JP3424366B2 - Silencer - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a muffler using active noise control.

[0002]

2. Description of the Related Art In recent years, an active noise control method has been proposed in which air conditioner noise, indoor noise of factories and automobiles, and the like are output by a speaker using a digital signal processing technique to mute the control noise.

A conventional silencer will be described below with reference to the drawings. FIG. 18 is a block diagram of a conventional silencer. In FIG. 18, 1, 2
Is a noise detector and an error detector, 3 is an adaptive filter, 4 is a speaker, 5 is a FIR filter which is a first digital filter, and 8 is an LMS calculator which is a coefficient calculator. is there.

The operation of the muffler having the above structure will be described below. First, noise is detected by the microphone 1, and the detection signal is input to the adaptive filter 3 and the FIR filter 5. Then, the noise signal processed by the adaptive filter 3 is output from the speaker 4. Then, in the microphone 2 installed near the ear of the listener, the reproduced sound from the speaker 4 interferes with the noise from the noise source, and the noise is attenuated by changing the coefficient of the adaptive filter 3.

Now, considering the coefficient update for the adaptive filter 3, the detection signal of the microphone 2 arranged near the ear of the listener is input to the LMS calculator 8, and the detection sound and the output of the FIR filter 5 cause the microphone. Two
The LMS calculation (least squares method) is performed so that the detection signal of 1 is minimized, and the coefficient of the adaptive filter 3 is updated. As a result, the speaker 4 in the microphone 2
Noise is attenuated by the control sound from. Here, the transfer function B (jω) from the speaker 4 to the microphone 2 is approximated to the FIR filter 5 in advance as a coefficient b (n).

This method is based on the Filtered-x LMS algorithm (see, for example, B. Widrowand S. Stearns,
`` Adaptive Signal Processing '' (Prentice-Hall, Engle
wood Cliffs, NJ, 1985)). Using this, the coefficient update of the adaptive filter 3 can be expressed by a mathematical expression as follows.

Since e (n) = y ^T (n) B (n) + X ^T (n) G1 (n),

[0008]

[Equation 1]

[0009] ^{Here, w T (n) = {} w0 (n), w1 (n), ..., wN-1 (n)} r (n) = x T (n) b (n) x T (n) = { x (n), x (n -1), ..., x (n-n + 1)} b T (n) = {b0 (n), b1 (n), ..., bN-1 (n)} y T ( n) = {y (n) , y (n-1), ..., y (n-n + 1)} y (n) = x T (n) w (n) , however, w (n); adaptive filter 3 Coefficient (the number of taps is N) α; step parameter r (n); output signal e (n) of FIR filter 5; output signal B (n) of microphone 2; transfer function G1 (n) from speaker 4 to microphone 2 If the transfer function from the noise source to the microphone 2 (Equation 1) approaches e (n) ≈0, the frequency characteristic W (k) of the coefficient w (n) of the adaptive filter 3 is {-G1 (k) / B (k)} is approximated.

The coefficient b (n) of the FIR filter 5 is obtained before the above calculation, which is performed by generating white noise from the measurement noise generator 12 as shown in FIG. The output of the FIR filter 5 and the detection signal of the microphone 2 are subtracted by the subtractor 11, and the result is used as an error signal for LMS calculation to obtain the coefficient.

By the above signal processing, noise is attenuated at the listener's ear position, but in reality (Fig. 2
In many cases, the speaker 4 and the microphone 2 cannot be installed near the listener or near the ear like 0).
In FIG. 20, the coefficient of the FIR filter 5 is c (n) according to the identification shown in FIG. 21, and the noise characteristic from the noise source to the microphone 2 is G0 (jω). The frequency characteristic W (k) of the coefficient w (n) of the filter 3 is approximated to {-G0 (k) / C (k)}. Therefore, as compared with the case of (FIG. 18), the adaptive filter 3
In the configuration of FIG. 20, the noise attenuation effect deteriorates at the listener's ear position even if the microphone 2 is muted.

Next, consider noise control at a plurality of positions.
(FIG. 22) shows a case where two speakers are used for two-point control. In FIG. 22, 1 is a microphone that is a noise detector, 2a-2b are microphones that are error detectors, 3a-3b are adaptive filters, 4a-4b are speakers, and 5a-5d are first. FIR filter, which is a digital filter, 8
a to 8d are LMS calculators, which are coefficient calculators, 1
3a to 13b are coefficient adders.

The operation of the silencer configured as above will be described below. First, noise is detected by the microphone 1, and the detection signal is input to the adaptive filters 3a to 3b and the FIR filters 5a to 5d. The noise signal processed by the adaptive filter 3a is output from the speaker 4a, and the noise signal processed by the adaptive filter 3b is output by the speaker 4b.
Is output from. Then, in the microphones 2a to 2b installed at the control positions, the reproduced sounds from the speakers 4a to 4b interfere with the noise from the noise source, and the noise is attenuated by changing the coefficients of the adaptive filters 3a to 3b.

Now, considering the coefficient update for the adaptive filter 3a, the detection signals of the microphones 2a-2b are input to the LMS calculators 8a-8b, respectively, and the detected sound and the outputs of the FIR filters 5a-5b cause the microphone 2a to be detected. And the LMS calculation is performed so that the detection signal of the microphone 2b is minimized, and the respective coefficients thereof are added by the coefficient adder 13a to update the coefficient of the adaptive filter 3a. Thereby, the microphones 2a-2
In b, the noise is attenuated by the control sound from the speaker 4a. Similarly, the control sound from the speaker 4b changes the coefficient of the adaptive filter 3b so as to attenuate the noise in the microphones 2a to 2b. Where FI
The R filter 5a includes a speaker 4a to a microphone 2a.
Is approximated as a coefficient c11 (n), and the transfer function C12 (jω) from the speaker 4a to the microphone 2b is coefficient c12 in the FIR filter 5b.
is approximated as (n), and the FIR filter 5c has a transfer function C21 from the speaker 4b to the microphone 2a.
(jω) is approximated as a coefficient c21 (n), and the transfer function C22 (jω) from the speaker 4b to the microphone 2b is approximated as a coefficient c22 (n) in the FIR filter 5d.

This method is called Multiple Error Filtered-x LM
S-algorithm (eg SJ Elliot as a reference)
t, IM Stothers and PA Nelson, ("A multiple er
rorLMS algorithm and its application to the active
control of sound and vibration. "IEEE Trans. Acous
t. Speech Signal Process. ASSP-35, pp1423-1434 (198
7))). If this is generally expressed by a mathematical expression, it can be expressed as follows. Now, assuming that there are m control speakers and k control point position microphones for one noise,

[0016]

[Equation 2]

[0017] Applying this to (Fig. 22),

[0018]

[Equation 3]

Therefore, w1 (n + 1) = w1 (n) + α {r11 (n) e1 (n) + r12 (n)
e2 (n)} w2 (n + 1) = w2 (n) + α {r21 (n) e1 (n) + r22 (n)
e2 (n)} where w1 (n); coefficient of adaptive filter 3a (the number of taps is N) w2 (n); coefficient of adaptive filter 3b (the number of taps is N) α; step parameter r11 (n); FIR filter 5a output signal r12 (n); FIR filter 5b output signal r21 (n); FIR filter 5c output signal r22 (n); FIR filter 5d output signal e1 (n); microphone 2a output signal e2 (n) ); Output signal of the microphone 2b As described above, noise control at a plurality of control points becomes possible by using this algorithm.

[0020]

However, as shown in FIG.
It is rare that the speaker 4 and the microphone 2 can be installed in the vicinity of the listener as in 8), and in practice, they are installed in a place apart from the listener as in (FIG. 20) and (FIG. 22). In particular, if the microphones 2, 2a-2b are far from the listener's ears, the effects are deteriorated at the ears, even if noise is sufficiently attenuated in the microphones 2, 2a-2b. Therefore, the audible effect becomes insufficient.

The present invention is intended to solve the above problems, and a first object thereof is to mute noise at an evaluation point even when the error detection microphone and the speaker are installed apart from the actual evaluation point. It is to suppress the deterioration of the effect. A second purpose is to suppress deterioration of the sound deadening effect when there are a plurality of evaluation points such as human ears, even if the evaluation points are located apart from the microphone and the speaker, as in the first purpose. That is. The third purpose is that, in addition to the second purpose, it can be controlled even in the presence of a plurality of noise sources. A fourth purpose is to reduce the amount of calculation in addition to the third purpose, and a fifth purpose is to apply the first purpose to periodic noise. A sixth objective is to apply the second objective to periodic noise, and a seventh objective is the sixth objective.
Is to achieve the objective of Stable and accurately. An eighth object is to obtain the same muffling effect as when the error detecting microphone is installed at the evaluation point even when the error detecting microphone and the speaker are installed apart from the actual evaluation point. A ninth purpose is to detect an error in the evaluation point even when the evaluation point is located apart from the microphone and the speaker, when there are a plurality of evaluation points such as human ears, like the eighth purpose. This is to obtain the same muffling effect as when a microphone is installed. The tenth purpose is to apply the eighth purpose to periodic noise. Furthermore, an eleventh object is to apply the ninth object to periodic noise, and a twelfth object is to provide a silencer capable of achieving the ninth object stably and accurately.

[0022]

In order to achieve the first object, a silencer of the first invention adapts a noise detector for detecting noise from a noise source, and noise from the noise detector. An adaptive filter for controlling, a first digital filter for signal processing the noise from the noise source, a speaker for reproducing the output of the adaptive filter, a second digital filter for signal processing the output of the adaptive filter, A third digital filter for signal-processing the output of the adaptive filter, an error detector installed away from the noise control position where the optimum damping effect by noise control is obtained, and an output from the error detector A subtractor for subtracting the output of the third digital filter, the output of the subtractor and the output of the second digital filter An adder for adding, and a coefficient calculator for updating by calculating the coefficients of the adaptive filter from the output of said adder and an output of said first digital filter.

In order to achieve the second object, the silencer of the second invention comprises a noise detector for detecting noise from a noise source, m adaptive filters, and (m × n) first filters. Digital filter, m speakers, (m ×
n) second digital filters, (m × n) third digital filters, and n errors installed apart from the noise control position where the optimum damping effect by noise control is obtained A detector, n subtractors, (m ×
n) adders, (m × n) coefficient calculators, and m coefficient adders, and the noise detector output is j (j = 1,
The second, ..., m) th adaptive filter and jk (k =
The first (1, 2, ..., N) -th digital filter is input, and the output of a certain j (= J) -th adaptive filter is the n-th Jk-th second digital filter and n.
Input to the Jk-th third digital filter,
The output of the J-th adaptive filter is reproduced by the J-th speaker, and the output of the K-th error detector is output from the output of the K-th third digital filter by the k-th (= K) -th subtractor. -1) J'K (J '
≠ J) The third digital filter output is subtracted, and the output is input to the m jK-th adders, and the JK-th adder outputs the K-th subtractor output and the JK-th second output. The output of the digital filter is added, and the JK-th coefficient calculator calculates the coefficient for the J-th adaptive filter at the K-th noise control point by the output of the JK-th first digital filter and the output of the JK-th adder. And the output of the JK-th coefficient calculator and the output of the (n−1) JK ′ (K ′ ≠ K) -th coefficient calculator are added by the J-th coefficient adder to obtain the J-th adaptive filter. Is configured to update the coefficient of.

In order to achieve the third object, the silencer of the third invention comprises l noise detectors for detecting noise from l noise sources, and (l × m) adaptive filters. , (L × m × n) first digital filters,
m speakers, m control signal adders, (l × m
× n) second digital filters, (m × n) third digital filters, and n number of digital filters installed far from the noise control position where the optimum damping effect by noise control is obtained. Error detector, n subtractors, (l
Xm × n) adders, (l × m × n) coefficient calculators, and (l × m) coefficient adders, and i (i = 1,
Of the 2, ..., L) th noise detectors, the output of a certain i (= I) th noise detector is m Ij (j = 1, 2 ,.
m) th adaptive filter and (m × n) Ij
It is input to the k (k = 1, 2, ..., N) -th first digital filter, and the output of a certain Ij (= IJ) -th adaptive filter is combined with the n IJk-th second digital filters. It is inputted to the J-th control signal adder, and the output of the IJ-th adaptive filter and (l-1) I'J (I 'are inputted by the J-th control signal adder.
≠ I) Add the outputs of the adaptive filter and add J
The output of the th control signal adder is input to the n Jk th third digital filters and the J th speaker, and is output from the output of the K th error detector to JK by a certain k (= K) subtractor. The output of the th-th third digital filter and the output of the (m−1) th J′K (J ′ ≠ K) -th third digital filter are subtracted, and the output is (l × m) ijK The output of the Kth subtractor and the output of the IJKth second digital filter are added by the IJKth adder, and the IJKth coefficient calculator outputs the IJKth first digital signal. Filter output and I
By the output of the JK-th adder, I for the I-th noise by the J-th speaker at the K-th noise control point
The coefficient of the J-th adaptive filter is calculated, and the output of the IJK-th coefficient calculator and the output of the (n-1) IJK '(K' ≠ K) -th coefficient calculator are added by the IJ-th coefficient adder. Then, the coefficient of the IJ-th adaptive filter is updated.

In order to achieve the fourth object, the silencer of the fourth invention comprises l noise detectors for detecting noise from l noise sources, and (l × m) adaptive filters. , (L × m × n) first digital filters,
m speakers, m control signal adders, (m ×
n) second digital filters, (m × n) third digital filters, and n errors installed apart from the noise control position where the optimum damping effect by noise control is obtained A detector, n subtractors, (m ×
n) adders, (l × m × n) coefficient calculators,
With (l × m) coefficient adders, i (i = 1, 2,
The output of the i (= I) th noise detector among the (l) th noise detectors is m Ij (j = 1, 2, ..., M).
Th adaptive filter and (m × n) Ijk
The (k = 1, 2, ..., N) -th first digital filter is input, and the output of a certain Ij (= IJ) -th adaptive filter is input to the J-th control signal adder. Output of the IJ-th adaptive filter and (l-1) I'J
The outputs of the (I ′ ≠ I) -th adaptive filters are added, and the outputs of the J-th control signal adder are n Jk-th second digital filters and n Jk-th third digital filters. The signal is input to the Jth speaker and is output from the output of the Kth error detector by the k (= K) th subtractor to the output of the JKth third digital filter and (m-1) J'K ( The output of the (J ′ ≠ K) th third digital filter is subtracted, and the output is input to the m jKth adders, and the JKth adder outputs the output of the Kth subtractor and the JKth adder. The output of the second digital filter is added and the output is input to l iJKth coefficient calculators. The IJKth coefficient calculator is the output of the IJKth first digital filter and the JKth adder. Output of the Kth J-th calculated the coefficients of the IJ th adaptive filter for the I-th noise by the speaker, IJ th by a factor adder and the output of IJK th coefficient calculator (n-1) pieces of IJ in the sound control point
The output of the K '(K' ≠ K) th coefficient calculator is added to obtain IJ
It is configured to update the coefficients of the th adaptive filter.

In order to achieve the fifth object, the silencer of the fifth invention is an adaptive filter for adaptively controlling noise, a first digital filter for similarly signal-processing noise, and a reproduction of the output of the adaptive filter. Speaker, a second digital filter for signal-processing the output of the adaptive filter, a third digital filter for similarly signal-processing the output of the adaptive filter, and a noise control position where an optimum damping effect by noise control is obtained. From the output of the error detector and the output of the third digital filter, and the output is used as the noise input signal of the adaptive filter and the first digital filter. A subtractor, adding the output of the subtractor and the output of the second digital filter An adder for, and a coefficient calculator for updating by calculating the coefficients of the adaptive filter from the output of said adder and an output of said first digital filter.

