X _k X _k-1 X _k-2 ... X _kL ], the output Y _k of the adaptive filter 9 is Given in.

When the output of the delay circuit 7 is d _k , the residual component ε _k is ε _k = d _k −X _k ^T W _k . In the LMS (least mean square) method, the weight vector is updated according to the following equation. W _{k + 1} = W _k +2 με _k X _{k In the} above formula, μ is a gain factor that determines the stability of the adaptation speed, that is, a so-called step gain.

By updating the weighting vector as described above, an action is taken to minimize the output power of the system. Hereinafter, this operation will be described by formulating. For simplicity, when the delay circuit 7 is ignored, the residual component ε from the adder 8 is ε = S + n−Y.

The expected value of the square of (ε) is expressed by the following equation. E [ε ² ] = E [S ² ] + E [(n−Y) ² ] + 2E [S (n−Y)] Here, since S is uncorrelated with n and Y, in the above equation, , E [S (n−Y)] = 0. Therefore, the expected value E [ε ² ] of the square of (ε) is expressed by the following equation. E [ε ² ] = E [S ² ] + E [(n−Y) ² ]

The adaptive filter 9 is adjusted so that E [ε ² ] is minimized, but E [S ² ] is not affected, so the following equation is obtained. Emin [ε ² ] = E [S ² ] + Emin [(n−Y) ² ]

Since E [S ² ] is not affected,
Minimizing [ε ² ] means minimizing E [(n−Y) ² ]. Therefore, the output Y of the adaptive filter 9 is the best least squares estimation value of [n].

When E [(n−Y) ² ] is minimized,
Since [ε-S = n−Y], E [(ε−
S) ² ] is also minimized. Therefore, adjusting the adaptive filter 9 to minimize the total output power is equivalent to the residual component ε being the best least squares estimate of the speech signal component S.

The residual component ε is generally the voice signal component S with some noise component added, but the output noise component is given by (n−Y), so E [(n−
Minimizing Y) ² ] is equivalent to maximizing the signal-to-noise ratio of the output.

FIG. 6 shows a modification of the embodiment. This modification is different from the one in that the main input and the reference input are formed in an analog manner and then digitally processed. Therefore, the adder 5 is replaced with the analog adder 25, and the analog adder 25 is arranged between the microphone 2 and the A / D conversion circuit 4. In addition, other configurations, actions,
The contents of the effects and the like are the same as those in the above-described first embodiment, and therefore, the portions common to the first embodiment are denoted by the same reference numerals, and the duplicated description will be omitted.

According to this embodiment, the main input (S + n) and the reference input ((n *)-n) are formed on the basis of the outputs of the pair of microphones 1 and 2 arranged close to each other. Then, the residual component ε and the reference input ((n *)-n)
Is compared by the comparison circuit 21 to obtain the ratio Ri. The step gain setting circuit 22 obtains the step gain μi corresponding to the above-mentioned ratio Ri, and the step gain μi is supplied to the adaptive filter 9. The adaptive filter 9 uses the weighting factor W based on the above step gain μi.
_nk is updated and adaptive processing is performed.

As a result, the rate of noise level reduction [noise reduction speed] is controlled, the noise component n in the main input (S + n) is canceled, and the audio signal component S is output. Therefore, the voice signal component S, the reference input ((n *)-
Optimal processing can be performed according to the level of n).

For example, when the level of the audio signal component S is higher than the level of the reference input ((n *)-n), noise is masked and cannot be heard, so no processing is performed or the processing speed is slowed. It becomes possible to do. Further, when the level of the audio signal component S is smaller than the level of the reference input ((n *)-n), it is possible to perform processing according to the state such as speeding up the adaptive processing.

Another embodiment is shown in FIGS. The other embodiment shown in FIG. 7 is different from the above-mentioned one embodiment in that the level of the reference input ((n *)-n) input to the adaptive filter 9 is controlled. .. That is, an amplifier 31 is provided between the adder 5 and the adaptive filter 9 in place of the step gain setting circuit 22 of the above-described embodiment, and the gain of the amplifier 31 is controlled by the control signal output from the comparison circuit 21. That is, it is controlled based on SR. The same parts as those in the above-described embodiment are designated by the same reference numerals, and the duplicated description will be omitted.

FIG. 8 shows a modification of another embodiment. This modification is different from the other embodiments described above.
This means that only the signal processing in the adaptive noise canceller 6 is performed with digital signals, and all other processing is performed with analog signals. Therefore, the adder 5, the comparison circuit 21, and the amplifier 31 shown in the other embodiments are replaced with the analog adder 35, the comparison circuit 37, and the amplifier 36, and the analog adder 35 and the amplifier 36 are also provided.
Is disposed between the microphone 2 and the A / D conversion circuit 4. The contents of other configurations, actions, effects, etc. are the same as those of the other embodiments described above, and the common parts are denoted by the same reference numerals and the duplicated description will be omitted.

