JPH11126082A - Muffling device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a muffling device capable of reducing noise even when a noise transferring time is short by reducing the processing delay time when an active noise control is used. SOLUTION: This device is a muffling device in which an active noise control digitally processing the noise from a noise source is used. Here, a transfer function from a speaker 9 till a microphone 1b being an error detector is preliminarily identified in an Fx filter 10 being a transfer function corrector and the sound pressuer of the sound pressure frequency characteristic of the speaker 9 is made small in a band higher than a Nyquist frequency and the processing delay time is reduced by making the Nyquist frequency equal to or smaller than the sound pressure level at the position of the microphone 1b.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

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

[0002]

2. Description of the Related Art In recent years, there has been proposed an active noise apparatus which outputs a control sound from a loudspeaker using a digital signal processing technology to mitigate air conditioning noise or indoor noise of a factory or an automobile and mute the noise.

Hereinafter, a conventional silencer will be described with reference to FIG. FIG. 12 shows a block diagram of a conventional silencer. In the figure, reference numerals 1a and 1b denote microphones, which are a noise detector and an error detector, respectively.
a, 2b are microphone amplifiers, 3a, 3b, 3c are low-pass filters, 4a, 4b are sample-and-hold circuits, 5a,
5b is an AD converter, 6 is an adaptive filter, 7
Denotes a DA converter, 8 denotes a power amplifier, 9a denotes a speaker, 10 denotes an Fx filter which is a transfer function corrector, and 11 denotes an LMS calculator which is a coefficient calculator.

[0004] The operation of the thus configured silencer will be described below. First, noise is detected by the microphone 1a, and the detection signal is transmitted to the microphone amplifier 2
After being amplified by a, the high-pass component is removed by the low-pass filter 3a. The output of the low-pass filter 3a is sampled by the sample-and-hold circuit 4a at each sampling cycle, and the output is converted to a digital signal by the AD converter 5a. The output of the AD converter 5a is input to the adaptive filter 6 and the Fx filter 10.
Then, the noise signal processed by the adaptive filter 6 is restored to an analog signal by the DA converter 7, and after the high-frequency component is removed by the low-pass filter 3c,
The signal is amplified by the power amplifier 8 and output from the speaker 9a.

[0005] The microphone 1b installed at the noise control point detects a sound in which the reproduced sound from the speaker 9a and the noise from the noise source interfere with each other. And the microphone 1b
Is amplified by the microphone amplifier 2b,
The high-pass component is removed by the low-pass filter 3b. The output of the low-pass filter 3b is sampled by the sample-and-hold circuit 4b at each sampling cycle, and the output is converted to a digital signal by the AD converter 5b. On the other hand, the Fx filter 10 has the AD converter 5b from the DA converter 7 via the speaker 9a and the microphone 1b.
The transfer function up to AD is previously obtained as a coefficient,
The signal from converter 5a is processed using this coefficient, and input to LMS calculator 11. LMS calculator 11
Is the output from the Fx filter 10 and the AD converter 5
Using the output from b, the coefficient of the adaptive filter 6 is updated by performing an LMS operation (least square method) so that the output signal from the AD converter 5b is minimized. As a result, noise is attenuated in the microphone 1b by the control sound from the speaker 9a.

[0006] This method is applied to the Filtered-x LMS algorithm (for example, see B. Widrow and S. Stearns, "A
daptive Signal Processing ”(Prentice-Hall, Englewo
od Cliffs, NJ, 1985). Using this, the coefficient update of the adaptive filter 6 can be expressed as follows by using a mathematical expression.

[0007] w (n + 1) = w (n) + αxT (n) c (n) e (n) = w (n) + αr (n) e (n)

Here, wT (n) = {w0 (n), w1 (n),..., WN-1 (n)} r (n) = xT (n) c (n) xT (n) = ｛ x (n), x (n−1),..., x (n−N + 1)｝ cT (n) = {c0 (n), c1 (n),..., cN−1 (n)} yT ( n) = {y (n), y (n-1),..., y (n-N + 1)} y (n) = xT (n) w (n) where w (n); coefficient of the adaptive filter 6 (The number of taps is N) α; step parameter r (n); output signal of Fx filter 10 e (n); output signal of microphone 1b c (n); transfer function from DA converter 7 to AD converter 5b

[0009]

However, as shown in FIG. 12, the noise of the noise source is detected by the microphone 1a, and the adaptive filter 6 generates a control signal.
a to reduce the output. In the case of the so-called feedforward control, as shown in FIG. 13, the time for the noise to transmit the distance from the noise source to the control point (in this case, the position of the microphone 1b) and the speaker 9a after the noise is detected by the microphone 1a. More control sound is output and the microphone 1b
Must be equal (the adaptive filter 6 performs the fine adjustment to equalize).