In order to achieve the sixth object, the silencer of the sixth invention comprises m adaptive filters and (m ×
n) first digital filters, m speakers, (m × n) second digital filters, and (m)
× n) third digital filters, n error detectors installed apart from the noise control position where the optimum damping effect is obtained by noise control, n subtractors, and (m Xn) adders, (m × n) coefficient calculators, and m coefficient adders, and the noise signal is j (j =
1, 2, ..., M) th adaptive filter and jk
The signal is processed by the (k = 1, 2, ..., N) -th first digital filter, and the output of a certain j (= J) -th adaptive filter is the n-th Jk-th second digital filter and n. Are input to the Jk-th third digital filter, and the output of the J-th adaptive filter is reproduced by the J-th speaker, and a certain k (=
The (K) th subtracter subtracts the output of the Kth error detector from the output of the JKth third digital filter and (m−
1) The outputs of the J'K (J '≠ J) th third digital filters are subtracted, and the outputs are used as noise signals of the Jth adaptive filter and n Jkth first digital filters. Further, the output of the Kth subtractor is input to m jKth adders, and the output of the Kth subtractor and the output of the JKth second digital filter are added by the JKth adder. Then, the JKth coefficient calculator outputs the output of the JKth first digital filter and the JKth coefficient.
The coefficient of the J-th adaptive filter at the K-th noise control point is obtained from the output of the n-th adder, and the output of the JK-th coefficient calculator and (n-1) JK 'are calculated by the J-th coefficient adder. The output of the (K ′ ≠ K) th coefficient calculator is added to update the coefficient of the Jth adaptive filter.

In order to achieve the seventh object, the silencer of the seventh invention comprises (m × n) adaptive filters,
(M × n) first digital filters, m speakers, (m × n) second digital filters, (m × n) third digital filters, and noise control , N error detectors, n subtractors, (m × n) adders, and (m × n) The number of coefficient calculators and the number of control signal adders are m, and the noise signal is jk (j = 1, 2, ..., M 2; k = 1, 2, ...
Signal processing is performed by the th adaptive filter and the jk first digital filter, and a certain jk (= J
The output of the (K) -th adaptive filter and the output of the (n-1) JK '(K' ≠ K) -th adaptive filter are added by the J-th control signal adder, and the output of the J-th control signal adder is added. The output is input to the n-th Jk-th second digital filter and the n-th Jk-th third digital filter, and the output of the J-th control signal adder is reproduced by the J-th speaker and K-th. Of the output of the Kth error detector and the output of the JKth third digital filter and (m-1) J'K (J '≠
The output of the (J) th third digital filter is subtracted, and the output is converted into m jKth adaptive filters and m
Noise signals from the jKth first digital filters, and the output of the Kth subtractor is m jK
Is input to the th adder and K is added by the JK th adder.
The output of the second subtractor and the output of the JK-th second digital filter are added, and the JK-th coefficient calculator outputs the output of the JK-th first digital filter and the output of the JK-th adder to the J-th It is configured to find and update the coefficients of the adaptive filter.

Similarly, in order to achieve the seventh object, the silencer of the eighth invention comprises (m × n) adaptive filters, (m × n) first digital filters, and m.
Speakers, (m × n) second digital filters, (m × n) third digital filters,
N error detectors installed apart from the noise control position where the optimum damping effect by noise control is obtained,
n subtractors, (m × n) adders, and (m × n)
The number of coefficient calculators and the number of control signal adders are m, and the noise signals are jk (j = 1, 2, ..., M 2; k = 1, 2 ,.
The n) -th adaptive filter and the jk-th first digital filter respectively perform signal processing to obtain a certain jk
The output of the (= JK) th adaptive filter is input to the JKth second digital filter and the Jth control signal adder, and the output of the JKth adaptive filter and (n- 1) J
The outputs of the K ′ (K ′ ≠ K) th adaptive filter are added, the output of the Jth control signal adder is input to the n Jkth third digital filters, and the Jth control signal addition is performed. The output of the filter is reproduced by the Jth speaker, and the output of the KKth error detector to the output of the JKth third digital filter and (m-1) J'K (J The output of the '≠ J) th third digital filter is subtracted, and the output is input as noise signals of the m jK-th adaptive filters and the m jK-th first digital filters, and further the K-th The output of the subtractor is input to m jKth adders,
The output of the Kth subtractor and JK by the JKth adder
Add the outputs of the second digital filter of the th
The th coefficient calculator is configured to obtain and update the coefficient of the Jth adaptive filter by the output of the JKth first digital filter and the output of the JKth adder.

In order to achieve the eighth object, the silencer of the ninth invention is a noise detector for detecting noise from a noise source, an adaptive filter for adaptively controlling noise from the noise detector, and A first digital filter that processes the noise from the noise source, a speaker that reproduces the output of the adaptive filter, a second digital filter that processes the output of the adaptive filter, and an output of the adaptive filter. A third digital filter for processing, an error detector installed far from the noise control position where the optimum damping effect by noise control is obtained, and a fourth digital filter for processing the output of the error detector. And a subtractor for subtracting the output of the third digital filter from the output of the fourth digital filter , Wherein the output of the subtracter second
Of the digital filter, and a coefficient calculator for calculating and updating the coefficient of the adaptive filter from the output of the first digital filter and the output of the adder.

In order to achieve the ninth object, the silencer of the tenth invention comprises a noise detector for detecting noise from a noise source, m adaptive filters, and (m × n) first filters. Digital filter, m speakers, (m ×
n) second digital filters, (m × n) third digital filters, and n errors installed apart from the noise control position where the optimum damping effect by noise control is obtained Detector, n fourth digital filters, n subtractors, (m × n) adders, (m × n) coefficient calculators, and m coefficient adders , And the output of the noise detector is the j (j = 1, 2, ..., M) th adaptive filter and jk (k = 1, 2 ,.
It is input to the (n) -th first digital filter, and the output of a certain j (= J) -th adaptive filter is output to the n-th Jk-th second digital filter and the n-th Jk-th third digital filter. Also, the output of the J-th adaptive filter is reproduced by the J-th speaker, and the output of a certain k (= K) -th error detector is K.
The signal is processed by the fourth digital filter of the th, and the output of the fourth digital filter of the Kth is output by the third digital filter of the Kth by the subtractor of the Kth and (m−1) J′K The output of the (J '≠ J) th third digital filter is subtracted, and the output is input to the m jK-th adders, and the JK-th adder outputs the output of the K-th subtractor and the JK-th Output of the second digital filter is added, and the JK-th coefficient calculator outputs the output of the JK-th first digital filter and the output of the JK-th adder to the J-th adaptive filter at the K-th noise control point. And the output of the JK-th coefficient calculator and (n-1)
The outputs of the JK ′ (K ′ ≠ K) th coefficient calculators are added to update the coefficient of the Jth adaptive filter.

In order to achieve the tenth object, the silencer of the eleventh invention is an adaptive filter for adaptively controlling noise, a first digital filter for signal processing noise, and a reproduction of the output of the adaptive filter. A speaker, a second digital filter for signal processing the output of the adaptive filter, and a third digital filter for signal processing the output of the adaptive filter,
Of the error detector installed at a place distant from the noise control position for obtaining the optimum damping effect by the noise control, the fourth digital filter for signal processing the output of the error detector, and the fourth digital filter. A subtracter that subtracts the output of the third digital filter from the output and uses the output as a noise input signal of the adaptive filter and the first digital filter, and the output of the subtractor and the output of the second digital filter. An adder for adding
It is composed of a coefficient calculator for calculating and updating the coefficient of the adaptive filter from the output of the first digital filter and the output of the adder.

In order to achieve the eleventh object, a silencer of the twelfth invention comprises m adaptive filters and (m
× n) first digital filters, m speakers, and (m × n) second digital filters,
(M × n) third digital filters, n error detectors installed apart from the noise control position where the optimum damping effect by noise control is obtained, and n fourth digital filters Filter, n subtractors, (m ×
n) adders, (m × n) coefficient calculators, and m coefficient adders, and the noise signal is j (j = 1, 2, ...).
m) th adaptive filter and jk (k = 1, 2,
, N) -th first digital filter, and the output of a certain j (= J) -th adaptive filter is the n-th Jk-th second digital filter and the n-th Jk-th third digital filter. , The output of the Jth adaptive filter is reproduced by the Jth speaker, and the output of a certain k (= K) th error detector is signal-processed by the Kth fourth digital filter. , The output of the JK-th third digital filter from the output of the K-th fourth digital filter and the (m-1) J'K (J '≠ J) -th third output from the output of the K-th fourth digital filter. The output of the digital filter is subtracted, the output is input as the noise signal of the Jth adaptive filter and the n Jkth first digital filters, and the Kth subtraction is performed. The output of is input to the m jK-th adders, and the JK-th adder adds the output of the K-th subtractor and the output of the JK-th second digital filter. The coefficient of the Jth adaptive filter at the Kth noise control point is obtained from the output of the JKth first digital filter and the output of the JKth adder, and the JKth coefficient calculator is used by the Jth coefficient adder. Output and (n-1) JK '
The output of the (K ′ ≠ K) th coefficient calculator is added to update the coefficient of the Jth adaptive filter.

In order to achieve the twelfth object, the silencer of the thirteenth invention comprises (m × n) adaptive filters, (m × n) first digital filters, and m speakers. And (m × n) second digital filters, (m × n) third digital filters, and are installed away from the noise control position where the optimum damping effect by noise control is obtained. N error detectors, n
Fourth digital filters, n subtractors,
It has (m × n) adders, (m × n) coefficient calculators, and m control signal adders, and the noise signal is jk (j
= 1, 2, ..., M; k = 1, 2, ..., N) Adaptive filter and jk-th first digital filter perform signal processing respectively, and output of a certain jk (= JK) -adaptive filter And (n-1) JK '(K'
≠ K) The output of the adaptive filter is added by the J-th control signal adder, and the output of the J-th control signal adder is n Jk-th second digital filters and n Jk-th second digital filters. 3, the output of the J-th control signal adder is reproduced by the J-th speaker, and the output of the K-th error detector is signal-processed by the K-th fourth digital filter. From the output of the Kth fourth digital filter to the output of the JKth third digital filter and (m-1) J'K (J '≠ J) th third digital filter Is subtracted, the output is input as noise signals of the m jKth adaptive filters and the m jKth first digital filters, and the output of the Kth subtractor is Is inputted to the number of jK th adder,
The output of the Kth subtractor and JK by the JKth adder
Add the outputs of the second digital filter of the th
The th coefficient calculator is configured to obtain and update the coefficient of the Jth adaptive filter by the output of the JKth first digital filter and the output of the JKth adder.

Also in order to achieve the twelfth object, the first
The silencer of the invention of 4 is (m × n) adaptive filters, (m × n) first digital filters, m speakers, and (m × n) second digital filters. A filter, (m × n) third digital filters, n error detectors installed away from the noise control position where the optimum damping effect by noise control is obtained, and n number of error detectors. It has 4 digital filters, n subtractors, (m × n) adders, (m × n) coefficient calculators, and m control signal adders. jk
(J = 1, 2, ..., M; k = 1, 2, ..., N) -th adaptive filter and the jk-th first digital filter respectively perform signal processing, and a certain jk (= JK) -th adaptive filter. Is input to the JK-th second digital filter and the J-th control signal adder, and the J-th control signal adder outputs the output of the JK-th adaptive filter and (n-1) JK '(K '≠
The outputs of the (K) -th adaptive filter are added, and J
The output of the n-th control signal adder is input to the n Jk-th third digital filters, and the output of the J-th control signal adder is reproduced by the J-th speaker and the K-th error detector outputs. The output is signal-processed by the Kth fourth digital filter, and the output of the Kth fourth digital filter to the output of the JKth third digital filter and (m-1) J'K (J '
≠ J) The output of the third digital filter is subtracted, and the output is input as noise signals of the m jK-th adaptive filters and the m jK-th first digital filters, and the K-th The output of the subtracter is m j
It is input to the Kth adder, and the output of the Kth subtractor and the output of the JKth second digital filter are added by the JKth adder, and the JKth coefficient calculator outputs the JKth first The output of the digital filter and the output of the JK-th adder are used to obtain and update the coefficient of the J-th adaptive filter.

[0036]

According to the structure of the first invention, even if the error detector is installed at a position distant from the actual evaluation point, the adaptive filter can be operated by the correction of the control signal by the second and third digital filters. Since the coefficient is updated so as to control the evaluation point, noise can be sufficiently attenuated at the evaluation point distant from the error detector.

In the second invention, since the noise control is performed by using the plurality of speakers and the error detector, the effect of the first invention can be obtained at a plurality of evaluation points.

Since the configuration of the third invention is such that the configuration of the second invention can be controlled for a plurality of noise sources, the same configuration as the second invention can be applied to a plurality of noise sources. The effect can be obtained.

According to the fourth aspect of the invention, the second digital filter of the third aspect of the invention is configured to perform signal processing on the output from the control signal adder, thereby reducing the amount of calculation and the third aspect of the invention. The same effect can be obtained.

In the configuration of the fifth aspect of the invention, the hardware can be reduced by using the output of the subtractor as the input signal of the adaptive filter and the first digital filter, and the error can be corrected by the correction of the control signal by the third digital filter. The periodic noise can be sufficiently attenuated at the evaluation point distant from the detector.

According to the sixth aspect of the invention, the output of the subtractor is used as the input signal of the adaptive filter and the first digital filter, and noise control is performed using a plurality of speakers and an error detector. The effect of the invention of 5 can be obtained at a plurality of evaluation points.

In the seventh invention, the output of the subtracter for outputting the noise signal detected by the error detector controlled by the adaptive filter and the first digital filter is used as the input signal of the adaptive filter and the first digital filter,
Moreover, since the noise control is performed by using the plurality of speakers and the error detector, the effect of the sixth invention can be obtained more stably.

In the eighth aspect of the invention, since the error signal of each LMS calculator does not include the output of the adaptive filter other than the output of the adaptive filter required by the LMS calculator, each LMS calculator can efficiently calculate.

According to the ninth aspect of the invention, even if the error detector is installed at a position distant from the actual evaluation point, the error detection signal is corrected by the fourth digital filter,
The noise signal at the actual evaluation point can be detected, and the adaptive filter is updated by the correction of the control signal by the second and third digital filters so as to control the actual evaluation point. At the evaluation points, it is possible to obtain the same effect as if the error detector is installed at the evaluation points.

In the tenth aspect of the invention, since the noise control is performed by using a plurality of speakers and an error detector,
The effect of the ninth invention can be obtained at a plurality of evaluation points.

In the eleventh aspect of the invention, since the output of the subtractor is used as the input signal of the adaptive filter and the first digital filter, the hardware can be reduced, and the error detection signal can be corrected by the fourth digital filter. By correcting the control signal by the third digital filter, it is possible to obtain the same effect as when the error detector is installed at the evaluation point at the evaluation point distant from the error detector.

In the twelfth aspect of the invention, the output of the subtractor is used as the input signals of the adaptive filter and the first digital filter, and noise control is performed using a plurality of speakers and an error detector. The effects of the eleventh invention can be obtained at a plurality of evaluation points.

In the thirteenth invention, the output of the subtracter for outputting the noise signal detected by the error detector controlled by the adaptive filter and the first digital filter is used as the input signal of the adaptive filter and the first digital filter, Moreover, since the noise control is performed by using the plurality of speakers and the error detector, the effect of the twelfth invention can be obtained more stably.

In the fourteenth invention, since the error signal of each LMS calculator does not include the output of the adaptive filter other than the output of the adaptive filter required by the LMS calculator, each LMS calculator can efficiently calculate.

[0050]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the first invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of a silencer according to an embodiment of the first invention. In FIG. 1, 1 is a microphone, which is a noise detector, and 2
Is a microphone which is an error detector, 3 is an adaptive filter, 4 is a speaker, 5 is an FIR filter which is a first digital filter, 6 is an FIR filter which is a second digital filter, and 7 is a 3 is a FIR filter which is a digital filter, 8 is an LMS calculator which is a coefficient calculator, 9 is an adder, and 10 is a subtracter.

The operation of the silencer configured as above will be described below. First, noise is detected by the microphone 1, and the detection signal is input to the adaptive filter 3 and the FIR filter 5. Then, the noise signal processed by the adaptive filter 3 is FIR
It is input to the filter 6, the FIR filter 7, and the speaker 4. Then, in the microphone 2 installed at a position away from the actual evaluation point, the reproduced sound from the speaker 4 and the noise from the noise source are detected.