FIG. 9 and FIG. 10 show still another embodiment. In this further embodiment, the microphone 1,
2 shows a configuration suitable when the audio signal components S output from 2 are in-phase and the noise components are out of phase with each other.

The still another embodiment shown in FIG. 9 differs from the above-mentioned one embodiment in the following two points. (1) First, the adders 5 and 41 are arranged between the A / D conversion circuits 3 and 4 and the adaptive noise canceller 6. (2) Next, the output sides of the adders 5 and 41 and the adaptive noise canceller 6 are An amplifier 31 and a comparison circuit 21 are arranged in between,
The output of the comparison circuit 21 is supplied to the amplifier 31.

In the adder 41, the A / D conversion circuit 3,
Both digital signals supplied from 4 are provided with a plus sign. However, in the adder 5, A /
The digital signal supplied from the D conversion circuit 3 has a negative sign, and the digital signal supplied from the A / D conversion circuit 4 has a positive sign.

Since the noise components (n *) and n are opposite in phase to each other, the added output of the noise components becomes almost zero. Therefore, the adder 41 outputs the audio signal component [2S] doubled in level. The audio signal component [2S] is supplied to the comparison circuit 21 and is also used as a main input of the adaptive noise canceller 6.

Further, the output of the adder 5 is a noise component ((n *)-n) only after the in-phase audio signal component S is removed. Since the noise components (n *) and n are in opposite phase to each other, the noise components (n *) and n are further strengthened in level by being subtracted. The noise component ((n *)-n) is
It is supplied to the comparison circuit 21 and the amplifier 31.

The outputs of the adders 5 and 41 are supplied to the comparison circuit 21, and the control signal SR output from the comparison circuit 21 is supplied to the amplifier 31. In the amplifier 31, the control signal S
After the level of the reference input ((n *)-n) is controlled based on R, it is supplied to the adaptive noise canceller 6. still,
The same parts as those in the above-described embodiment are designated by the same reference numerals, and the duplicated description will be omitted.

FIG. 10 shows a modification of still another embodiment. This modification is different from the above-mentioned other embodiments in that the above-mentioned audio signal component [2S] and noise component ((n *)-n) are formed in an analog manner. Therefore, the adders 5 and 41 are replaced by analog adders 42 and
This means that the adders 42 and 45 are replaced with 45, and the adders 42 and 45 are arranged in front of the A / D conversion circuits 3 and 4. The contents of other configurations, operations, effects, and the like are the same as those of the above-mentioned other embodiments, and the common parts are denoted by the same reference numerals and the duplicated description will be omitted.

The adaptive processing apparatus shown in this embodiment can be applied to recording systems in various fields. For example, it can be applied to a small portable video camera device or a single microphone. Furthermore, the pair of microphones 1 and 2 shown in this embodiment can be used with or without directivity.

[0056]

According to the adaptive processing apparatus of the first aspect of the present invention, there is an effect that it is possible to perform the optimum processing according to the levels of the voice signal component and the noise component.

For example, when the level of the voice signal component is higher than the level of the noise component, the noise is masked and cannot be heard, so no processing is performed or the processing speed is slowed down. When is smaller than the level of the noise component, it is possible to perform processing according to the state such as increasing the speed of adaptive processing.

According to the adaptive processing device of the second aspect of the present invention, it is possible to obtain the same effect as that of the first aspect.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of an adaptive filter.

FIG. 3 is a diagram showing a correspondence relationship between a ratio and a step gain.

FIG. 4 is a diagram showing a noise level corresponding to a step gain.

FIG. 5 is a waveform diagram showing a noise reduction effect.

FIG. 6 is a block diagram showing a modified example of the embodiment.

FIG. 7 is a block diagram showing another embodiment of the present invention.

FIG. 8 is a block diagram showing a modified example of another embodiment.

FIG. 9 is a block diagram showing still another embodiment of the present invention.

FIG. 10 is a block diagram showing a modified example of still another embodiment.

[Explanation of symbols]

 1, 2 Microphone 6 Adaptive noise canceller 8 Adder 9 Adaptive filter 21, 37 Comparison circuit 22 Step gain setting circuit 31, 36 Amplifier ε Residual component μ Step gain

Claims (2)

[Claims] 1. A one of a pair of input signals is supplied to an adaptive filter, and the other signal of the pair of signals is subtracted from the output of the adaptive filter to obtain a subtraction. An adaptive processing device for outputting an output as a residual component, comprising: means for comparing the residual component with a reference input; and control means for controlling a coefficient of the residual component based on the comparison result, An adaptive processing apparatus characterized in that the adaptive filter is controlled according to the output of the control means. 2. One of a pair of input signals is supplied to an adaptive filter, the other signal of the pair of signals is subtracted from the output of the adaptive filter, and a subtraction is performed. An adaptive processing apparatus for outputting an output as a residual component, means for comparing the residual component with a reference input, and control for controlling the level of an input signal supplied to the adaptive filter based on the comparison result. And an adaptive processing device.