JPH06276595A - Adaptive noise reduction system - Google Patents

Abstract

PURPOSE:To make it possible to perform a proper cancel operation for all inputted noise. CONSTITUTION:A noise signal having the correlation with the noise signal in a main input signal is supplied as a reference input signal to an adaptive filter means 24, the output signal of the adaptive filter means 24 is subtracted from the main input voice signal and noise is cancelled. The adaptive filter means 24 updates the weight coefficient of each stage of the filter adaptively so that the output power of a subtraction means 14 may be minimized. The value of a proper gain factor is preliminarily determined when an adaptive operation is singly performed for individual noise signals. For the individual noise signals, each determined value is set as the gain factor of the stage (tap) whose coefficient absolute value is large within the filter.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an adaptive noise reduction system suitable for use in, for example, a voice pickup device for a VTR with a built-in camera.

[0002]

2. Description of the Related Art Conventionally, as this type of adaptive noise reduction system, an adaptive noise canceller as shown in FIG. 5 has been known.

FIG. 5 shows the basic configuration of this adaptive noise canceller, in which 1 is a main input terminal and 2 is a reference input terminal. The main input signal input through the main input terminal 1 is supplied to the combining circuit 4 via the delay circuit 3.
Further, the reference input signal input through the reference input terminal 2 is supplied to the combining circuit 4 via the adaptive filter circuit 5 and subtracted from the signal from the delay circuit 3. The output of the synthesis circuit 4 is fed back to the adaptive filter circuit 5 and is also led to the output terminal 6.

In this adaptive noise canceller, a desired signal s and a noise signal n0 uncorrelated with the desired signal s are added as a main input signal. On the other hand, the noise signal n1 is input as the reference input signal. The noise signal n1 of the reference input has no correlation with the desired signal s, but has a correlation with the noise signal n0.

The adaptive filter circuit 5 filters the reference input noise signal n1 and outputs a signal y that approximates the noise signal n0. In this case, in the adaptive filter circuit 5, the synthesizing circuit 4 is processed by a predetermined adaptive algorithm.
The filter coefficient for filtering the reference input noise signal n1 is updated so that the subtraction output (residual output) e.

As the output signal y of the adaptive filter circuit 5, a signal having a phase opposite to that of the noise signal n0 and an equal amplitude can be obtained. The delay circuit 3 is for compensating for the time delay required for the arithmetic processing in the adaptive filter circuit 5, the propagation time in the adaptive filter, and the like, and for adjusting the time with the signal to be subtracted.

The principle of the adaptive noise canceller will be described below.

Now, assuming that the desired signal s, noise n0, noise n1, and output signal y are statistically stationary and the average value is zero, the residual output e is e = s + n0-y. The expected value despite the squared it, because the desired signal s is the noise n0, and the output y ^{uncorrelated, E [e 2] = E} [s 2] + E [(n0 -y) 2] + 2E [s (n0 -y)] = a ^{E [s 2] + E [} (n0 -y) 2].

Assuming that the adaptive filter circuit 5 converges, the adaptive filter circuit 5 updates the filter coefficient so that E [e ² ] is minimized. At this time, since E [s ² ] is not affected, Emin [e ² ] = E [s ² ] + Emin [(n0-y) ² ].

[0010] That is, E E by [e ^2] is minimized [(n0 -y) ^2] is minimized, the output y of the adaptive filter circuit 5 will estimate of the noise signal n0.
The expected value of the output from the synthesis circuit 4 is the desired signal s
Will only be. That is, adjusting the adaptive filter circuit 5 to minimize the total output power is equivalent to the subtraction output e becoming the least-squares estimated value of the desired speech signal s.

The output e is generally the signal s with some noise remaining, but since the output noise is given by n−y, minimizing E [(n−y) ² ] is Is equivalent to maximizing the signal-to-noise ratio of.