The detection signal of the microphone 2 is supplied to the subtractor 10
Is subtracted from the output signal of the FIR filter 7. Here, in the FIR filter 7, the transfer function C (jω) from the speaker 4 to the microphone 2 is identified in advance as a coefficient c (n). Therefore, the output component of the adaptive filter 3 is removed from the output of the subtractor 10, and as a result, only the noise signal detected by the microphone 2 is obtained. Then, the output of the subtractor 10 is added to the output of the FIR filter 6 in the adder 9 and input to the LMS calculator 8. Here, the transfer function B (jω) from the speaker 4 to the actual evaluation point (for example, the ear of the listener) is identified in the FIR filter 6 in advance as a coefficient b (n). Therefore, the output of the adder 9 becomes approximately equal to the detection signal of the microphone installed at the actual evaluation point.

The LMS calculator 8 updates the coefficient of the adaptive filter 3 by the outputs of the FIR filter 5 and the adder 9 so that the output of the adder 9 becomes the minimum. If this is expressed by a mathematical formula, it can be expressed as follows.

[0055] e3 (n) = y3 (n ) + e2 (n) = y1 T (n) b (n) + e1 (n) -y2 (n) = y1 T (n) b (n) + X T (n) ^{G0 (n) + y1 T (} n) C (n) -y1 T (n) c (n) here, if the C (n) ≒ c (n ), e3 (n) = y1 T (n) b ( n) + x ^T (n) G0
(N), so

[0056]

[Equation 4]

However, w (n); coefficient α of the adaptive filter 3; step parameter r (n); output signal e3 (n) of the FIR filter 5; output signal e3 (n) of the adder 9 (Equation 4) When approaching ≈0, the frequency characteristic W (k) of the coefficient w (n) of the adaptive filter 3 is approximated to {-G0 (k) / B (k)}. Therefore, (Fig. 1
8) and (FIG. 20), the coefficient W (k) of the adaptive filter 3 is {-G1 (k) /
Only the numerator is different from B (k)}, and the difference is smaller than the case where the denominator and the numerator in (FIG. 20) are different. From this, the effect at the actual evaluation point is not affected by the transfer function C (jω), but only by the noise characteristic. When the noise contains a lot of low frequency components, deterioration of the silencing effect at the evaluation point is suppressed, and particularly in the case of low frequency noise in which the difference in distance between the speaker 4 and the microphone 2 can be ignored, the effect is almost the same as that of (FIG. 18). The effect can be obtained.

Now, the coefficients b (n) of the FIR filters 5 to 7,
c (n) is calculated before performing the above calculation, but FIR filter 7
The coefficient c (n) of is obtained in the same manner as in the case of (FIG. 21), and the coefficient b (n) of the FIR filters 5 to 6 is obtained in the same manner as in the case of (FIG. 19). However, the FIR filter 5 may be obtained with the configuration of (FIG. 2).

Next, an embodiment of the second invention will be described with reference to the drawings. FIG. 3 is a block diagram of a silencer according to an embodiment of the second aspect of the invention, in which a plurality of control points are present. In FIG. 3, 1 is a microphone, which is a noise detector, and 2
a to 2b are microphones, which are error detectors,
a to 3b are adaptive filters, 4a to 4b are speakers, 5a to 5d are FIR filters which are first digital filters, 6a to 6d are FIR filters which are second digital filters, and 7a to 7d are first filters. FIR filter which is a digital filter of No. 3, LMS calculators 8a to 8d which are coefficient calculators, 9a to 9d adders, 10a to 10d subtractors, 1
3a to 13b are coefficient adders.

The operation of the silencer configured as above will be described below. First, noise is detected by the microphone 1, and the detection signal is input to the adaptive filters 3a to 3b and the FIR filters 5a to 5d. The noise signal processed by the adaptive filter 3a is input to the FIR filters 6a to 6b, the FIR filters 7a to 7b and the speaker 4a, and the noise signal processed by the adaptive filter 3b is FIR filters 6c to 6d and the FIR filter. 7c to 7d and speaker 4
Input to b. Then, the microphones 2a to 2b installed at the control positions detect the reproduced sound from the speakers 4a to 4b and the noise from the noise source. The detection signal of the microphone 2a is supplied to the FIR filter 7a in the subtractor 10a.
Is subtracted from the output signal of. Here, the FIR filter 7a
In advance, the transfer function C11 (jω) from the speaker 4a to the microphone 2a is identified in advance as the coefficient c11 (n). Therefore, the output component of the adaptive filter 3a is removed from the output of the subtractor 10a.

Then, the output is F in the subtractor 10c.
It is subtracted from the output signal of the IR filter 7c. Where F
In the IR filter 7c, the transfer function C21 (jω) from the speaker 4b to the microphone 2a is identified in advance as a coefficient c21 (n). Therefore, the output component of the adaptive filter 3b is removed from the output of the subtractor 10c, resulting in only the noise signal detected by the microphone 2a.

Similarly, the output components of the adaptive filters 3a to 3b are removed from the output of the subtractor 10b, leaving only the noise signal detected by the microphone 2b. Then, the output of the subtractor 10c is added to the output of the FIR filter 6a in the adder 9a and input to the LMS calculator 8a. Here, in the FIR filter 6a, the transfer function B11 (jω) from the speaker 4a to the actual evaluation point (for example, the right ear of the listener) is identified in advance as a coefficient b11 (n). Therefore, the output of the adder 9a becomes approximately equal to the detection signal of the microphone installed at the right ear which is the actual evaluation point.

Similarly, the outputs of the adders 9b and 9d are approximately equal to the detection signal of the microphone installed in the left ear, and the output of the adder 9c is approximately equal to the detection signal of the microphone installed in the right ear. LMS
The computing unit 8a calculates the coefficient of the adaptive filter 3a based on the outputs of the FIR filter 5a and the adder 9a so that the output of the adder 9a is minimized. Therefore, the LMS calculator 8a obtains a coefficient when the adaptive filter 3a controls the right ear by using the speaker 4a.

Similarly, the LMS calculator 8b calculates the coefficient of the adaptive filter 3a from the outputs of the FIR filter 5b and the adder 9b so that the output of the adder 9b becomes the minimum, and the coefficient is calculated by the adaptive filter 3a. 4a is used to control the left ear region. Then, in the coefficient adder 13a, the coefficient calculated by the LMS calculator 8a and the coefficient calculated by the LMS calculator 8b are added,
The coefficient of the adaptive filter 3a is updated. The same applies to the adaptive filter 3b.

From the above, as in the case of (FIG. 1), even if the adaptive filters 3a to 3b are installed at positions apart from the evaluation points by the speakers 4a to 4b and the microphones 2a to 2b, it is as if the evaluation points Since it operates as if it were controlled, it is possible to suppress the deterioration of the silencing effect at the evaluation point. Further, for example, focusing on the LMS calculator 8a, only the noise signal detected by the microphone 2a and the output signal of the adaptive filter 3a processed by the FIR filter 6a are output to the adder 9a which is an error input of the LMS calculation. Since it is included, LMS calculation is efficient,
It is also executed with high accuracy.

In this embodiment, the number of error detecting microphones and the number of speakers are two. However, the number may be increased according to the application, and in that case, the algorithm may be naturally expanded for the increased amount. In addition, the subtractor 10a and the subtractor 10
c together and output from the microphone 2a to FIR
The outputs of the filters 7a and 7c are subtracted, and the subtracter 10
b and the subtractor 10d are combined into a microphone 2b.
The outputs of the FIR filters 7b and 7d may be subtracted from the output of.

Next, an embodiment of the third invention will be described with reference to the drawings. FIG. 4 is a block diagram of a silencer according to an embodiment of the third aspect of the present invention, in which a plurality of noises are controlled by a plurality of control points.
In FIG. 4, 1a to 1b are microphones that are noise detectors, 2a to 2b are microphones that are error detectors, 3a to 3d are adaptive filters, 4a to 4b are speakers, and 5a to 5h are first microphones. FIR filters 6a to 6 which are digital filters of No. 1
h is a FIR filter that is a second digital filter, 7a to 7d are FIR filters that are third digital filters, 8a to 8h are LMS calculators that are coefficient calculators, and 9a to 9h are additions. Bowl, 10a
-10d is a subtracter, 13a-13d is a coefficient adder, 14
Reference symbols a to 14b are control signal adders.

The operation of the muffler having the above structure will be described below. First, noise is detected by each of the microphones 1a and 1b, and the detection signals are detected by the adaptive filters 3a to 3d and the FIR filter 5.
a to 5h, respectively. First, consider the control by the speaker 4a. The noise signal processed by the adaptive filter 3a is input to the FIR filters 6a to 6b, and the noise signal processed by the adaptive filter 3c is input to the FIR filters 6e to 6f. The outputs of the adaptive filters 3a and 3c are added by the control signal adder 14a. And the output is FIR
It is input to the filters 7a-7b and the speaker 4a. Then, the microphones 2a to 2b installed at the control positions detect the reproduced sound from the speaker 4a and the noise from the noise source. The detection signal of the microphone 2a is subtracted from the output signal of the FIR filter 7a in the subtractor 10a. Here, in the FIR filter 7a, the transfer function C11 (jω) from the speaker 4a to the microphone 2a has a coefficient c11 in advance.
Identified as (n). Therefore, the output components of the adaptive filters 3a and 3c are removed from the output of the subtractor 10a.

Then, the output is F in the subtractor 10c.
It is subtracted from the output signal of the IR filter 7c. Where F
In the IR filter 7c, the transfer function C21 (jω) from the speaker 4b to the microphone 2a is identified in advance as a coefficient c21 (n). Therefore, the output components of the adaptive filters 3b and 3d are removed from the output of the subtractor 10c, resulting in only the noise signal detected by the microphone 2a. Similarly, the output of the subtractor 10b is only the noise signal detected by the microphone 2b. Then, the output of the subtractor 10c is added to the outputs of the FIR filters 6a and 6e in the adders 9a and 9e, and is input to the LMS calculators 8a and 8e. Here, the FIR filter 6
In a and 6e, the transfer function B11 (jω) from the speaker 4a to the actual evaluation point (for example, the right ear of the listener) is identified in advance as a coefficient b11 (n). Therefore, the adder 9a,
The output of 9e is approximately equal to the detection signal of the microphone installed at the right ear, which is the actual evaluation point.

Similarly, the outputs of the adders 9b and 9f are approximately equal to the detection signal of the microphone installed near the left ear. The LMS calculator 8a uses the outputs of the FIR filter 5a and the adder 9a to calculate the coefficient of the adaptive filter 3a so that the output of the adder 9a is minimized. Therefore, in the LMS calculator 8a, the adaptive filter 3a
Calculates the coefficient when the right ear is controlled using the speaker 4a.

Similarly, the LMS calculator 8b calculates the coefficient of the adaptive filter 3a by the outputs of the FIR filter 5b and the adder 9b so that the output of the adder 9b becomes the minimum, and the coefficient is calculated by the adaptive filter 3a. 4a is used to control the left ear region. Then, the coefficient adder 13a adds the coefficient calculated by the LMS calculator 8a and the coefficient calculated by the LMS calculator 8b to update the coefficient of the adaptive filter 3a. Similarly, the adaptive filter 3c also operates to control the ears with the speaker 4a.

The above is the control by the speaker 4a. By operating the adaptive filters 3b, 3d in the same manner, the control of the ears by the speaker 4b becomes possible.

Therefore, even when there are a plurality of noise sources, the adaptive filters 3a to 3d are connected to the speaker 4a.
4b and the microphones 2a to 2b are installed apart from the evaluation points, they operate as if they were controlling the evaluation points, so that it is possible to suppress deterioration of the silencing effect at the evaluation points. Further, for example, focusing on the LMS calculator 8a, the noise signal detected by the microphone 2a and the FI are output to the output of the adder 9a which is an error input of the LMS calculation.
Since only the output signal of the adaptive filter 3a that has been signal-processed by the R filter 6a is included, the LMS calculation is executed efficiently and accurately.

In the present embodiment, the case where there are two noise sources has been described as an example, but when the number of noise sources increases more than that, the algorithm is expanded by that number and each filter is expanded. The outputs may be combined into the number of speakers by the control signal adder. In addition, the subtractor 10a and the subtractor 10
c into one, and the FI is output from the microphone 2a.
The outputs of the R filters 7a and 7c are subtracted, and the subtracter 1
0b and subtractor 10d are combined into one microphone 2
The configuration may be such that the outputs of the FIR filters 7b and 7d are subtracted from the output of b.

Next, an embodiment of the fourth invention will be described with reference to the drawings. FIG. 5 is a block diagram of a silencer according to an embodiment of the fourth aspect of the present invention, in which a plurality of noises are controlled by a plurality of control points.
In FIG. 5, 1a to 1b are microphones that are noise detectors, 2a to 2b are microphones that are error detectors, 3a to 3d are adaptive filters, 4a to 4b are speakers, and 5a to 5h are first. FIR filters 6a to 6 which are digital filters of No. 1
d is a FIR filter that is a second digital filter, 7a to 7d are FIR filters that are third digital filters, 8a to 8h are LMS calculators that are coefficient calculators, and 9a to 9d are additions. Bowl, 10a
-10d is a subtracter, 13a-13d is a coefficient adder, 14
Reference symbols a to 14b are control signal adders.

In the configuration of FIG. 5, the control signal adder 14
The number of FIR filters is halved by using the outputs of a to 14b as the inputs of the FIR filters 6a to 6d. Therefore, the error signals of the LMS calculators 8a to 8h include more signals than in the case of (FIG. 4), but there is no problem in stability because it is the same as the original Multiple Error Filtered-x LMS algorithm. .

Therefore, even in this configuration, the adaptive filters 3a to 3d are the same as the speakers 4a to 4b and the microphone 2.
Even if a to 2b are installed apart from the evaluation points, they operate as if they were controlling the evaluation points, so that it is possible to suppress deterioration of the silencing effect at the evaluation points. Further, since the number of FIR filters can be halved, the amount of calculation can be reduced and the hardware can be downsized.

It should be noted that the subtractor 10a and the subtractor 10c are combined into one, the outputs of the FIR filters 7a and 7c are subtracted from the output of the microphone 2a, and similarly the subtractor 10b and the subtractor 10d are combined into one. , The outputs of the FIR filters 7b and 7d may be subtracted from the output of the microphone 2b.

Next, an embodiment of the fifth invention will be described with reference to the drawings. FIG. 6 is a block diagram of a silencer according to an embodiment of the fifth aspect of the present invention, in which periodic noise is controlled. In FIG. 6, 2 is a microphone, which is an error detector, 3
Is an adaptive filter, 4 is a speaker, 5 is an FIR filter which is a first digital filter, 6 is an FIR filter which is a second digital filter, and 7 is an FI which is a third digital filter.
R filter, 8 is an LMS calculator which is a coefficient calculator, 9 is an adder, and 10 is a subtractor.

The operation of the silencer configured as above will be described below. First, since the microphone 2 serves both as a noise detector and an error detector, the microphone 2 detects noise. The detected noise signal is subtracted by the subtractor 10 from the output of the FIR filter 7. Here, since the transfer function C (jω) from the speaker 4 to the microphone 2 is identified as the coefficient c (n) in the FIR filter 7 in advance, the output of the subtractor 10 is the same as in the case of (FIG. 1). Therefore, the output component of the adaptive filter 3 is removed, resulting in only the noise signal detected by the microphone 2. Then, this noise signal is input to the adaptive filter 3 and the FIR filter 5. The noise signal processed by the adaptive filter 3 is input to the FIR filter 6, the FIR filter 7, and the speaker 4. Then, in the microphone 2 installed at a position away from the actual evaluation point, the speaker 4
The reproduced sound from the sound source and the noise from the noise source interfere with each other.

On the other hand, the output of the subtractor 10 is added to the output of the FIR filter 6 in the adder 9 and input to the LMS calculator 8. Here, in the FIR filter 6,
The transfer function B (jω) from the speaker 4 to the actual evaluation point (for example, the ear of the listener) is previously identified as the coefficient b (n). Therefore, the output of the adder 9 becomes approximately equal to the detection signal of the microphone installed at the actual evaluation point. The LMS calculator 8 updates the coefficient of the adaptive filter 3 based on the outputs of the FIR filter 5 and the adder 9 so that the output of the adder 9 is minimized. Therefore, the periodic noise is sufficiently attenuated at the actual evaluation point. If this is expressed by a mathematical formula, it can be expressed as follows.