[0010] If the noise transmission time, the microphone 1
If the control sound is output from the loudspeaker 9a and reaches the microphone 1b after the noise is detected by a, it is not possible to reduce the noise, but there is a risk of adding the noise. is there. However, in most cases of noise control, the time required for each element shown in FIG. 13 must be reduced because the noise transmission time is short (the distance from the noise source to the control point is short). Speaker 9a
Since it is difficult to adjust the group delay and the spatial propagation time from the speaker 9a to the microphone 1b, it is necessary to reduce the processing delay time in other circuit parts.

For example, the low-pass filters 3a and 3c
Since it is an anti-aliasing filter, as shown in FIG.
Since a sharp cutoff characteristic (usually a fifth- or sixth-order Chebyshev characteristic is used) having an amplitude characteristic sufficiently attenuated at the Nyquist frequency fn is used, the group delay increases. On the other hand, in order to reduce the group delay of the low-pass filters 3a and 3c, the ADC 5a and the DAC 7 are each used as a primary or secondary filter by using a delta-sigma 1-bit AD converter or a delta-sigma 1-bit DA converter. However, since the delta-sigma type has a large delay, the above problem cannot be solved.

In order to reduce the operation delay time, F
If the number of taps of the x filter is reduced, it is not possible to accurately form the coefficient in the low-frequency component of the Fx filter.

Furthermore, a high-frequency signal outside the noise control frequency band makes the growth of the adaptive filter unstable, and the adaptive filter coefficient may diverge.

SUMMARY OF THE INVENTION It is an object of the present invention to reduce the processing delay time of the silencer, to reduce noise even when the noise transmission time is short or when the number of taps of the Fx filter is small, and to operate the silencer stably. I do.

[0015]

SUMMARY OF THE INVENTION In order to solve this problem, a noise suppressor according to the present invention includes a noise detector for detecting noise from a noise source and a high-frequency component of the noise from the noise detector. A first low-pass filter, a first AD conversion circuit that samples an analog output of the first low-pass filter and quantizes the digital signal, and adaptively controls the digital signal from the first AD conversion circuit. An adaptive filter, a transfer function corrector for signal processing the digitally converted noise signal from the first AD conversion circuit, a DA converter for converting the output of the adaptive filter into an analog signal, and an output of the DA converter And an error detector installed at a control position that reduces noise due to interference between control sound from the speaker and noise from a noise source. A second low-pass filter that removes a high-frequency component of the error signal detected by the error detector, and a second that samples an analog output of the second low-pass filter and quantizes the digital signal. An AD converter, and a coefficient calculator for calculating and updating coefficients of the adaptive filter from an output of the transfer function corrector and an output of the second AD converter circuit, wherein the transfer function corrector includes the speaker The transfer function from the to the error detector is identified as a coefficient in advance, and the sound pressure frequency characteristics of the speaker are reduced sufficiently in a band higher than half the sampling frequency (Nyquist frequency) to reduce the sound pressure. The sound pressure frequency characteristics indicate that the sound pressure level in the frequency band above the Nyquist frequency is lower than the sound pressure level in the frequency band in which noise is reduced at the position of the error detector. The level is obtained by reducing the processing delay time as equal to or less than.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS According to a first aspect of the present invention, there is provided a noise reduction device for detecting a noise from a noise source and a high frequency component of the noise from the noise detector. 1 low-pass filter, a first AD conversion circuit that samples an analog output of the first low-pass filter and quantizes the digital signal, and an adaptive control that adaptively controls the digital signal from the first AD conversion circuit. A filter, a transfer function corrector that also processes the digitally converted noise signal from the first AD conversion circuit, a DA converter that converts the output of the adaptive filter into an analog signal, and an output of the DA converter. A speaker to be reproduced; an error detector installed at a control position for reducing noise by interference between control noise from the speaker and noise from a noise source; A second low-pass filter for removing high-frequency components of the error signal detected by the error detector, and a second AD converter quantizes the digital signal by sampling the analog output of said second low-pass filter,
A coefficient calculator for calculating and updating coefficients of the adaptive filter from an output of the transfer function corrector and an output of the second AD conversion circuit, wherein the transfer function corrector includes an error detector from the speaker. The transfer function of the speaker is previously identified as a coefficient, and the sound pressure frequency characteristic of the speaker is sufficiently reduced in a band higher than half the sampling frequency (Nyquist frequency), and the sound pressure frequency characteristic of the speaker is reduced. The processing delay time is reduced by setting the sound pressure level in a band equal to or higher than the Nyquist frequency equal to or lower than the sound pressure level in the frequency band in which noise is reduced at the position of the error detector.