In some cases, the synthesis circuit 4 serves as a sound synthesis means. That is, the adaptive filter circuit 5 forms a noise canceling voice signal -y having a phase opposite to that of noise and an equal amplitude, and supplies this to a speaker or the like to acoustically add to the main voice to reduce noise. To do. The residual e in this case will be picked up by the residual detection microphone.

The adaptive filter circuit 5 can be realized by either an analog signal processing circuit or a digital signal processing circuit, but in general, D
It is configured as a digital processing circuit using an SP (digital signal processor). FIG. 6 shows an example of the configuration of the adaptive filter circuit 5 in the case of the configuration of the digital processing circuit.

In this example, the adaptive filter circuit 5 comprises an FIR filter type adaptive linear combiner 100 and a filter coefficient update calculation means 110. The adaptive filter circuit 5 can be configured as software by a DSP equipped with a microcomputer. In this example, the algorithm for updating the filter coefficient is described as a case where the LMS (Least Mean Squares) method, which is widely used because of its small calculation amount and practicality, is used.

The LMS method will be described with reference to FIG. In this case, as shown in FIG. 6, an FIR filter type adaptive linear combiner 100 is used. The adaptive linear combiner 100 includes a plurality of delay circuits DL1, DL2, each having a delay time Z ⁻¹ of a unit sampling time.
...... DLm (m is a positive integer), a weighting circuit MX0 for multiplying the input noise n1 and the output signals of the delay circuits DL1, DL2, ... DLm and the weighting coefficient (filter coefficient).
MX1, MX2, ... MXm and weighting circuits MX0 to MX
An adder circuit 101 for adding the outputs of m is provided. The output of the adder circuit 101 is the signal y described in FIG.

The weighting coefficients supplied to the weighting circuits MX0 to MXm are formed in the filter coefficient calculation circuit 110 by the LMS algorithm based on the residual signal e from the combining circuit 14 and the reference input n1. The algorithm executed by the filter coefficient calculation circuit 110 is as follows.

Now, the input vector X _k at time k is
As shown in FIG. 6, X _k = [x _0k x _1k x _2k ... x _mk ] ^T , the output is y _k , and the weighting coefficient is w _jk (j = 0,1,2, ... m). Then, the relationship between the input and output is as shown in the following Equation 1.

[0018]

[Equation 1] Becomes

Then, the weight vector W _{k at} time _k
The, if defined as _{W k = [w 0k w 1k} w 2k ··· w mk] T, the input-output relationship is given by ^{_{y k = X k T · W}} k ... (1). Here, if the desired response is d _k , the residual e _k is expressed by e _k = d _k −y _k = d _k −X _k ^T · W _k (2).

In the LMS method, the update of the weight vector is sequentially performed by the equation (3) of W _{k + 1} = W _k +2 μ · e _k · X _k (3). The initial value of the weighting coefficient is set to a constant value or a random value. Here, μ is a gain factor (step gain) that determines the speed and stability of adaptation.

In the above formula (3), a vector for modifying the coefficient vector W _k at a certain time k, is a second term of the right side of the equation (3), a scalar gain factor μ and instantaneous error e _k Both directly affect the modified value. Similarly, since the reference input vector X _k also works in the form of a product, this also affects the correction value. The average convergence time constant τ _a is represented by τ _a = (n + 1) / 4 μ · trE [X _i X _j ^T ]. Here, n is the order of the reference input vector (F
Corresponding to the tap number of IR filter), trE [X _i X _j ^T ]
Is the average power of the reference input. That is, the larger the number of taps of the FIR filter, the slower the convergence speed, and the larger the gain factor μ, the faster the convergence speed.

In the case of a stationary signal, the final residual noise level is large when the convergence speed is fast, and conversely the final noise level is small when the convergence is slow. However, when the target signal fluctuates like a voice, its properties change before it completely converges. Therefore, the amount of cancellation increases as the convergence speed increases to some extent.