X (n) = e2 (n) = e1 (n) -y2
(N) = N (n) + y1 ^T (n) C (n) -y1
^T (n) c (n) where C (n) ≈c (n), x (n) = e2 (n) = g0 (n) N (n) = N '
(N), so e3 (n) = y3 (n) + e2 (n) = y1 ^T (n) b
(N) + N '(n). Therefore,

[0083]

[Equation 5]

Where w (n); coefficient α of adaptive filter 3; step parameter r (n); output signal e3 (n) of FIR filter 5; output signal N '(n) of adder 9; periodic noise Comparing (Equation 5) and (Equation 4), the noise signal is N '(n), x
Only (n) is different. Therefore, in the case of (FIG. 6), the control can be performed similarly to the case of (FIG. 1). However, as can be seen from the above description, in the configuration of (FIG. 6), the noise of the noise source is not detected in advance by the noise detection microphone 1 as in the case of (FIG. 1), but the error detection microphone 2 is used.
In order to detect the noise that has reached the control point, the noise control is delayed. Therefore, this algorithm is effective only for noise with strong repeating characteristics such as periodic noise.

From the above, in the case of low frequency periodic noise,
As in the case of (FIG. 1), deterioration of the muffling effect at the evaluation point is suppressed, and the output of the subtractor 10 is shared by the input signal to the adder 9 and the input signals of the adaptive filter 3 and the FIR filter 5. As a result, it is possible to reduce the noise detection microphone and hardware associated therewith (for example, an anti-aliasing filter and an A / D converter). Even if there are a plurality of periodic noise sources, the same configuration can be used for control.

Next, an embodiment of the sixth invention will be described with reference to the drawings. FIG. 7 is a block diagram of a silencer according to an embodiment of the sixth aspect of the invention, in which periodic noise is controlled by a plurality of control points.
In FIG. 7, 2a to 2b are microphones which are error detectors, 3a to 3b are adaptive filters, 4a to 4b are speakers, 5a to 5d are FIR filters which are first digital filters, and 6a to 6
d is a FIR filter that is a second digital filter, 7a to 7d are FIR filters that are third digital filters, 8a to 8d are LMS calculators that are coefficient calculators, and 9a to 9d are additions. Bowl, 10a
-10d are subtractors, and 13a-13b are coefficient adders.

The operation of the muffling device having the above structure will be described below. First, since the microphones 2a and 2b serve both as a noise detector and an error detector, noise is detected by the microphones 2a and 2b.
The noise signal detected by the microphone 2a is subtracted by the subtractor 10a.
Thus, the output of the FIR filter 7a is subtracted. Here, in the FIR filter 7a, the transfer function C11 (jω) from the speaker 4a to the microphone 2a has a coefficient c11 in advance.
Since it is identified as (n), the output component of the adaptive filter 3a is removed from the output of the subtractor 10a as in the case of (FIG. 3).

Next, the output is F in the subtractor 10c.
It is subtracted from the output signal of the IR filter 7c. Where F
In the IR filter 7c, the transfer function C21 (jω) from the speaker 4b to the microphone 2a is identified in advance as a coefficient c21 (n). Therefore, the output component of the adaptive filter 3b is removed from the output of the subtractor 10c, resulting in only the noise signal detected by the microphone 2a. Similarly, the output components of the adaptive filters 3a to 3b are removed from the output of the subtractor 10b, leaving only the noise signal detected by the microphone 2b. Then, each of these noise signals is transmitted to the adaptive filter 3a.
3b and FIR filters 5a to 5d. The noise signal processed by the adaptive filter 3a is F
The noise signals that are input to the IR filters 6a to 6b, the FIR filters 7a to 7b, and the speaker 4a and processed by the adaptive filter 3b are input to the FIR filters 6c to 6d, the FIR filters 7c to 7d, and the speaker 4b. Then, the microphones 2a to
In 2b, the reproduced sound from the speakers 4a to 4b and the noise from the noise source interfere with each other.

On the other hand, the output of the subtractor 10c is added to the output of the FIR filter 6a in the adder 9a and input to the LMS calculator 8a. Here, in the FIR filter 6a, the transfer function B11 (jω) from the speaker 4a to the actual evaluation point (for example, the right ear of the listener) is beforehand provided.
Is identified as the coefficient b11 (n). Therefore, the adder 9
The output of a is approximately equal to the detection signal of the microphone installed near the right ear which is the actual evaluation point.

Similarly, the outputs of the adders 9b and 9d are approximately equal to the detection signal of the microphone installed in the left ear, and the output of the adder 9c is approximately equal to the detection signal of the microphone installed in the right ear. LMS
The computing unit 8a calculates the coefficient of the adaptive filter 3a based on the outputs of the FIR filter 5a and the adder 9a so that the output of the adder 9a is minimized. Therefore, in the LMS calculator 8a, the adaptive filter 3a is connected to the speaker 4a.
Is used to determine the coefficient for controlling the right ear.

Similarly, the LMS calculator 8b calculates the coefficient of the adaptive filter 3a by the outputs of the FIR filter 5b and the adder 9b so that the output of the adder 9b becomes the minimum, and the coefficient is calculated by the adaptive filter 3a. 4a is used to control the left ear region. Then, in the coefficient adder 13a, the coefficient calculated by the LMS calculator 8a and the coefficient calculated by the LMS calculator 8b are added,
The coefficient of the adaptive filter 3a is updated. The same applies to the adaptive filter 3b. Therefore, the periodic noise is sufficiently attenuated at both ears of the listener.

As can be understood from the above description, in the configuration of (FIG. 7), the noise of the noise source is not detected in advance by the noise detecting microphone 1 as in the case of (FIG. 3), but the error detecting microphone 2a is used. Due to the detection of the noise that has reached the control point by 2b, the noise control is delayed. Therefore, this algorithm is effective only for noise with strong repeating characteristics such as periodic noise. Also (Fig. 7) and (Fig. 6)
Comparing the above, the configuration of (FIG. 7) is based on the configuration of (FIG. 6), and the algorithm is naturally expanded in the same manner as in the case of (FIG. 3) so as to enable control at a plurality of points. is there.

As described above, in the case of low frequency periodic noise, even if there are a plurality of control points (FIG. 6), deterioration of the silencing effect at the evaluation points is suppressed, and the subtractor 10b, 1
The output of 0c is input to the adders 9a to 9d, the adaptive filters 3a to 3b, and the FIR filters 5a to
By sharing it with the input signal of 5d, it is possible to reduce the noise detection microphone and the hardware associated therewith (for example, an anti-alias filter and an A / D converter).

In this embodiment, the number of error detecting microphones and the number of speakers are two. However, the number may be increased according to the application, and in this case, the algorithm may be naturally expanded for the increased amount. In addition, the subtractor 10a and the subtractor 10
c into one, and the FI is output from the microphone 2a.
The outputs of the R filters 7a and 7c are subtracted, and the subtracter 1
0b and subtractor 10d are combined into one microphone 2
The configuration may be such that the outputs of the FIR filters 7b and 7d are subtracted from the output of b. Further, even if there are a plurality of periodic noise sources, the control can be performed with the same configuration.

Next, an embodiment of the seventh invention will be described with reference to the drawings. FIG. 8 is a block diagram of a silencer according to an embodiment of the seventh aspect of the present invention, in which periodic noise is controlled by a plurality of control points.
In FIG. 8, 2a to 2b are microphones that are error detectors, 3a to 3d are adaptive filters, 4a to 4b are speakers, 5a to 5d are FIR filters that are first digital filters, and 6a to 6
d is a FIR filter that is a second digital filter, 7a to 7d are FIR filters that are third digital filters, 8a to 8d are LMS calculators that are coefficient calculators, and 9a to 9d are additions. Bowl, 10a
-10d are subtractors, and 14a-14b are control signal adders.

The operation of the muffler having the above structure will be described below. Similar to (FIG. 7), since the microphones 2a and 2b serve both as a noise detector and an error detector, noise is detected by the microphones 2a and 2b. The output of the FIR filter 7a is subtracted from the noise signal detected by the microphone 2a by the subtractor 10a. Here, the FIR filter 7a includes a transfer function C11 from the speaker 4a to the microphone 2a in advance.
Since (jω) is identified as the coefficient c11 (n), the output component of the adaptive filter 3a is removed from the output of the subtractor 10a as in the case of (FIG. 7).

Then, the output is F in the subtractor 10c.
It is subtracted from the output signal of the IR filter 7c. Where F
In the IR filter 7c, the transfer function C21 (jω) from the speaker 4b to the microphone 2a is identified in advance as a coefficient c21 (n). Therefore, the output component of the adaptive filter 3b is removed from the output of the subtractor 10c, resulting in only the noise signal detected by the microphone 2a.

Similarly, the output components of the adaptive filters 3a to 3b are removed from the output of the subtractor 10b, leaving only the noise signal detected by the microphone 2b. The output of the subtractor 10c is the adaptive filter 3a,
3c and the FIR filters 5a and 5c, and the subtracter 1
0b output is adaptive filter 3b, 3d and FIR
It is input to the filters 5b and 5d. The noise signal from the subtractor 10c processed by the adaptive filter 3a is added to the noise signal from the subtractor 10b processed by the adaptive filter 3b by the control signal adder 14a, and the output is added to the FIR filters 6a to 6b. And the FIR filters 7a and 7b and the speaker 4a.

Similarly, the noise signal from the subtractor 10c processed by the adaptive filter 3c is added to the noise signal from the subtractor 10b processed by the adaptive filter 3d by the control signal adder 14b, and the output is output. Are FIR filters 6c-6d and FIR filters 7c-7d
Is input to the speaker 4b. And in the microphones 2a-2b installed at the control positions, the speakers 4a-4b
The reproduced sound from the sound source and the noise from the noise source interfere with each other.

On the other hand, the output of the subtractor 10c is added to the output of the FIR filter 6a in the adder 9a,
The LMS calculator 8a is input to the LMS calculator 8a and is added to the output of the FIR filter 6c in the adder 9c.
Entered in. Here, in the FIR filter 6a, the transfer function B11 (jω) from the speaker 4a to the actual evaluation point (for example, the right ear of the listener) is identified in advance as a coefficient b11 (n). Therefore, the output of the adder 9a becomes approximately equal to the detection signal of the microphone installed near the right ear which is the actual evaluation point.

Similarly, the outputs of the adders 9b and 9d are approximately equal to the detection signal of the microphone installed in the left ear, and the output of the adder 9c is approximately equal to the detection signal of the microphone installed in the right ear. LMS
The computing unit 8a calculates the coefficient of the adaptive filter 3a based on the outputs of the FIR filter 5a and the adder 9a so that the output of the adder 9a is minimized. Therefore, the LMS calculator 8a obtains a coefficient when the adaptive filter 3a controls the right ear by using the speaker 4a.

Similarly, the LMS calculator 8b calculates the coefficient of the adaptive filter 3b from the outputs of the FIR filter 5b and the adder 9b so that the output of the adder 9b becomes the minimum, and the coefficient is calculated by the adaptive filter 3b. 4a is used to control the left ear region. The same applies to the adaptive filters 3c to 3d. Therefore, the periodic noise is sufficiently attenuated at both ears of the listener.

As can be seen from the above description, the configuration of (FIG. 8) has the LMS calculators 8a to 8d in the configuration of (FIG. 7).
Each of them has an adaptive filter. Therefore, although the calculation amount is increased as compared with (FIG. 7), the noise signals processed by the adaptive filters 3a to 3d and the coefficients obtained by the LMS calculators 8a to 8d are supplied to the FIR filters 5a to 5d. Since it is obtained using the detection signals of the corresponding microphones 2a to 2b,
Noise control can be executed accurately and stably. Further, similarly to the case of (FIG. 7), the deterioration of the silencing effect at the evaluation point is suppressed, and the outputs of the subtracters 10b and 10c are input to the adders 9a to 9d and the adaptive filter 3a.
.About.3d and the input signals of the FIR filters 5a to 5d, the noise detection microphone and hardware associated therewith (for example, anti-aliasing filter and A / D converter) can be reduced.

In this embodiment, the number of error detecting microphones and the number of speakers are two. However, the number may be increased according to the application, and in this case, the algorithm may be naturally expanded for the increased amount. In addition, the subtractor 10a and the subtractor 10
c into one, and the FI is output from the microphone 2a.
The outputs of the R filters 7a and 7c are subtracted, and the subtracter 1
0b and subtractor 10d are combined into one microphone 2
The configuration may be such that the outputs of the FIR filters 7b and 7d are subtracted from the output of b. Further, even if there are a plurality of periodic noise sources, the control can be performed with the same configuration.

FIG. 9 is a block diagram of a silencer according to the eighth embodiment of the present invention, in which periodic noise is controlled by a plurality of control points. In FIG. 9, 2a to 2b are microphones which are error detectors, 3a to 3d are adaptive filters, 4a to 4b are speakers, 5a to 5d are FIR filters which are first digital filters, and 6a to 6d is a FIR filter which is a second digital filter, and 7a to 7
d is a FIR filter that is a third digital filter, and 8a to 8d are LMSs that are coefficient calculators.
Calculators, 9a to 9d are adders, 10a to 10d are subtractors, and 14a to 14b are control signal adders.

The operation of the silencer configured as above will be described below. Comparing (FIG. 9) and (FIG. 8), in FIG.
6d is configured to perform signal processing on the outputs of the control signal adders 14a to 14b, but in FIG. 9 the FIR filters 6a to 6d are configured to perform signal processing on the outputs of the adaptive filters 3a to 3d. Is.

Therefore, the same effect as in the case of (FIG. 8) can be obtained, and further, the LMS operation units 8a to 8d perform the L operation.
Since the error signal of the MS calculation includes only the output of the adaptive filters 3a to 3d for which the respective LMS calculators 8a to 8d calculate coefficients and the noise signal detected by the microphones 2a to 2b (for example, in the case of the LMS calculator 8a, , The output signals of the adaptive filters 3b to 3d other than the output signal from the adaptive filter 3a and the noise signal from the microphone 2a are not included), and the LMS calculation is performed efficiently and accurately.

In this embodiment, the number of error detecting microphones and the number of speakers are two. However, the number may be increased according to the application, and in that case, the algorithm may be naturally expanded for the increased amount. In addition, the subtractor 10a and the subtractor 10
c into one, and the FI is output from the microphone 2a.
The outputs of the R filters 7a and 7c are subtracted, and the subtracter 1
0b and subtractor 10d are combined into one microphone 2
The configuration may be such that the outputs of the FIR filters 7b and 7d are subtracted from the output of b. Further, even if there are a plurality of periodic noise sources, the control can be performed with the same configuration.

An embodiment of the ninth invention will be described below with reference to the drawings. FIG. 23 is a block diagram of a silencer according to an embodiment of the ninth invention. In FIG. 23, 1 is a microphone that is a noise detector, 2 is a microphone that is an error detector, 3 is an adaptive filter, 4 is a speaker, and 5 is a speaker.
Is a FIR filter which is a first digital filter, and 6 is an F filter which is a second digital filter.
IR filter, 7 is a FIR filter that is a third digital filter, 8 is an LMS calculator that is a coefficient calculator, 9 is an adder, 10 is a subtractor, and 17 is a fourth
FIR filter which is a digital filter of.

The operation of the muffler having the above structure will be described below. First, noise is detected by the microphone 1, and the detection signal is input to the adaptive filter 3 and the FIR filter 5. Then, the noise signal processed by the adaptive filter 3 is FIR
It is input to the filter 6, the FIR filter 7, and the speaker 4. Then, in the microphone 2 installed at a position away from the actual evaluation point, the reproduced sound from the speaker 4 and the noise from the noise source are detected.