According to a second aspect of the present invention, a digital IIR filter for correcting a low-frequency transfer function is added to an input side or an output side of the transfer function corrector according to the first aspect. That is, even if the number of taps of the Fx filter used as the transfer function corrector is reduced, the coefficient in the low-frequency component of the Fx filter can be accurately formed.

According to a third aspect of the present invention, a high-frequency adaptive filter coefficient outside a desired noise control frequency band is removed by a coefficient calculator by removing a high-frequency component on an input side or an output side of a transfer function corrector. A first digital IIR filter that prevents updating is added, and high-frequency components are removed from the output of the second AD converter, and the coefficient arithmetic unit prevents updating of high-frequency adaptive filter coefficients outside the desired noise control frequency band. Second
, And the first digital IIR filter and the second digital IIR filter.
The R filter has substantially the same characteristics, removes high-frequency components, and can prevent the coefficient calculator from updating high-frequency adaptive filter coefficients other than the desired noise control frequency band.

According to a fourth aspect of the present invention, the sound pressure frequency characteristic of the loudspeaker according to the first to third aspects is more than the sound pressure level in the frequency band where noise is reduced at the position of the error detector.
The sound pressure level in the band equal to or higher than the Nyquist frequency is set to be equal to or lower than the same, so that the processing delay time of the silencer can be reduced by the group delay of the low-pass filter 3a and the cost can be reduced.

According to a fifth aspect of the present invention, the analog signals of the first and second AD converters are converted into digital signals and input to the adaptive filter and the transfer function corrector. The conversion time is less than half the sampling period (= reciprocal of the sampling frequency), and the delay time can be reduced.

According to a sixth aspect of the present invention, the first and second AD converters according to the first to third aspects are each of a successive approximation type, and the delay time can be reduced.

According to a seventh aspect of the present invention, the time required for converting a digital signal from the adaptive filter of the DA converter into an analog signal is set to be less than half the sampling period (= reciprocal of the sampling frequency). Thus, the delay time can be reduced.

[0023] The invention according to claim 8 provides the DA of claim 7
The converter is of a ladder type, and the delay time can be reduced.

FIG. 1 shows an embodiment of the present invention.
11 will be described with reference to FIG. (Embodiment 1) FIG. 1 is a block diagram showing a muffler according to an embodiment of the present invention. In the figure, 1
Microphones a and 1b are a noise detector and an error detector, 2a and 2b are microphone amplifiers, 3a and 3b
Is a low-pass filter, 4a and 4b are sample hold circuits, 5a and 5b are AD converters, 6 is an adaptive filter, 7 is a DA converter, 8 is a power amplifier, 9 is a speaker, and 10 is an Fx filter which is a transfer function corrector. , 11 are LMS calculators which are coefficient calculators.

The operation of the muffler configured as described above will be described below. First, noise is detected by the microphone 1a, and the detection signal is transmitted to the microphone amplifier 2
After being amplified by a, the high-pass component is removed by the low-pass filter 3a. The output of the low-pass filter 3a is sampled by the sample-and-hold circuit 4a at each sampling cycle, and the output is converted to a digital signal by the AD converter 5a. The output of the AD converter 5a is input to the adaptive filter 6 and the Fx filter 10. The noise signal processed by the adaptive filter 6 is restored to an analog signal by the DA converter 7, amplified by the power amplifier 8, and output from the speaker 9.

Here, the sound pressure frequency characteristic of the speaker 9 is as follows.
For example, as shown in FIG. 2, since the sound pressure level in a band equal to or higher than the Nyquist frequency fn is sufficiently lower than the sound pressure level in the frequency bands f1 and f2 having a noise control effect, the output of the DA converter 7 Even if there is no limiting low-pass filter, the signal is not reproduced from the speaker 9 (or the reproduction level is small), so that the influence of aliasing (characteristic indicated by a bold line with fn being axially symmetric) does not matter.