The value of the gain factor μ approaches the one where the output y of the adaptive filter circuit 5 cancels the noise n0 and converges so that the output of the device becomes equivalent to the desired signal s. Need to be satisfied. When 0 <μ <(power of signal) / (number of taps of FIR filter + 1) (4) μ is larger than the range of Expression (4), the output y of the adaptive filter circuit 5 diverges, and It makes a lot of noise as an output.

Conventionally, the value of the gain factor μ is set in consideration of the number of taps (filter order) of the adaptive filter, which directly affects the quality of the processed signal, and the magnitude (power) of the reference input signal. It is set to a constant value so that the adaptive filter circuit 5 operates normally (converges) by satisfying the equation (4).

Conventionally, the value of the step gain μ is calculated from the reference input power as one common value for the weighting factors w _0k , w _1k , w _2k , ... W _mk of each stage (tap). Generally, to compensate for system stability,
Calculated from the maximum power of the reference input.

[0026]

As described above, when the noise signal is input to the main input and the reference input of the adaptive noise canceller with an arbitrary time difference, the noise signal in the main input due to the adaptive operation. The components are reduced and removed.

If the noise to be reduced is a noise signal from a single noise source, the adaptive noise canceller can quickly follow the fluctuation of the noise signal, so that the appropriate step gain value By the adaptive operation, in the adaptive filter circuit 5, the adaptive linear combiner 100 including a component equivalent to the noise signal component n in the main input signal, that is, a component having the strongest correlation (component having the smallest delay) is provided. The absolute value of the coefficient at the tap position of becomes relatively large.

However, for example, a camera-integrated VTR
Considering a system such as the voice pickup device of No. 1, the noise source for the desired voice is not a specific noise source, and does not always exist alone. If it is possible to examine the characteristics of the input noise and its change in advance or at an instant, you can set or switch parameters such as appropriate step gains accordingly, but in general, , It is difficult to examine the characteristics of the input noise and its change in advance, so it is difficult to switch parameters such as the step gain,
Not realistic.

Therefore, as described above, the LMS is conventionally used.
In the method, the step gain is determined within a stable range for noise signals as far as it can be predicted, that is, as a constant value obtained from the maximum power of the reference input, and the step gain of any tap of the adaptive linear combiner 100 is determined. Is adopted. However, for that reason, since the step gain value is not appropriate for a signal with low power, there is a problem that a sufficient cancellation amount cannot be obtained.

As the adaptive algorithm, a learning identification method or the like is used in addition to the LMS method. In this learning identification method, the step gain is normalized by the reference input signal power in the adaptive linear combiner 100. Thus, although the non-stationarity of the signal is taken into consideration, the power fluctuation of the reference input signal adversely affects when the number of taps is small. Further, the learning identification method has drawbacks such as being weak against disturbance.

[0031]

In order to solve the above problems, in the adaptive noise reduction system according to the present invention, when the reference symbols of the embodiments described later are made to correspond to each other, the main input signal including the desired voice and the noise signal is A noise signal having a correlation with the noise signal in the main input signal is supplied to the adaptive filter means 24 as a reference input signal, and the output signal of the adaptive filter means 24 in the combiner 14 is supplied to the main input. In a system in which the noise signal in the main input signal is subtracted from the signal to be reduced and removed, the adaptive filter means 24 includes the noise signal in the main input,
It has a filter 240 of an order corresponding to the maximum delay time with respect to the noise signal of the reference input, and adaptively updates the weighting coefficient of each stage of the filter so as to minimize the output power of the combining means 14. The reference input signal is used to form the output signal that approximates the noise signal in the main input signal. Each of the adaptive filter means 24 updates the weighting coefficient of each stage of the filter. The gain factor for each stage is such that the delay time between the noise signal input to the main input terminal and the noise signal input to the reference input terminal is the delay to each stage of the filter as the noise signal to be reduced. It is characterized in that things which are considered to be equal to the time are set to the values for performing the optimum adaptive processing operation when input respectively.