The noise characteristic of the detection signal of the microphone 2 is corrected by the FIR filter 17, and the subtracter 1
At 0, it is subtracted from the output signal of the FIR filter 7.
Here, in the FIR filter 17, the ratio of the transfer function G0 (jω) from the noise source to the microphone 2 and the transfer function G1 (jω) from the noise source to the actual evaluation point is previously calculated as a coefficient g1 / g0.
In the FIR filter 7, the transfer function from the speaker 4 to the FIR filter 17 is previously identified as the coefficient cg1 / g0 (n) in the FIR filter 7.

Here, the coefficient of the FIR filter 17 is obtained by the configuration shown in (FIG. 24). In (Fig. 24),
2 is a microphone installed at the error detection position, 17 is F
An IR filter, 18 is a microphone placed at an actual evaluation point, 19 is an LMS calculator, and 20 is a subtractor. In this way, by using the microphone 2 for noise detection and the microphone 18 for error detection, the transfer function G0 (jω) from the noise source to the microphone 2 and the transfer function G1 (jω) from the noise source to the actual evaluation point are The coefficient g1 / g0 (n), which is the ratio, can be obtained. Similarly, for the coefficient of the FIR filter 7, after obtaining the coefficient of the FIR filter 17, (FIG. 25)
It is calculated with the configuration shown in.

From the above, the output component of the adaptive filter 3 is removed from the output of the subtractor 10, and as a result, only the noise signal detected by the microphone 2 is obtained. Further, this noise signal is output by the FIR filter 17. This means that the transfer function G0 (jω) is removed and only the transfer function G1 (jω) is affected. Then, the output of the subtractor 10 is added to the output of the FIR filter 6 in the adder 9 and input to the LMS calculator 8. Here, the transfer function B (jω) from the speaker 4 to the actual evaluation point (for example, the ear of the listener) is identified in the FIR filter 6 in advance as a coefficient b (n). Therefore, the output of the adder 9 becomes approximately equal to the detection signal of the microphone installed at the actual evaluation point.

The LMS calculator 8 updates the coefficient of the adaptive filter 3 by the outputs of the FIR filter 5 and the adder 9 so that the output of the adder 9 becomes the minimum. If this is expressed by a mathematical formula, it can be expressed as follows.

If C (n) ≈c (n), then e4 (n) = y1 ^T (n) b (n) + x ^T (n) g1
(N), so

[0116]

[Equation 6]

However, w (n); coefficient α of the adaptive filter 3; step parameter r (n); output signal e4 (n) of the FIR filter 5; output signal e4 (n) of the adder 9 (Equation 6) When approaching ≈0, the frequency characteristic W (k) of the coefficient w (n) of the adaptive filter 3 is approximated to {-G1 (k) / B (k)}. Therefore, (Fig. 1
Compared with the case of 8), the coefficient W (k) of the adaptive filter 3 becomes equal to {-G1 (k) / B (k)} of (FIG. 18). Therefore, the same effect as the effect (FIG. 18) can be obtained not only in the low frequency region but also in the entire control band.

Next, regarding one embodiment of the tenth invention,
A description will be given with reference to the drawings. (FIG. 26) is a block diagram of a silencer according to an embodiment of the tenth aspect of the present invention, in which a plurality of control points are present. (Fig. 2
In 6), 1 is a microphone that is a noise detector, 2a-2b are microphones that are error detectors, 3a-3b are adaptive filters, and 4a-4b.
Is a speaker, 5a to 5d are FIR filters that are first digital filters, 6a to 6d are FIR filters that are second digital filters, and 7a to
7d is the FIR which is the third digital filter
Filters, 8a to 8d are LMs which are coefficient calculators
S calculators, 9a to 9d are adders, 10a to 10d are subtractors, 13a to 13b are coefficient adders, and 17a to 17b are FIR filters which are fourth digital filters.

The operation of the muffling device having the above structure will be described below. First, noise is detected by the microphone 1, and the detection signal is input to the adaptive filters 3a to 3b and the FIR filters 5a to 5d. The noise signal processed by the adaptive filter 3a is input to the FIR filters 6a to 6b, the FIR filters 7a to 7b and the speaker 4a, and the noise signal processed by the adaptive filter 3b is FIR filters 6c to 6d and the FIR filter. 7c to 7d and speaker 4
Input to b. Then, the microphones 2a to 2b installed at the control positions detect the reproduced sound from the speakers 4a to 4b and the noise from the noise source. The noise characteristic of the detection signal of the microphone 2a is corrected by the FIR filter 17a, and is further subtracted from the output signal of the FIR filter 7a in the subtractor 10a.

Here, the transfer function from the noise source is (
In the case as shown in 2), the transfer function G0 (j) from the noise source to the microphone 2a is preset in the FIR filter 17a.
ω) and the noise source and the actual evaluation point (for example, the right ear of the listener)
The ratio of the transfer function G2 (jω) up to is identified as a coefficient g2 / g0 (n), and the transfer function from the speaker 4a to the FIR filter 17a is previously included in the FIR filter 7a.
It has been identified as c11g2 / g0 (n). Here, the coefficients of the FIR filters 17a and 17b may be obtained separately as in the case of (FIG. 24).

From the above, the output component of the adaptive filter 3a is removed from the output of the subtractor 10a. Next, the output is subtracted from the output signal of the FIR filter 7c in the subtractor 10c. Here, the FIR filter 7c includes the speaker 4b and the FIR filter 1 in advance.
The transfer function up to 7a is identified as the coefficient c21g2 / g0 (n). Therefore, the output component of the adaptive filter 3b is removed from the output of the subtractor 10c, resulting in only the noise signal detected by the microphone 2a. Further, this noise signal has the transfer function G0 (jω) removed by the FIR filter 17a, so that only the transfer function G2 (jω) is affected.

Similarly, the output components of the adaptive filters 3a to 3b are removed from the output of the subtractor 10b to become only the noise signal detected by the microphone 2b. The noise signal is transferred by the FIR filter 17b to the transfer function G1 ( j
Since ω) is removed, only the transfer function G3 (jω) is affected. The output of the subtractor 10c is the adder 9
a, the output of the FIR filter 6a is added, and LM
It is input to the S calculator 8a. Here, the FIR filter 6
In a, the transfer function B11 (jω) from the speaker 4a to the actual evaluation point (for example, the right ear of the listener) is previously calculated as a coefficient b11.
Identified as (n). Therefore, the output of the adder 9a becomes approximately equal to the detection signal of the microphone installed at the right ear which is the actual evaluation point.

Similarly, the outputs of the adders 9b and 9d are approximately equal to the detection signal of the microphone installed in the left ear, and the output of the adder 9c is approximately equal to the detection signal of the microphone installed in the right ear. LMS
The computing unit 8a calculates the coefficient of the adaptive filter 3a based on the outputs of the FIR filter 5a and the adder 9a so that the output of the adder 9a is minimized. Therefore, the LMS calculator 8a obtains a coefficient when the adaptive filter 3a controls the right ear by using the speaker 4a.

Similarly, the LMS calculator 8b calculates the coefficient of the adaptive filter 3a from the outputs of the FIR filter 5b and the adder 9b so that the output of the adder 9b becomes the minimum, and the coefficient is calculated by the adaptive filter 3a. 4a is used to control the left ear region. Then, in the coefficient adder 13a, the coefficient calculated by the LMS calculator 8a and the coefficient calculated by the LMS calculator 8b are added,
The coefficient of the adaptive filter 3a is updated. The same applies to the adaptive filter 3b.

From the above, as in the case of (FIG. 23), even if the adaptive filters 3a to 3b are installed at positions apart from the evaluation points by the speakers 4a to 4b and the microphones 2a to 2b, it is as if the evaluation points Since it operates as if it were controlled, it is possible to obtain the same effect as when the microphones 2a to 2b are installed at the evaluation points. Further, for example, focusing on the LMS calculator 8a, only the noise signal detected by the microphone 2a and the output signal of the adaptive filter 3a processed by the FIR filter 6a are output to the adder 9a which is an error input of the LMS calculation. Since it is included, the LMS operation is executed efficiently and accurately.

In this embodiment, the number of error detecting microphones and the number of speakers are two. However, the number may be increased according to the application, and in that case, the algorithm may be naturally expanded for the increased amount. In addition, the subtractor 10a and the subtractor 10
c from the output of the FIR filter 17a
The outputs of the IR filters 7a and 7c may be subtracted, the subtractors 10b and 10d may be combined into one, and the outputs of the FIR filters 7b and 7d may be subtracted from the output of the FIR filter 17b.

Next, an embodiment of the eleventh invention will be described.
A description will be given with reference to the drawings. (FIG. 27) is a block diagram of a silencer according to an embodiment of the eleventh aspect of the present invention, in which periodic noise is controlled. (Fig. 2
In 7), 2 is a microphone that is an error detector, 3 is an adaptive filter, 4 is a speaker, 5 is a FIR filter that is a first digital filter, and 6 is a FI that is a second digital filter.
R filter, 7 is a FIR filter which is a third digital filter, and 8 is an L filter which is a coefficient calculator.
MS calculator, 9 is an adder, 10 is a subtractor, and 17 is a FIR filter which is a fourth digital filter.

The operation of the muffling device having the above structure will be described below. First, since the microphone 2 serves both as a noise detector and an error detector, the microphone 2 detects noise. The detected noise signal is subjected to signal processing by the FIR filter 17, and F
The output of the IR filter 17 is subtracted from the output of the FIR filter 7 by the subtractor 10. Here, in the FIR filter 17, the ratio of the transfer function G0 (jω) from the noise source to the microphone 2 and the transfer function G1 (jω) from the noise source to the actual evaluation point is set as a coefficient g1 / g0 (n) in advance. The transfer function from the speaker 4 to the FIR filter 17 is previously identified in the FIR filter 7 by the coefficient cg1 / g0 (n).
Has been identified as.

Here, the coefficient of the FIR filter 17 is obtained by the configuration shown in (FIG. 28). In (Fig. 28),
2 is a microphone installed at the error detection position, 17 is F
An IR filter, 18 is a microphone placed at an actual evaluation point, 19 is an LMS calculator, and 20 is a subtractor. In this way, by using the microphone 2 for noise detection and the microphone 18 for error detection, the transfer function G0 (jω) from the noise source to the microphone 2 and the transfer function G1 (jω) from the noise source to the actual evaluation point. Coefficient that is the ratio of g1 / g0 (n)
Can be asked. It should be noted here that only the frequency component of the periodic noise N (t) has a coefficient.
It is obtained as g1 / g0 (n), which is different from the case of (FIG. 23).

From the above, as in the case of (FIG. 23), the output component of the adaptive filter 3 is removed from the output of the subtractor 10, resulting in only the noise signal detected by the microphone 2. Furthermore, this noise signal is F
The transfer function G0 (jω) is removed by the IR filter 17, and only the transfer function G1 (jω) is affected.
Then, this noise signal is transmitted to the adaptive filter 3 and the FI.
It is input to the R filter 5. The noise signal processed by the adaptive filter 3 is input to the FIR filter 6, the FIR filter 7, and the speaker 4. Then, in the microphone 2 installed at a position away from the actual evaluation point, the reproduced sound from the speaker 4 and the noise from the noise source interfere with each other.

On the other hand, the output of the subtracter 10 is added to the output of the FIR filter 6 in the adder 9 and input to the LMS calculator 8. Here, in the FIR filter 6,
The transfer function B (jω) from the speaker 4 to the actual evaluation point (for example, the ear of the listener) is previously identified as the coefficient b (n). Therefore, the output of the adder 9 becomes approximately equal to the detection signal of the microphone installed at the actual evaluation point. The LMS calculator 8 updates the coefficient of the adaptive filter 3 based on the outputs of the FIR filter 5 and the adder 9 so that the output of the adder 9 is minimized. Therefore, the periodic noise is sufficiently attenuated at the actual evaluation point. If this is expressed by a mathematical formula, it can be expressed as follows.

Here, if C (n) ≈c (n), then x (n) = e3 (n) = g1 (n) N (n) = N '
And since (n), the e4 (n) = y3 (n ) + e3 (n) = y1 T (n) b (n) + N '(n). Therefore,

[0133]

[Equation 7]

Where w (n); coefficient α of adaptive filter 3; step parameter r (n); output signal e4 (n) of FIR filter 5; output signal N '(n) of adder 9; periodic noise Comparing (Equation 7) and (Equation 6), the noise signal is N '(n), x
Only (n) is different. Therefore, in the case of (FIG. 27), control can be performed in the same manner as in the case of (FIG. 23). However, as can be understood from the above description, in the configuration of (FIG. 27) (FIG.
As in the case of 3), the noise of the noise source is not detected in advance by the noise detecting microphone 1, but the noise reaching the control point is detected by the error detecting microphone 2, so that the noise control is delayed. Therefore, this algorithm is effective only for noise with strong repeating characteristics such as periodic noise.

From the above, the same effect can be obtained when the microphone 2 is installed at the evaluation points in the entire control band, as in the case of (FIG. 23). Further, by sharing the output of the subtracter 10 with the input signal to the adder 9 and the input signals of the adaptive filter 3 and the FIR filter 5, noise detection microphone and hardware accompanying it (for example, anti-alias filter and A / D converter). Etc.) can be reduced.

Next, regarding an embodiment of the twelfth invention,
A description will be given with reference to the drawings. (FIG. 29) is a block diagram of a silencer according to an embodiment of the twelfth invention, in which periodic noise is controlled by a plurality of control points. In FIG. 29, 2a to 2b are microphones which are error detectors, 3a to 3b are adaptive filters, 4a to 4b are speakers, 5a to 5d are FIR filters which are first digital filters, and 6a.
˜6d is FI which is the second digital filter
R filter, 7a to 7d are FIR filters which are third digital filters, 8a to 8d are LMS calculators which are coefficient calculators, 9a to 9d are adders, 1
0a to 10d are subtractors, 13a to 13b are coefficient adders,
17a to 17b are FIR filters which are the fourth digital filters.

The operation of the muffling device having the above structure will be described below. First, since the microphones 2a and 2b serve both as a noise detector and an error detector, noise is detected by the microphones 2a and 2b.
The noise signal detected by the microphone 2a is subjected to signal processing by the FIR filter 17a, and the FIR filter 17a
Is subtracted from the output of the FIR filter 7a by the subtractor 10a. Here, when the transfer function from the noise source is as shown in (FIG. 30), the transfer function G0 (j from the noise source to the microphone 2a is previously stored in the FIR filter 17a.
ω) and the noise source and the actual evaluation point (eg, the ear of the listener)
The transfer function from the speaker 4a to the FIR filter 17a has a coefficient c11g2 / g0 (n) in advance in the FIR filter 7a. Has been identified as. FIR here
The coefficients of the filters 17a and 17b may be obtained separately as in the case of (FIG. 28).

It should be noted here that only the frequency component of the periodic noise N (t) has the coefficient g2 / g0 (n).
Alternatively, it can be obtained as g3 / g1 (n), which is different from the case of (FIG. 26).

From the above, the output component of the adaptive filter 3a is removed from the output of the subtractor 10a.

Then, the output is F in the subtractor 10c.
It is subtracted from the output signal of the IR filter 7c. Where F
In the IR filter 7c, the transfer function from the speaker 4b to the FIR filter 17a is identified in advance as a coefficient c21g2 / g0 (n). Therefore, the output component of the adaptive filter 3b is removed from the output of the subtractor 10c, and further the transfer function G0 is obtained by the FIR filter 17a.
(jω) is removed, and only the noise signal is affected by only the transfer function G2 (jω). Similarly, the output components of the adaptive filters 3a to 3b are removed from the output of the subtractor 10b, and the transfer function G1 (j
ω) is removed, and only the noise signal is affected by only the transfer function G3 (jω). Then, the noise signals are transmitted to the adaptive filters 3a and 3b and the FIR filter 5, respectively.
a to 5d. The noise signal processed by the adaptive filter 3a is input to the FIR filters 6a to 6b, the FIR filters 7a to 7b, and the speaker 4a,
The noise signals signal-processed by the adaptive filter 3b are FIR filters 6c to 6d and FIR filters 7c to 7d.
Is input to the speaker 4b. And in the microphones 2a-2b installed in the control position, the speakers 4a-4
The reproduced sound from b and the noise from the noise source interfere with each other.

On the other hand, the output of the subtractor 10c is added to the output of the FIR filter 6a in the adder 9a and input to the LMS calculator 8a. Here, in the FIR filter 6a, the transfer function B11 (jω) from the speaker 4a to the actual evaluation point (for example, the right ear of the listener) is beforehand provided.
Is identified as the coefficient b11 (n). Therefore, the adder 9
The output of a is approximately equal to the detection signal of the microphone installed near the right ear which is the actual evaluation point.

Similarly, the outputs of the adders 9b and 9d are approximately equal to the detection signal of the microphone installed in the left ear, and the output of the adder 9c is approximately equal to the detection signal of the microphone installed in the right ear. LMS
The computing unit 8a calculates the coefficient of the adaptive filter 3a based on the outputs of the FIR filter 5a and the adder 9a so that the output of the adder 9a is minimized. Therefore, in the LMS calculator 8a, the adaptive filter 3a is connected to the speaker 4a.
Is used to determine the coefficient for controlling the right ear.

Similarly, the LMS calculator 8b calculates the coefficient of the adaptive filter 3a from the outputs of the FIR filter 5b and the adder 9b so that the output of the adder 9b becomes the minimum, and the coefficient is calculated by the adaptive filter 3a. 4a is used to control the left ear region. Then, in the coefficient adder 13a, the coefficient calculated by the LMS calculator 8a and the coefficient calculated by the LMS calculator 8b are added,
The coefficient of the adaptive filter 3a is updated. The same applies to the adaptive filter 3b. Therefore, the periodic noise is sufficiently attenuated at both ears of the listener.

As can be seen from the above description (FIG. 29)
In the configuration (1), the noise detection microphone 1 does not detect the noise of the noise source in advance as in the case of (FIG. 26), but the error detection microphones 2a to 2b detect the noise reaching the control point. There is a delay in noise control.
Therefore, this algorithm is effective only for noise with strong repeating characteristics such as periodic noise. Further, comparing (FIG. 29) and (FIG. 27), the configuration of (FIG. 29) is (FIG.
Based on the configuration of 7), the algorithm is naturally expanded as in the case of (FIG. 26) so as to enable control at a plurality of points.

From the above, the same effect as in the case where the microphones 2a to 2b are installed at the evaluation points is obtained as in the case of (FIG. 27), and the outputs of the subtracters 10b and 10c are supplied to the adders 9a to 9d. Input signal and adaptive filter 3a
.. 3b and the input signals of the FIR filters 5a to 5d, the noise detection microphone and hardware associated therewith (for example, antialiasing filter and A / D converter) can be reduced.

In this embodiment, the number of error detecting microphones and the number of speakers are two. However, the number may be increased according to the application, and in that case, the algorithm may be naturally expanded to the increased amount. In addition, the subtractor 10a and the subtractor 10
c are combined into one, and the outputs of the FIR filters 7a and 7c are subtracted from the output of the FIR filter 17a. Similarly, the subtractor 10b and the subtractor 10d are combined into one, and the output of the FIR filter 17b is converted into the FIR filter 7b. , 7d may be subtracted.

Next, regarding one embodiment of the thirteenth invention,
A description will be given with reference to the drawings. FIG. 31 is a block diagram of a silencer according to an embodiment of the thirteenth invention, in which periodic noise is controlled by a plurality of control points. In FIG. 31, 2a to 2b are microphones that are error detectors, 3a to 3d are adaptive filters, 4a to 4b are speakers, 5a to 5d are FIR filters that are first digital filters, and 6a.
˜6d is FI which is the second digital filter
R filter, 7a to 7d are FIR filters which are third digital filters, 8a to 8d are LMS calculators which are coefficient calculators, 9a to 9d are adders, 1
0a to 10d are subtractors, 14a to 14b are control signal adders, and 17a to 17b are FIR filters which are fourth digital filters.

The operation of the silencer configured as above will be described below. Similarly to (FIG. 29), since the microphones 2a and 2b serve both as a noise detector and an error detector, noise is detected by the microphones 2a and 2b. The noise signal detected by the microphone 2a is subjected to signal processing by the FIR filter 17a, and FI
The output of the R filter 17a is FIR by the subtractor 10a.
The output of the filter 7a is subtracted. Here, in the case where the transfer function from the noise source is as shown in (FIG. 30), the transfer function G0 (jω) from the noise source to the microphone 2a and the actual evaluation point (for example, reception The ratio to the transfer function G2 (jω) up to the listener's ears) is identified as a coefficient g2 / g0 (n), and the FIR filter 7a includes the speaker 4a to the FIR filter 17a in advance.
The transfer functions up to are identified as the coefficients c11g2 / g0 (n). Here, the coefficients of the FIR filters 17a and 17b may be separately calculated in the same manner as in the case of (FIG. 28).

It should be noted here that only the frequency component of the periodic noise N (t) has the coefficient g2 / g0 (n).
Alternatively, it can be obtained as g3 / g1 (n), which is different from the case of (FIG. 26).

From the above, the output component of the adaptive filter 3a is removed from the output of the subtractor 10a.

Then, the output is F in the subtractor 10c.
It is subtracted from the output signal of the IR filter 7c. Where F
In the IR filter 7c, the transfer function from the speaker 4b to the FIR filter 17a is identified in advance as a coefficient c21g2 / g0 (n). Therefore, the output component of the adaptive filter 3b is removed from the output of the subtractor 10c, and further the transfer function G0 is obtained by the FIR filter 17a.
(jω) is removed, and only the noise signal is affected by only the transfer function G2 (jω). Similarly, the output components of the adaptive filters 3a to 3b are removed from the output of the subtractor 10b, and the transfer function G1 (j
ω) is removed, and only the noise signal is affected by only the transfer function G3 (jω). Then, the noise signals are transmitted to the adaptive filters 3a and 3b and the FIR filter 5, respectively.
a to 5d. The noise signal signal-processed by the adaptive filter 3a is added to the output of the adaptive filter 3b by the control signal adder 14a, and then F
It is input to the IR filters 6a-6b, the FIR filters 7a-7b, and the speaker 4a. Further, the noise signal processed by the adaptive filter 3b is the control signal adder 14b.
Is added to the output of the adaptive filter 3d by the FIR filter 6c-6d and the FIR filter 7c-
7d and the speaker 4b. Then, in the microphones 2a to 2b installed at the control position, the speaker 4a
The reproduced sound from 4b and the noise from the noise source interfere with each other.

On the other hand, the output of the subtractor 10c is added to the output of the FIR filter 6a in the adder 9a,
The LMS calculator 8a is input to the LMS calculator 8a and is added to the output of the FIR filter 6c in the adder 9c.
Entered in. Here, in the FIR filter 6a, the transfer function B11 (jω) from the speaker 4a to the actual evaluation point (for example, the right ear of the listener) is identified in advance as a coefficient b11 (n). Therefore, the output of the adder 9a becomes approximately equal to the detection signal of the microphone installed near the right ear which is the actual evaluation point.

Similarly, the outputs of the adders 9b and 9d are approximately equal to the detection signal of the microphone installed in the left ear, and the output of the adder 9c is approximately equal to the detection signal of the microphone installed in the right ear. LMS
The computing unit 8a calculates the coefficient of the adaptive filter 3a based on the outputs of the FIR filter 5a and the adder 9a so that the output of the adder 9a is minimized. Therefore, the LMS calculator 8a obtains a coefficient when the adaptive filter 3a controls the right ear by using the speaker 4a.

Similarly, the LMS calculator 8b calculates the coefficient of the adaptive filter 3b from the outputs of the FIR filter 5b and the adder 9b so that the output of the adder 9b becomes the minimum, and the coefficient is calculated by the adaptive filter 3b. 4a is used to control the left ear region. The same applies to the adaptive filters 3c to 3d. Therefore, the periodic noise is sufficiently attenuated at both ears of the listener.

As can be seen from the above description (FIG. 31)
The configuration of the LMS calculator 8a to the configuration of (FIG. 29)
An adaptive filter is provided for each 8d. Therefore, although the amount of calculation increases compared to (FIG. 29), the noise signal processed by each adaptive filter 3a to 3d and the coefficient obtained by the LMS calculators 8a to 8d are supplied to each FIR filter 5a to 5d. Since it is obtained by using the detection signals of the corresponding microphones 2a to 2b, noise control can be performed accurately and stably. And again, as in the case of (FIG. 29), the microphones 2a ...
The same effect as when 2b is installed at the evaluation point is obtained.
Further, the outputs of the subtractors 10b and 10c are added to the adders 9a to 9a.
By sharing the input signal to d and the input signals of the adaptive filters 3a to 3d and the FIR filters 5a to 5d, it is possible to reduce the noise detection microphone and hardware associated therewith (for example, antialiasing filter and A / D converter). .

In this embodiment, the number of error detecting microphones and the number of speakers are two. However, the number may be increased depending on the application, and in that case, the algorithm may be naturally expanded for the increased amount. In addition, the subtractor 10a and the subtractor 10
c are combined into one, and the outputs of the FIR filters 7a and 7c are subtracted from the output of the FIR filter 17a. Similarly, the subtractor 10b and the subtractor 10d are combined into one, and the output of the FIR filter 17b is converted into the FIR filter 7b. , 7d may be subtracted.

FIG. 32 is a block diagram of a silencer according to an embodiment of the fourteenth aspect of the present invention in which periodic noise is controlled by a plurality of control points. In FIG. 32, 2a and 2b are microphones that are error detectors, 3a and 3d are adaptive filters, and 4a and 4a.
b is a speaker, 5a to 5d are FIR filters which are first digital filters, 6a to 6d are FIR filters which are second digital filters, and 7a
˜7d is FI which is the third digital filter
R filter, 8a to 8d are L, which are coefficient calculators
MS calculators, 9a to 9d adders, 10a to 10d subtractors, 14a to 14b control signal adders, 17a to 17
b is an FIR filter which is the fourth digital filter.

The operation of the silencer configured as above will be described below. When (FIG. 32) is compared with (FIG. 31), the FIR filter 6 is shown in (FIG. 31).
Although a to 6d are configured to process the outputs of the control signal adders 14a to 14b, in FIG. 32, the FIR filters 6a to 6d are configured to process the outputs of the adaptive filters 3a to 3d. I'm just there.

Therefore, the same effect as in the case of (FIG. 31) can be obtained, and the error signal of the LMS calculation performed by the LMS calculators 8a to 8d includes the LMS calculators 8a to 8d.
Include only the outputs of the adaptive filters 3a to 3d for obtaining the coefficients and the noise signals detected by the microphones 2a to 2b (for example, in the case of the LMS calculator 8a, the output signal from the adaptive filter 3a and the microphone 2a).
Filters 3b to 3d other than noise signals from the
Output signal is not included), and the LMS operation is executed efficiently and accurately.

In this embodiment, the number of error detecting microphones and the number of speakers are two. However, the number may be increased according to the application, and in this case, the algorithm may be naturally expanded for the increased amount. In addition, the subtractor 10a and the subtractor 10
c are combined into one, and the outputs of the FIR filters 7a and 7c are subtracted from the output of the FIR filter 17a. Similarly, the subtractor 10b and the subtractor 10d are combined into one, and the output of the FIR filter 17b is converted into the FIR filter 7b. , 7d may be subtracted.

By the way, (FIG. 3), (FIG. 4), (FIG. 5),
(FIG. 7), (FIG. 8), (FIG. 9), (FIG. 26), (FIG. 2)
9), (FIG. 31), and (FIG. 32), the speaker 4
Although a case where two a-4b and two microphones 2a-2b are used is shown, as one application thereof (Fig. 10), the speakers 4a-4b and the microphones 2a-2b are shown.
It is conceivable that the chair is installed on the headrest portion of the chair 15. In this case, the transfer functions C11 (jω) and C12 (j) from the speakers 4a-4b to the microphones 2a-2b.
ω), C21 (jω), C22 (jω), and the speaker 4a to
4b to the listener's binaural transfer functions B11 (jω), B
12 (jω), B21 (jω), and B22 (jω) are (Fig. 11)
become that way. Thus, in an actual application, the speaker 4
The a to 4b and the microphones 2a to 2b are installed near the head of the listener, such as the headrest portion of the chair 15, but some distance from the ears. At this time, according to the conventional control method shown in FIG.
chair 1 covering a-4b and microphones 2a-2b
Due to the effect of the cushioning material and the front cover of No. 5, and the difference in distance between the microphones 2a and 2b and both ears, even if a sufficient noise reduction effect is obtained with the microphones 2a and 2b, the effect deteriorates at the listener's ears. . However, (FIG. 3), (FIG. 4), (FIG. 5), (FIG. 7), (FIG. 8), (FIG. 9), (FIG. 26), (FIG. 29), (FIG. 31), (FIG.
In the configuration of 2), the speakers 4a and 4b and the microphone 2 are used.
Even if a to 2b are installed in the chair 15, the sound deadening effect can be sufficiently obtained in the ear of the listener as described above.

Further, (FIG. 12), (FIG. 13), (FIG. 1)
As shown in 4), by providing a recess for positioning the listener's head on the headrest portion of the chair 15, both ears of the listener can always be guided to the optimum control position. Further, as shown in FIGS. 10 and 12, the chair 15 has a structure in which the backrest portion and the headrest portion are integrally formed. However, even in the case where the headrest portion is detachable as in a car seat, the headrest portion can be removed. Similar effects can be obtained by installing speakers 4a-4b and microphones 2a-2b in the section and operating them in a state where the headrest section and the backrest section are connected (normal use state). Furthermore, speakers 4a-4b and microphones 2a-2b are installed on the pillow 16 shown in (FIG. 15), (FIG. 3), (FIG. 4), (FIG. 5), (FIG. 7),
(FIG. 8), (FIG. 9), (FIG. 26), (FIG. 29), (FIG. 3)
By performing the controls 1) and (FIG. 32), it is possible to reduce the snoring of the person next to the person at the ear position.

When controlling only a considerably low frequency, even if the control of (FIG. 1) or (FIG. 6) is performed by using one speaker 4 and one microphone 2 as shown in (FIG. 16). good. In addition, as shown in (FIG. 17), the speakers 4a to 4c are provided on a wall or ceiling covering the listener.
Alternatively, the microphones 2a to 2c may be installed.

[0164]

As is apparent from the above description, according to the first aspect of the present invention, even if the error detector is installed at a position distant from the actual evaluation point, the control signal of the second and third digital filters can be controlled. By the correction, the adaptive filter updates the coefficient so as to control the actual evaluation point, so that the noise can be sufficiently attenuated at the evaluation point away from the error detector.

In the second invention, noise control is performed by using a plurality of speakers and an error detector, and, like the first invention, the error detector is installed at a position away from an actual evaluation point. However, since the adaptive filter is configured to update the coefficients so as to control the actual evaluation points by the correction of the control signals by the second and third digital filters, a plurality of error detectors separated from the error detectors can be obtained. Noise can be sufficiently attenuated at the evaluation point.

In the third invention, noise control is performed for a plurality of noise sources using a plurality of speakers and an error detector, and the error detector is separated from an actual evaluation point as in the first invention. Even if it is installed at the position, the adaptive filter is configured to update the coefficients so as to control the actual evaluation points by the correction of the control signals by the second and third digital filters, so that each error detection It is possible to sufficiently dampen a plurality of noises at a plurality of evaluation points apart from the container.

A fourth aspect of the invention is similar to the third aspect of the invention in that the second digital filter of the third aspect of the invention is configured to perform signal processing on the output from the control signal adder, while reducing the amount of calculation. In the fifth invention, the output of the subtractor is used as the input signals of the adaptive filter and the first digital filter, so that the hardware can be reduced, and the control signal of the third digital filter can be reduced. By the correction, one or a plurality of periodic noises can be sufficiently attenuated at the evaluation point distant from the error detector.

In the sixth invention, the output of the subtractor is used as the input signal of the adaptive filter and the first digital filter, and the noise control is performed by using the plurality of speakers and the error detector. It is possible to realize a silencer having an excellent effect that one or a plurality of periodic noises can be sufficiently attenuated at a plurality of evaluation points apart from the error detector.