[0032]

According to the present invention having the above-described structure, the gain factor of each stage of the filter of the adaptive filter means is equal to the delay time to each stage of the filter at the main input terminal and the reference input terminal in advance. It is set to be appropriate when a single noise signal is input with a delay time difference.

Therefore, during the main input and the reference input,
When the above noise is input, the weighting coefficient of the filter stage having a delay amount equal to the delay time difference between the noise in the main input and the noise in the reference input takes a dominant value compared to the weighting factors of the other stages. The appropriate adaptive processing is performed to reduce and remove the noise.

[0034]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the adaptive noise reduction system according to the present invention will be described below with reference to FIGS.

In FIG. 1, 11 is a main input microphone for picking up a desired voice, and 21 is a reference input microphone for picking up unnecessary voice and ambient noise in a direction to be removed as noise. In this example, the arrival direction of the desired voice is a direction from the upper side to the lower side (hereinafter referred to as the front direction) in the figure as shown by an arrow AR in FIG. It is an example of realizing a device that does not pick up the voice from () as noise.

In the case of this example, the main input microphone 11 is composed of an omnidirectional microphone as shown in FIG. On the other hand, as shown in FIG. 3, the reference input microphone 21 is a unidirectional microphone having low sensitivity in the desired voice arrival direction and high sensitivity in the rear direction.

Further, in the case of this example, these microphones 11 and 21 are arranged along the direction of the arrow AR, and the distance d between them is the unit delay time τ of the adaptive filter circuit described later in this example. It is selected to be equal to the distance (d = 2cτ) that the voice is propagated during the doubling. Note that c is the speed of sound propagation. Also, the unit delay time τ
Corresponds to the unit sampling time of the A / D converters 13 and 23 and the D / A converter 15, and τ = 1 / fs when the sampling frequency is fs.

Then, the sound signal picked up by the main input microphone 11 and converted into an electric signal is supplied to the A / D converter 13 via the amplifier 12, converted into a digital signal, and subtracted. It is supplied to the circuit 14. An audio signal obtained by collecting the sound by the reference input microphone 21 and converting it into an electric signal is supplied to the A / D converter 23 via the amplifier 22, converted into a digital signal, and then an adaptive filter circuit. 24.

In this embodiment, the adaptive filter circuit 24
2 has a filter order of 3 as shown in FIG. 2, and has a delay amount of 0 as a main input as shown in FIG. That is, the adaptive filter circuit 24 is
Two delay circuits 24 each having a unit delay time (one sample) τ
1, 242 and three weighting circuits 243, 244, 245
And a three-tap (third-order) FI including
It is composed of an R filter 240 and an LMS arithmetic circuit 247 which outputs weighting coefficients to be supplied to the weighting circuits 243, 244 and 245 of the FIR filter 240.

The adaptive filter circuit 24 can be composed of a DSP (digital signal processor) equipped with a microcomputer. Then, the digital signal from the A / D converter 23 is supplied to the arithmetic circuit 247 and also to the subtraction circuit 14 via the FIR filter 240.

The output signal of the subtraction circuit 14 is the arithmetic circuit 24.
The analog signal is returned to the analog signal by the D / A converter 15 and is led to the output terminal 16. Alternatively, the D / A converter 15 may be omitted, and the output signal of the subtraction circuit 14 may be derived as a digital signal to the output terminal 16.

In this case, the LMS of the adaptive filter circuit 24
Weighting coefficients w _0k , w supplied from the arithmetic circuit 247 to the weighting circuits 243, 244, 245 of the respective taps of the FIR filter.
_1k and w _2k are the following arithmetic expressions: w _{0k + 1} = w _0k + 2μ ₀ · e _k · x _0k (5) w _{1k + 1} = w _1k +2 μ ₁ · e _k · x _1k (6) w _{2k + 1} = w _2k + 2μ ₂ · e _k · x _2k (7)

In this example, this (5),
Each step gain μ ₀ in the equations (6) and (7),
μ ₁ and μ ₂ are appropriate values that have been obtained and set in advance as described below. That is, in this case, in the update equation of the weighting coefficient in the LMS operation circuit 247 of the adaptive filter circuit 24, a constant value step gain is not assigned to all taps, but the tap position and the main input terminal of the noise signal. And an appropriate step gain is assigned according to the input time difference to the reference input terminal.
An example of a method of determining the step gain value will be described below.

[Method of Experimentally Determining Step Gain] First, a method of experimentally determining the step gain will be described.

According to this method, in the system, an appropriate step gain value for independently performing adaptive processing on each noise signal is obtained in advance. Then, with respect to each of the individual noise signals, the step gain of the tap having a large coefficient absolute value in the FIR filter, that is, the tap having a high correlation between the main and reference inputs is set to the obtained value.

A further explanation will be given by applying it to the example of FIG. In this case, as shown in FIG. 3, the desired voice arrival method (direction of arrow AR) is the front direction (0 degree direction), and the front direction is 1
The direction that makes an angle of 80 degrees is 90 degrees with the back direction and the front direction.
The direction that forms an angle of degrees is called the side.

In the example of FIG. 1, the inventors of the present invention have shown that when a noise signal is independently incident from the side or the back direction,
We investigated appropriate step gains for signals with various characteristics and signal levels. As a result, it became clear that the appropriate step gain value depends only on the incident direction, and the value increases as the incident direction approaches the front direction.

In the omnidirectional microphone 11 and the unidirectional microphone 21 used for the main input and the reference input, the power during the reference input decreases as the incident angle of the noise signal becomes in the front direction. Since the canceling effect is impaired, it is considered that the appropriate step gain value becomes larger. That is, it can be said that the appropriate step value according to the incident direction is approximately proportional to the sensitivity ratio of the two input microphones 11 and 21.

As a result of investigating the weighting coefficient of the filter in each case, the coefficient of the tap corresponding to the input time difference of the main and reference of the noise signal has a very dominant value compared with the coefficients of the other taps. Therefore, it was found that an appropriate step gain should be given to update the coefficient of the tap corresponding to the input time difference of each signal.

Therefore, the step gain of the weighting coefficient of each tap is set as follows. In this case, consider a case where a noise signal is incident from the side to be reduced to the back side.

For example, when a single noise signal is incident from the side and the system of FIG. 1 is adaptively operated, this noise signal is simultaneously applied to both microphones 11 and 21 (without delay).
Sound is picked up, so that when the system adaptively reduces this noise, the coefficients of the first tap of the FIR filter 240 dominate over the other taps. Therefore, an appropriate step gain value obtained as a system at that time is set to the step gain μ ₀ of the tap at the first stage.

Next, by injecting a single noise signal from the rear direction to adaptively operate the system of FIG. 1, the position of the main input microphone 11 and the reference input microphone 2 are set.
As described above, the distance d from the position of 1 is set to 2 samples per unit sampling time. Therefore, the signal from the rear direction is 2 samples faster than the reference input, The sound is picked up by the microphones 11 and 21,
At this time, the coefficient of the third tap of the FIR filter 240 is superior to the other taps. Therefore, an appropriate step gain value at that time is set to the step gain μ ₃ of the _third tap.

When the noise signal arrives from the middle of the range from the side to the rear, the microphones 11 and 21 pick up the sound with the reference input faster by 0 to 2 samples. As a result, the coefficient of the tap in the second stage becomes superior to other taps. Therefore, the step gain μ ₂ of the tap at the second stage is, for example, an appropriate noise obtained by applying a noise signal to the system of FIG. Set a different step gain value. In addition, this 2
The step gain μ ₂ of the tap at the stage may be set to an intermediate value between the set step gain μ ₀ and the set step gain μ ₃ .