According to a seventh aspect of the invention, the output of the subtracter for outputting the noise signal detected by the error detector controlled by the adaptive filter and the first digital filter is used as the input signal of the adaptive filter and the first digital filter. Moreover, since the noise control is performed by using the plurality of speakers and the error detector, the effect of the sixth invention can be obtained more stably.

In the eighth invention, since the error signal of each LMS calculator does not include the output of the adaptive filter other than the output of the adaptive filter required by the LMS calculator, each LMS calculator can efficiently calculate. .

The ninth aspect of the invention is to correct the control signal by the second and third digital filters and correct the error signal by the fourth digital filter even if the error detector is installed at a position distant from the actual evaluation point. By the correction, the adaptive filter updates the coefficient so as to control the actual evaluation point, so that it is possible to obtain the same effect as if the error detector is installed at the evaluation point.

In the tenth invention, noise control is performed by using a plurality of speakers and an error detector, and, like the ninth invention, the error detector is installed at a position apart from an actual evaluation point. Also, by the correction of the control signal by the second and third digital filters and the correction of the error signal by the fourth digital filter, the adaptive filter is configured to update the coefficients so as to control the actual evaluation points. Therefore, it is possible to obtain the same effect as if each error detector is installed at each evaluation point.

In the eleventh aspect of the invention, the hardware can be reduced by using the output of the subtractor as the input signal of the adaptive filter and the first digital filter, and the correction of the control signal by the third digital filter and the fourth aspect of the invention. By correcting the error signal by the digital filter, the periodic noise can be sufficiently attenuated as if the error detector was installed at the evaluation point.

In the twelfth aspect of the invention, the output of the subtractor is used as the input signals of the adaptive filter and the first digital filter, and noise control is performed using a plurality of speakers and an error detector. By the correction of the control signal by the digital filter of 3 and the correction of the error signal by the fourth digital filter, at the plurality of evaluation points apart from each error detector, it is as if each error detector is installed at each evaluation point. The effect of can be obtained.

In a thirteenth aspect of the invention, the output of the subtractor that outputs the noise signal detected by the error detector controlled by the adaptive filter and the first digital filter is used as the input signal of the adaptive filter and the first digital filter.
Moreover, the noise control is performed by using a plurality of speakers and an error detector, and the control signal is corrected by the third digital filter and the error signal is corrected by the fourth digital filter. The effect can be obtained more stably.

In the fourteenth invention, since the error signal of each LMS calculator does not include the output of the adaptive filter other than the output of the adaptive filter required by the LMS calculator, each LMS calculator can efficiently calculate. .

[Brief description of drawings]

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

FIG. 2 is a block diagram showing a method for identifying coefficients of the FIR filter 5 according to the embodiment of the first invention.

FIG. 3 is a block diagram of an embodiment of the second invention.

FIG. 4 is a block diagram of an embodiment of the third invention.

FIG. 5 is a block diagram of an embodiment of the fourth invention.

FIG. 6 is a block diagram of an embodiment of the fifth invention.

FIG. 7 is a block diagram of an embodiment of the sixth invention.

FIG. 8 is a block diagram of an embodiment of the seventh invention.

FIG. 9 is a block diagram of an embodiment of the eighth invention.

FIG. 10 is a diagram showing a chair 15 in which speakers 4a-4b and microphones 2a-2b are installed in the embodiments of the second to fourth and sixth to eighth inventions.

FIG. 11 is a speaker 4a in the chair 15 (FIG. 10).
4b to microphones 2a to 2b showing transmission paths

FIG. 12 is a view showing a chair 15 in which speakers 4a-4b and microphones 2a-2b are installed in the embodiments of the second to fourth and sixth to eighth inventions, and in which a recess is provided.

FIG. 13 is a speaker 4a in the chair 15 (FIG. 12).
4b to microphones 2a to 2b showing transmission paths

FIG. 14 is a speaker 4a in the chair 15 (FIG. 12).
4b to microphones 2a to 2b showing transmission paths

FIG. 15 is a view showing a pillow 16 on which speakers 4a-4b and microphones 2a-2b are installed in the second to fourth and sixth to eighth embodiments of the invention.

FIG. 16 is a view showing a chair 15 on which a speaker 4 and a microphone 2 according to the first and fifth embodiments of the invention are installed.

FIG. 17 shows speakers 4a-4c and microphones 2a-2.
The figure which shows the example which installed c in the wall and the ceiling

FIG. 18 is a block diagram showing a conventional silencer.

FIG. 19 is a block diagram showing a method for identifying coefficients of the FIR filter 5 in (FIG. 18).

FIG. 20 is a block diagram showing a conventional silencer.

FIG. 21 is a block diagram showing a method for identifying coefficients of the FIR filter 5 in FIG. 20.

FIG. 22 is a block diagram showing a conventional silencer.

FIG. 23 is a block diagram of an embodiment of the ninth invention.

FIG. 24 is a block diagram showing a method for identifying coefficients of an FIR filter 17 according to an embodiment of the ninth invention.

FIG. 25 is a block diagram showing a method of identifying coefficients of an FIR filter 7 according to an embodiment of the ninth invention.

FIG. 26 is a block diagram of an embodiment of the tenth invention.

FIG. 27 is a block diagram of an eleventh embodiment of the invention.

FIG. 28 is a block diagram showing a method for identifying a coefficient of the FIR filter 17 in the embodiment of the eleventh invention.

FIG. 29 is a block diagram of an embodiment of the twelfth invention.

FIG. 30 is a diagram showing a transfer function from a noise source to each control point when the periodic noise is controlled by a plurality of control points.

FIG. 31 is a block diagram of an embodiment of the thirteenth invention.

FIG. 32 is a block diagram of an embodiment of the fourteenth invention.

[Description of Reference Signs] 1, 1a, 1b Microphones 2, 2a, 2b, 2c Microphones 3, 3a, 3b, 3c, 3d Adaptive filters 4, 4a, 4b, 4c Speakers 5, 5a, 5b, 5c, 5d, 5e, 5f, 5g, 5h
FIR filters 6, 6a, 6b, 6c, 6d, 6e, 6f, 6g, 6h
FIR filters 7, 7a, 7b, 7c, 7d FIR filters 8, 8a, 8b, 8c, 8d, 8e, 8f, 8g, 8h
LMS calculators 9, 9a, 9b, 9c, 9d, 9e, 9f, 9g, 9h
Adder 10, 10a, 10b, 10c, 10d Subtractor 11 Subtractor 12 Measurement noise generator 13a, 13b, 13c, 13d Coefficient adder 14a, 14b Control signal adder 15 Chair 16 Pillow 17, 17a, 17b FIR filter 18 Microphone 19 LMS calculator 20 Subtractor 21 Control circuit block of silencer

─────────────────────────────────────────────────── ─── Continuation of front page (56) Reference JP-A-6-250674 (JP, A) JP-A-6-195089 (JP, A) JP-A-6-195087 (JP, A) JP-A-6- 161466 (JP, A) (58) Fields surveyed (Int.Cl. ⁷ , DB name) G10K 11/178 F01N 1/00 G01H 3/00 H03H 21/00

Claims (19)