The signal from the front direction is picked up by both microphones 11 and 21 with the reference input being delayed by two samples. Therefore, the time difference between the front direction signals that are not to be canceled is outside the FIR filter 240, and a more ideal cancel operation can be obtained.

As described above, the step gains μ ₀ , μ ₂ and μ ₃ of the respective taps of the FIR filter 240 are set in advance and the results of the adaptive noise reduction processing are performed. FIG. 4 shows a comparison with the result of the adaptive noise reduction process (conventional method) that gives a gain.

FIG. 4 shows the amount of cancellation of the respective noise signals when the noise signals arrive from the side and the back side alternately in terms of time. Here, the amount of cancellation is the power level of the noise mixed in the main input before processing,
It is a ratio with the noise power level remaining in the system after processing, and a negative sign means a decrease in power, and the larger the absolute value, the larger the cancellation amount.

As is apparent from FIG. 4, in the case of the noise signal from the rear direction, the example of FIG.
Although the difference is not so remarkable, the amount of cancellation of the noise signal from the side is about 6 in the case of the example of FIG. 1 compared with the conventional example.
It can be seen that the dB is improved.

The value of the step gain in the case of the conventional example is a value obtained from the long-term average power of the reference input signal. When a unidirectional microphone facing the back direction is used as the reference input microphone, the power in the back direction is larger than that in other directions such as the side direction, and is dominant to the long-term average power. Therefore, the stability limit value of the step gain, which is proportional to the reciprocal of the power, becomes small, and also becomes 1/10 of the generally appropriate value. From this, even if the step gain value calculated in this way is an appropriate value in the back direction, it becomes too small for the signal from the side with small power during reference input, and is not appropriate.

On the other hand, in the case of the present invention, since the step gain is set to a value suitable for each of the lateral direction and the rear surface direction, the lateral cancel amount is improved as compared with the conventional method. -ing This indicates that the noise from the side can be effectively canceled without impairing the amount of noise cancellation in the back direction, and it can be seen that this is very effective.

[Theoretical Method of Determining Step Gain Value] As described above, it can be said that the appropriate step gain value according to the incident direction is approximately proportional to the sensitivity ratio of the two input microphones 11 and 21. Therefore, if the directional characteristics of the microphones 11 and 21 are known, the value of the step gain of each tap of the FIR filter 240 can be theoretically set in advance using the directional characteristics.

That is, the value of the step gain is proportional to the value of the sensitivity ratio of the main input microphone 11 to the reference input microphone 21 in the incident direction of the noise signal,
It is given as the value of the step gain of the tap corresponding to the difference in the incident time between the main input determined by the incident direction of the noise signal and the reference input.

For example, as shown in FIG. 3, when the maximum sensitivity value of the directional characteristics of the microphone 11 and the microphone 21 is "1", the microphone 21 having unidirectionality is used.
As shown in FIG. 3, the sensitivity of the directional characteristics of the back direction is “1” in the back direction, but the side sensitivity is
It is set to about "0.5", and the sensitivity in the angular range between the lateral direction and the back direction is "0.5" to "1".

Therefore, for example, when the input direction of the noise signal is the rear direction, the sensitivities of both microphones 11 and 21 are both "1", and the sensitivity ratio is also 1. As the step gain at this time, for example, a value substantially equal to the conventional value can be set. Since the delay time difference between the main and reference inputs when the noise signal input direction is the back direction is 2 samples, this value is 3 of the FIR filter 240.
Set as the step gain value μ ₂ of the tap at the stage.

When the input direction of the noise signal is diagonally backward, the sensitivity of the directional characteristic of the microphone 11 which is omnidirectional in that direction is "1", but the directional characteristic of the microphone 21 which is unidirectional. The sensitivity of the characteristic is “0.8”, for example. Therefore, the value of the sensitivity ratio of both is "1.25". Therefore, a value that is 1.25 times the step gain value μ ₂ is set as the step gain μ ₁ of the second tap.

Further, when the input direction of the noise signal is lateral, the sensitivity of the directional characteristics of the microphone 21 is
It is “0.5”. Therefore, the microphone 1
The value of the sensitivity ratio of 1 and 21 is “2”. Therefore, a value that is twice the step gain value μ ₂ is set as the step gain μ ₀ of the first tap.

In the above example, the main input microphone and the reference input microphone are not limited to those having the directivity of the above-described example, and for example, the main input microphone has the maximum in the desired voice direction. A unidirectional microphone arranged so as to have sensitivity can be used, and as the reference input microphone, a bidirectional microphone having little sensitivity in the desired voice arrival direction may be used.

The directional microphone can be constructed by using a plurality of omnidirectional microphone units.

[0068]

As described above, according to the present invention, an appropriate step gain value is obtained in advance for each noise signal input, and the delay between the main and reference inputs of the noise signal is delayed. Since it is set as the step gain for updating the filter coefficient of the tap that becomes the delay output corresponding to the quantity, it is not necessary to successively obtain an appropriate step gain for the noise signal input at the moment.

Since an appropriate step gain value is preset for each noise signal, an appropriate reduction / removal operation is always performed on the noise signal. That is, when noise is input to the main input and the reference input, the weighting coefficient of the filter stage having the delay amount equal to the delay time difference between the noise in the main input and the noise of the reference input is the weighting coefficient of the other stage. Takes a predominant value compared to, and appropriate adaptive processing is performed,
The noise is reduced and removed.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of an adaptive noise reduction system according to the present invention.

FIG. 2 is a block diagram of an example of the adaptive filter circuit of the example of FIG.

FIG. 3 is a diagram for explaining directional characteristics of microphones 11 and 21 in the example of FIG.

FIG. 4 is a diagram for explaining an effect of the present invention in comparison with a conventional one.

FIG. 5 is a block diagram showing an outline of an adaptive noise reduction system.

FIG. 6 is a block diagram of an example of an adaptive filter circuit.

[Explanation of symbols]

11 main input microphone 14 subtraction circuit 21 reference input microphone 24 adaptive filter circuit 240 FIR filter 243 to 245 weighting circuit w _{0k to} w _2k weighting coefficient

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁵ Identification code Office reference number FI technical display location H04R 1/40 320 Z

Claims (2)

[Claims] 1. A main input signal including a desired voice and a noise signal is supplied to a synthesizing means, and a noise signal correlated with a noise signal in the main input signal is supplied to an adaptive filter means as a reference input signal. In the synthesizing system, the output signal of the adaptive filter means is subtracted from the main input signal to reduce and remove a noise signal in the main input signal. A filter having an order corresponding to the maximum delay time between the signal and the noise signal at the reference input is provided, and the weighting coefficient of each stage of the filter is adaptively adjusted so as to minimize the output power of the combining means. To form the output signal that is approximated to the noise signal in the main input signal from the reference input signal. The gain factor for each stage for updating the weighting coefficient of each stage of the filter is the noise signal between the noise signal input to the main input terminal and the noise signal input to the reference input terminal as the noise signal to be reduced. The delay time of the filter is set equal to the delay time up to each stage of the filter, and is set to a value for performing the optimum adaptive processing operation when each is input. Noise reduction system. 2. The adaptive noise reduction system according to claim 1, wherein the main input and the reference input are signals from a voice sensor, and at least the voice sensor for reference input has sensitivity in a desired voice arrival direction. Is low, the gain factor for each stage for updating the weighting coefficient of each stage of the filter of the adaptive filter means, the main input terminal and the reference input terminal, the delay to each stage of the filter The adaptive noise is set to have a value according to the sensitivity ratio of the main input and reference input voice sensors in the direction in which a noise signal is input with a delay time difference equal to time. Reduction system.