(57) [Claims] 1. A noise detector for detecting noise from a noise source, an adaptive filter for adaptively controlling the noise from the noise detector, and a first digital filter for signal-processing the noise from the noise source. A speaker that reproduces the output of the adaptive filter, a second digital filter that processes the output of the adaptive filter, a third digital filter that processes the output of the adaptive filter, and an optimum attenuation effect by noise control An error detector installed apart from the obtained noise control position, a subtractor for subtracting the output of the third digital filter from the output of the error detector, an output of the subtractor and the second Adder for adding the outputs of the digital filters, and the output of the first digital filter and the output of the adder. In a muffler comprising a coefficient calculator for calculating and updating the coefficient of the adaptive filter from the above, a noise control position at which the first digital filter and the second digital filter obtain an optimum attenuation effect from the speaker The muffling apparatus is characterized in that the transfer functions up to are identified in advance as coefficients, and the transfer function from the speaker to the error detector is identified in advance as coefficients in the third digital filter. 2. A noise suppressor for controlling noise at n points using m speakers, wherein the noise suppressor detects noise from a noise source, and m adaptive filters. (M × n) first digital filters, m speakers, and (m
× n) second digital filters, (m × n) third digital filters, and n number of digital filters installed far from the noise control position where the optimum damping effect by noise control is obtained. Error detector, n subtractors, (m
× n) adders, (m × n) coefficient calculators, and m
The output of the noise detector is the j (j = 1, 2, ..., M) th adaptive filter and the jk (k = 1, 2, ..., N) th first digital filter. It is input to the filter and some j (=
The output of the (J) -th adaptive filter is the n-th Jk-th second digital filter and the n-th Jk-th third digital filter.
Of the J-th adaptive filter, and the output of the J-th adaptive filter is reproduced by the J-th speaker. The k-th (= K) -th subtractor subtracts the output of the K-th error detector from the output of the K-th error detector. The output of the digital filter and the output of the (m-1) th J'K (J '≠ J) th third digital filter are subtracted, and the output is input to the m number of jKth adders to obtain JK The output of the Kth subtractor and the output of the JKth second digital filter are added by the th adder, and the JKth coefficient calculator is the output of the JKth first digital filter and the JKth adder. The coefficient for the J-th adaptive filter at the K-th noise control point is obtained from the output of, and the output of the JK-th coefficient calculator and (n
−1) The outputs of the JK ′ (K ′ ≠ K) th coefficient calculators are added to update the coefficient of the Jth adaptive filter, and the JKth first digital filter and the JKth Second
The transfer function from the J-th speaker to the K-th noise control position where the optimum attenuation effect is obtained is identified as a coefficient in advance in the digital filter of No. 3, and the J-th third digital filter has the J-th transfer function. A silencer characterized in that the transfer function from the speaker to the Kth error detector is identified in advance as a coefficient. 3. A muffler for controlling l noises at n points by using m speakers, wherein the muffler detects noises from l noise sources.
Noise detectors, (l × m) adaptive filters, (l × m × n) first digital filters, m speakers, m control signal adders, and l
× m × n) second digital filters and (m × n)
n) third digital filters, n error detectors installed away from the noise control position where the optimum damping effect by noise control is obtained, n subtractors, and (l × m × n) adders, (l × m × n) coefficient calculators, and (l × m) coefficient adders, i (i = 1, 2, ..., L) Of the i-th noise detector, the output of the i-th (= I) -th noise detector is m Ij (j =
The 1, 2, ..., M) th adaptive filter and (m ×)
n) Ijk (k = 1, 2, ..., N) -th first digital filters are input to a certain Ij (= I)
The output of the (J) th adaptive filter is n IJk
The second digital filter and the Jth control signal adder, and the Jth control signal adder outputs the output of the IJth adaptive filter and (l-1) I'J (I '≠ The output of the (I) th adaptive filter is added, and the output of the Jth control signal adder is n Jk
The third (third) digital filter and the Jth speaker are input to the output of the Kth error detector and the output of the JKth third digital filter (m- 1) The output of the J′K (J ′ ≠ K) th third digital filter is subtracted, and the output is input to the (l × m) ijKth adder, and the IJKth adder is added. The output of the K-th subtractor and the output of the IJK-th second digital filter are added, and the IJK-th coefficient calculator outputs K by the output of the IJK-th first digital filter and the output of the IJK-th adder. The coefficient of the IJ-th adaptive filter for the I-th noise by the J-th speaker at the noise control point is calculated, and the output of the IJK-th coefficient calculator and (n- 1) The outputs of the IJK ′ (K ′ ≠ K) th coefficient calculators are added to update the coefficients of the IJth adaptive filter, and the IJKth first digital filter and the IJKth first digital filter are updated. In the second digital filter, the transfer function from the Jth speaker to the Kth noise control position where the optimum attenuation effect is obtained is identified in advance as a coefficient, and in the JKth third digital filter, the transfer function is identified. A silencer characterized in that a transfer function from the Jth speaker to the Kth error detector is identified as a coefficient in advance. 4. A muffler for controlling l noises at n points using m speakers, wherein the muffler detects noises from l noise sources.
Noise detectors, (l × m) adaptive filters, (l × m × n) first digital filters, m speakers, m control signal adders, and m
× n) second digital filters, (m × n) third digital filters, and n number of digital filters installed far from the noise control position where the optimum damping effect by noise control is obtained. Error detector, n subtractors, (m
× n) adders, (l × m × n) coefficient calculators, and (l × m) coefficient adders, and i (i = 1, 2, ..., L) th Of the noise detectors, the output of the i (= I) th noise detector is m Ij (j =
The 1, 2, ..., M) th adaptive filter and (m ×)
n) Ijk (k = 1, 2, ..., N) -th first digital filters are input to a certain Ij (= I)
The output of the (J) th adaptive filter is input to the Jth control signal adder, and the output of the IJth adaptive filter and (l−
1) The outputs of the I'J (I '≠ I) -th adaptive filters are added, and the output of the J-th control signal adder is n Jk-th second digital filters and n Jk-th Of the Kth error detector from the output of the Kth error detector and (m-1) from the output of the Kth error detector by a certain k (= K) th subtractor. ) The output of the J'K (J '≠ K) th third digital filter is subtracted, and the output is m
To the jK-th adder, the JK-th adder adds the output of the K-th subtractor and the output of the JK-th second digital filter, and the output is calculated as l iJK-th coefficient operation. To the IJK-th coefficient calculator
By the output of the JK-th first digital filter and the output of the JK-th adder, J at the K-th noise control point
The coefficient of the IJ-th adaptive filter with respect to the I-th noise caused by the th-th speaker is calculated, and the output of the IJK-th coefficient calculator and (n-
1) The outputs of IJK ′ (K ′ ≠ K) th coefficient calculators are added to update the coefficient of the IJth adaptive filter, and the IJKth first digital filter and the JKth adaptive filter are updated. In the second digital filter, the transfer function from the Jth speaker to the Kth noise control position where the optimum attenuation effect is obtained is identified in advance as a coefficient, and in the JKth third digital filter, the transfer function is identified. A silencer characterized in that a transfer function from the Jth speaker to the Kth error detector is identified as a coefficient in advance. 5. An adaptive filter for adaptively controlling noise, and a first digital filter for signal processing noise,
A speaker that reproduces the output of the adaptive filter, a second digital filter that processes the output of the adaptive filter, a third digital filter that processes the output of the adaptive filter, and an optimum attenuation effect by noise control An error detector installed far from the noise control position where it is obtained, and the output of the third digital filter is subtracted from the output of the error detector, and the output is output to the adaptive filter and the first filter.
A subtractor that uses the noise input signal of the digital filter of
An adder for adding the output of the subtractor and the output of the second digital filter, and a coefficient calculator for calculating and updating the coefficient of the adaptive filter from the output of the first digital filter and the output of the adder In a silencer for controlling periodic noise, the first digital filter and the second digital filter have a coefficient of a transfer function from the speaker to a noise control position where an optimum damping effect is obtained. And a transfer function from the speaker to the error detector is identified in advance as a coefficient in the third digital filter. 6. A silencer for controlling noise at n points using m speakers for periodic noise, wherein the silencer includes m adaptive filters and (m
× n) first digital filters, m speakers, and (m × n) second digital filters,
(M × n) third digital filters, n error detectors installed apart from the noise control position where the optimum damping effect by noise control is obtained, and n subtractors, It has (m × n) adders, (m × n) coefficient calculators, and m coefficient adders, and the noise signal is the j (j = 1, 2, ..., M) -th Signals are respectively processed by the adaptive filter and the jk (k = 1, 2, ..., N) -th first digital filter, and a certain j
The output of the (= J) th adaptive filter is n J
It is input to the k-th second digital filter and n Jk-th third digital filters, and the output of the J-th adaptive filter is reproduced by the J-th speaker and the k-th (= K) -th subtraction is performed. Subtracts the output of the JK-th third digital filter and the output of the (m-1) J'K (J '≠ J) -th third digital filter from the output of the K-th error detector. The output is J
Input as the noise signal of the nth adaptive filter and the nth Jk-th first digital filter, and K
The output of the th subtractor is input to the m jK-th adders, and the JK-th adder adds the output of the K-th subtractor and the output of the JK-th second digital filter,
The JK-th coefficient calculator calculates the coefficient for the J-th adaptive filter at the K-th noise control point from the output of the JK-th first digital filter and the output of the JK-th adder, and the J-th coefficient adder By JK
The output of the th coefficient calculator and (n−1) JK ′ (K ′ ≠
The K) th coefficient calculator output is added to update the coefficient of the Jth adaptive filter, and the JKth first digital filter and the JKth second filter are added.
The transfer function from the J-th speaker to the K-th noise control position where the optimum attenuation effect is obtained is identified as a coefficient in advance in the digital filter of No. 3, and the J-th third digital filter has the J-th transfer function. A silencer characterized in that the transfer function from the speaker to the Kth error detector is identified in advance as a coefficient. 7. A silencer for controlling noise at n locations using m speakers for periodic noise, wherein the silencer comprises (m × n) adaptive filters and (m × n) adaptive filters. A first digital filter, m speakers, (m × n) second digital filters, (m × n) third digital filters, and an optimum attenuation effect by noise control N error detectors installed away from the noise control position of
The number of subtracters, the number of (m × n) adders, the number of (m × n) coefficient calculators, and the number of m control signal adders are included. Noise signals are jk (j = 1, 2). ,,, m; k = 1, 2,
..., n) -th adaptive filter and jk-th first
Signal is processed by each digital filter of
The output of the k (= JK) th adaptive filter and (n
−1) The outputs of the JK ′ (K ′ ≠ K) th adaptive filters are added by the Jth control signal adder,
The output of the Jth control signal adder is the nth Jkth second
, And the nth Jk-th third digital filter, and the output of the J-th control signal adder is reproduced by the J-th speaker, and the K-th subtractor subtracts the K-th error detector. From the output of the JK-th third digital filter and (m-1) J '
The output of the K (J ′ ≠ J) th third digital filter is subtracted, and the output is input as noise signals of the m jKth adaptive filters and the m jKth first digital filters, Further, the output of the Kth subtractor is input to the m jKth adders, and the JKth adder adds the output of the Kth subtractor and the output of the JKth second digital filter to JKth adder. The th coefficient calculator is configured to obtain and update the coefficient of the Jth adaptive filter by the output of the JKth first digital filter and the output of the JKth adder. And JK th second
The transfer function from the J-th speaker to the K-th noise control position where the optimum attenuation effect is obtained is identified as a coefficient in advance in the digital filter of No. 3, and the J-th third digital filter has the J-th transfer function. A silencer characterized in that the transfer function from the speaker to the Kth error detector is identified in advance as a coefficient. 8. A silencer for controlling noise at n points using m speakers for periodic noise, wherein the silencer comprises (m × n) adaptive filters and (m × n) adaptive filters. A first digital filter, m speakers, (m × n) second digital filters, (m × n) third digital filters, and an optimum attenuation effect by noise control N error detectors installed away from the noise control position of
The number of subtracters, the number of (m × n) adders, the number of (m × n) coefficient calculators, and the number of m control signal adders are included. Noise signals are jk (j = 1, 2). ,,, m; k = 1, 2,
..., n) -th adaptive filter and jk-th first
Signal is processed by each digital filter of
The output of the k (= JK) th adaptive filter is JK
The second digital filter and the Jth control signal adder are input to the Jth control signal adder.
The output of the th adaptive filter and (n-1) J
The outputs of the K ′ (K ′ ≠ K) th adaptive filter are added, the output of the Jth control signal adder is input to the n Jkth third digital filters, and the Jth control signal addition is performed. The output of the filter is reproduced by the Jth speaker, and the output of the KKth error detector to the output of the JKth third digital filter and (m-1) J'K (J The output of the '≠ J) th third digital filter is subtracted, and the output is input as noise signals of the m jK-th adaptive filters and the m jK-th first digital filters, and further the K-th The output of the subtractor is input to m jKth adders,
The output of the Kth subtractor and JK by the JKth adder
Add the outputs of the second digital filter of the th
The th coefficient calculator is configured to obtain and update the coefficient of the Jth adaptive filter by the output of the JKth first digital filter and the output of the JKth adder. And JK th second
The transfer function from the J-th speaker to the K-th noise control position where the optimum attenuation effect is obtained is identified as a coefficient in advance in the digital filter of No. 3, and the J-th third digital filter has the J-th transfer function. A silencer characterized in that the transfer function from the speaker to the Kth error detector is identified in advance as a coefficient. 9. A noise detector for detecting noise from a noise source, an adaptive filter for adaptively controlling the noise from the noise detector, and a first digital filter for signal-processing the noise from the noise source. A speaker that reproduces the output of the adaptive filter, a second digital filter that processes the output of the adaptive filter, a third digital filter that processes the output of the adaptive filter, and an optimum attenuation effect by noise control An error detector installed away from the noise control position to obtain, a fourth digital filter for signal processing the output of the error detector, and an output of the fourth digital filter from the third digital filter. A subtractor for subtracting the output of the filter, the output of the subtractor and the second
A silencer comprising an adder for adding the outputs of the digital filters, and a coefficient calculator for calculating and updating the coefficient of the adaptive filter from the output of the first digital filter and the output of the adder, The first digital filter and the second digital filter previously identify the transfer function from the speaker to the noise control position where the optimum attenuation effect is obtained as a coefficient, and the third digital filter identifies the transfer function from the speaker to the noise control position. The transfer function from the noise detector to the fourth digital filter is identified in advance as a coefficient, and the transfer function from the noise source to the error detector and the optimum attenuation effect from the noise source are included in the fourth digital filter. Is characterized in that the ratio with the transfer function up to the noise control position where Sound equipment. 10. A noise suppressor for controlling noise at m points using m speakers, wherein the noise suppressor detects noise from a noise source, and m adaptive filters. (M × n) first digital filters, m speakers, and (m
× n) second digital filters, (m × n) third digital filters, and n number of digital filters installed far from the noise control position where the optimum damping effect by noise control is obtained. Error detector, n fourth digital filters, n subtractors, (m × n) adders, (m × n) coefficient calculators, and m coefficient adders The output of the noise detector is input to the j (j = 1, 2, ..., M) th adaptive filter and the jk (k = 1, 2, ..., N) th first digital filter. , Some j (=
The output of the (J) -th adaptive filter is the n-th Jk-th second digital filter and the n-th Jk-th third digital filter.
, The output of the Jth adaptive filter is reproduced by the Jth speaker, and the output of a certain k (= K) th error detector is signal-processed by the Kth fourth digital filter. , The Kth subtractor outputs the output of the Kth fourth digital filter to the output of the JKth third digital filter (m
−1) The output of the J′K (J ′ ≠ J) th digital filter is subtracted, and the output is input to the m jK th adder, and the JK th adder outputs the K th The output of the subtractor of J.sub.K and the output of the JK.sup.th second digital filter are added, and the JK.sup.th coefficient calculator outputs the K.sup. The coefficient for the J-th adaptive filter at the control point is calculated, and the output of the JK-th coefficient calculator and (n-1) J-th coefficient are calculated by the J-th coefficient adder.
The output of the K ′ (K ′ ≠ K) th coefficient calculator is added to update the coefficient of the Jth adaptive filter. The JKth first digital filter and the JKth second filter are configured.
The transfer function from the J-th speaker to the K-th noise control position where the optimum attenuation effect is obtained is identified as a coefficient in advance in the digital filter of No. 3, and the J-th third digital filter has the J-th transfer function. The transfer function from the speaker to the K-th fourth digital filter through the K-th error detector is identified in advance as a coefficient, and the K-th fourth digital filter includes a noise source from the K-source. A silencer characterized in that the ratio of the transfer function up to the th error detector and the transfer function up to the Kth noise control position where the optimum damping effect is obtained from the noise source is identified as a coefficient in advance. 11. An adaptive filter for adaptively controlling noise, a first digital filter for signal processing of noise, a speaker for reproducing the output of the adaptive filter, and a second digital for signal processing of the output of the adaptive filter. A filter, and a third digital filter for signal processing the output of the adaptive filter,
Of the error detector installed at a place distant from the noise control position for obtaining the optimum damping effect by the noise control, the fourth digital filter for signal processing the output of the error detector, and the fourth digital filter. A subtracter that subtracts the output of the third digital filter from the output and uses the output as a noise input signal of the adaptive filter and the first digital filter, and the output of the subtractor and the output of the second digital filter. An adder for adding
A silencer for controlling periodic noise, comprising: a coefficient calculator for calculating and updating a coefficient of the adaptive filter from an output of the first digital filter and an output of the adder. The transfer function up to the noise control position where the optimum attenuation effect is obtained from the speaker is previously identified as a coefficient in the filter and the second digital filter, and the error detector from the speaker is included in the third digital filter. The transfer function up to the fourth digital filter is previously identified as a coefficient, and the transfer function from the noise source to the error detector and the optimum attenuation effect from the noise source are obtained in the fourth digital filter. A silencer characterized in that a ratio with a transfer function up to a noise control position is identified as a coefficient in advance. 12. A method for producing periodic noise using m speakers
In a muffler for controlling noise at a location, the muffler includes m adaptive filters and (m
× n) first digital filters, m speakers, and (m × n) second digital filters,
(M × n) third digital filters, n error detectors installed apart from the noise control position where the optimum damping effect by noise control is obtained, and n fourth digital filters Filter, n subtractors, (m ×
n) adders, (m × n) coefficient calculators, and m coefficient adders, and the noise signal is the j (j = 1, 2, ..., M) th adaptive filter The signal is processed by the jk (k = 1, 2, ..., N) -th first digital filter, and a certain j
The output of the (= J) th adaptive filter is n J
The k-th second digital filter and n Jk-th third digital filters are input, and the output of the J-th adaptive filter is reproduced by the J-th speaker, and the k-th (= K) -th error The output of the detector is subjected to signal processing by the Kth fourth digital filter, and the output of the Kth fourth digital filter to the output of the JKth third digital filter and (m-1) are output by the Kth subtractor. ) The output of the J′K (J ′ ≠ J) th third digital filter is subtracted, and the output is input as the noise signal of the Jth adaptive filter and n Jkth first digital filters. Further, the output of the K-th subtractor is input to the m jK-th adders, and the JK-th adder outputs the output of the K-th subtractor and the JK-th second digital filter. The JK-th coefficient calculator calculates the coefficient for the J-th adaptive filter at the K-th noise control point from the output of the JK-th first digital filter and the output of the JK-th adder. , The output of the JK-th coefficient calculator and the output of the (n-1) JK '(K' ≠ K) -th coefficient calculator are added by the J-th coefficient adder to obtain the coefficient of the J-th adaptive filter. To update the JK-th first digital filter and the JK-th second digital filter.
The transfer function from the J-th speaker to the K-th noise control position where the optimum attenuation effect is obtained is identified as a coefficient in advance in the digital filter of No. 3, and the J-th third digital filter has the J-th transfer function. The transfer function from the speaker to the K-th fourth digital filter through the K-th error detector is identified in advance as a coefficient, and the K-th fourth digital filter includes a noise source from the K-source. A silencer characterized in that the ratio of the transfer function up to the th error detector and the transfer function up to the Kth noise control position where the optimum damping effect is obtained from the noise source is identified as a coefficient in advance. 13. A method for producing periodic noise using m speakers
In a muffler for controlling noise at a location, the muffler includes (m × n) adaptive filters, (m × n) first digital filters, m speakers, and (m × n) ) Second digital filters, (m × n) third digital filters, and n error detections installed apart from the noise control position where the optimum damping effect by noise control is obtained Vessel, n
Fourth digital filters, n subtractors,
It has (m × n) adders, (m × n) coefficient calculators, and m control signal adders, and noise signals are jk (j = 1, 2, ..., M; k). = 1, 2,
..., n) -th adaptive filter and jk-th first
Signal is processed by each digital filter of
The output of the k (= JK) th adaptive filter and (n
−1) The outputs of the JK ′ (K ′ ≠ K) th adaptive filters are added by the Jth control signal adder,
The output of the Jth control signal adder is the nth Jkth second
To the Nth digital filter and the Nth Jkth third digital filter, the output of the Jth control signal adder is reproduced by the Jth speaker, and the output of the Kth error detector is input to the Kth error detector. The signal is processed by the fourth digital filter, and the output of the Kth fourth digital filter to the output of the JKth third digital filter and (m-1) J'K (J '≠ J)
The output of the th-th third digital filter is subtracted, and the output is input as the noise signals of the m-th jK-th adaptive filter and the m-th jK-th first digital filter. The output is input to the m-th jK-th adder, and the JK-th adder adds the output of the K-th subtractor and the output of the JK-th second digital filter. The JK-th first digital filter and the JK-th second digital filter are configured to obtain and update the coefficient of the J-th adaptive filter by the output of the th-th first digital filter and the output of the JK-th adder.
The transfer function from the J-th speaker to the K-th noise control position where the optimum attenuation effect is obtained is identified as a coefficient in advance in the digital filter of No. 3, and the J-th third digital filter has the J-th transfer function. The transfer function from the speaker to the K-th fourth digital filter through the K-th error detector is identified in advance as a coefficient, and the K-th fourth digital filter includes a noise source from the K-source. A silencer characterized in that the ratio of the transfer function up to the th error detector and the transfer function up to the Kth noise control position where the optimum damping effect is obtained from the noise source is identified as a coefficient in advance. 14. A periodic noise is generated by using m speakers to generate n
In a muffler for controlling noise at a location, the muffler includes (m × n) adaptive filters, (m × n) first digital filters, m speakers, and (m × n) ) Second digital filters, (m × n) third digital filters, and n error detections installed apart from the noise control position where the optimum damping effect by noise control is obtained Vessel, n
Fourth digital filters, n subtractors,
It has (m × n) adders, (m × n) coefficient calculators, and m control signal adders, and noise signals are jk (j = 1, 2, ..., M; k). = 1, 2,
..., n) -th adaptive filter and jk-th first
Signal is processed by each digital filter of
The output of the k (= JK) th adaptive filter is JK
The second digital filter and the Jth control signal adder are input to the Jth control signal adder.
The output of the th adaptive filter and (n-1) J
The outputs of the K ′ (K ′ ≠ K) th adaptive filter are added, the output of the Jth control signal adder is input to the n Jkth third digital filters, and the Jth control signal addition is performed. The output of the Kth error detector is reproduced by the Jth speaker, the output of the Kth error detector is signal-processed by the Kth fourth digital filter, and the output of the Kth fourth digital filter is output by the Kth subtractor. From the output of the JK-th third digital filter and the output of the (m−1) th J′K (J ′ ≠ J) -th third digital filter to obtain the output of the m-th jK-th digital filter. The noise signals of the adaptive filter and the m-th jK-th first digital filters are input, and the output of the K-th subtractor is input to the m-th jK-th adder, and the JK-th adder outputs the K-th of The output of the subtractor and the output of the JKth second digital filter are added, and the JKth coefficient calculator outputs the output of the JKth first digital filter and the JKth coefficient.
The output of the th adder is used to obtain and update the coefficient of the J th adaptive filter. The JK th first digital filter and the JK th second filter are arranged.
The transfer function from the J-th speaker to the K-th noise control position where the optimum attenuation effect is obtained is identified as a coefficient in advance in the digital filter of No. 3, and the J-th third digital filter has the J-th transfer function. The transfer function from the speaker to the K-th fourth digital filter through the K-th error detector is identified in advance as a coefficient, and the K-th fourth digital filter includes a noise source from the K-source. A silencer characterized in that the ratio of the transfer function up to the th error detector and the transfer function up to the Kth noise control position where the optimum damping effect is obtained from the noise source is identified as a coefficient in advance. 15. The silencer according to any one of claims 1 to 14, wherein the speaker and the error detector are installed near the head. 16. The silencer according to claim 15, wherein the speaker and the error detector are installed on the headrest. 17. The muffler according to claim 16, wherein the headrest has a recess that inevitably fixes the head. 18. The silencer according to claim 16, wherein the headrest is provided on a chair. 19. The muffler according to claim 16, wherein the headrest is a pillow.