Next, the microphone 1b installed at the noise control point detects a sound in which the reproduced sound from the speaker 9 and the noise from the noise source interfere with each other. After the error signal from the microphone 1b is amplified by the microphone amplifier 2b, the high-frequency component is removed by the low-pass filter 3b. The output of the low-pass filter 3b is supplied to the sample-and-hold circuit 4
b, sampling is performed at each sampling cycle, and the output is A
It is converted into a digital signal by the D converter 5b.

On the other hand, in the Fx filter 10, a transfer function from the DA converter 7 to the AD converter 5b via the speaker 9 and the microphone 1b is obtained in advance as a coefficient, and a signal from the AD converter 5a is used as a coefficient. It is processed and input to the LMS calculator 11. LMS
The arithmetic unit 11 uses the output from the Fx filter 10 and the output from the AD converter 5b,
The LMS operation (least square method) is performed so that the output signal from is minimized, and the coefficient of the adaptive filter 6 is updated. Thereby, the speaker 9 in the microphone 1b
The noise is attenuated by the control sound from.

As described above, by setting the sound pressure frequency characteristic of the speaker 9 to a sufficiently low level in a band higher than the Nyquist frequency fn as shown in FIG. 2, the DA required in the conventional example shown in FIG. Since the low-pass filter 3c after the converter 7 can be omitted, the processing delay time of the silencer can be reduced by the group delay of the low-pass filter 3c in FIG. Thereby, the distance from the noise detection point to the control point can be shortened. Further, the cost can be reduced accordingly.

Note that the sound pressure level of the noise at the microphone 1a is set as shown in FIG.
When the sound pressure level in the band above the Nyquist frequency fn is smaller than the sound pressure level after noise reduction in the above, the low-pass filter 3a becomes unnecessary and the low-pass filter 3
The processing delay time of the silencer can be reduced by the group delay of a, and the cost can be reduced.

However, when the sound pressure level differences are almost equal as shown in FIG. 4, there is a risk of affecting the noise control effect. In this case, the amplitude frequency characteristic of the low-pass filter 3a is made to have a gradual cutoff characteristic as shown in FIG. 5, and the band above the Nyquist frequency fn is reduced to 10 dB.
By blocking as described above, unnecessary high-frequency components can be prevented from affecting the noise control effect, and the low-pass filter 3 is used because the sharp characteristics shown in FIG. 14 are not used.
The group delay of a can be minimized.

Further, the noise is transmitted from the noise source to the microphone 1.
In many cases, the high-frequency component is attenuated in the distance, or is sound-insulated or absorbed during transmission to the microphone 1b.
As shown in FIG. 3, since the sound pressure frequency characteristic of the noise detected in the step S1 has a sufficient sound pressure level difference between the noise control bands f1 and f2 and the band equal to or higher than the Nyquist frequency fn, the low-pass filter 3b becomes unnecessary. Cost can be reduced. When the noise characteristic of the microphone 1b is as shown in FIG.
As shown in FIG. 5, by cutting off the band above the Nyquist frequency fn by 10 dB or more by making the cut-off stage characteristic gentle, it is possible to prevent unnecessary high frequency components from affecting the noise control effect.

Next, as shown in FIG. 6, the time from when the noise signal is sampled by the sample and hold circuit 4a to when it is quantized by the AD converter 5a and input to the adaptive filter 6 is half the sampling period. Similarly, when the time from when the control signal is output from the adaptive filter 6 to when the analog signal is restored by the DA converter 7 is half the sampling period,
Since the input / output delay time of the adaptive filter 6 is exactly one sample period, the delay time can be stably reduced.

When the noise control band is sufficiently lower than the Nyquist frequency as shown in FIG. 3, the input / output delay time of the adaptive filter 6 is 1 as shown in FIG.
AD conversion time and DA
Even if the conversion time is reduced, the noise control effect is not adversely affected, and the processing delay time of the entire silencer can be reduced. Further, in order to reduce the AD conversion time, a delta-sigma type, which has recently become mainstream in the acoustic field, is required.
Since the bit AD converter has a long delay time, it is preferable to use the successive approximation type. Further, the data transfer between the adaptive filter 6 and the AD converter 5a and the data transfer between the adaptive filter 6 and the DA converter 7 are preferably not parallel serial transfer which requires a long time but parallel transfer which does not require a long transfer time.

In the present embodiment, the AD converter is constituted by a sample-hold circuit and an AD converter. However, an AD converter having a built-in sample-hold circuit may be used. Further, a delta-sigma 1-bit AD converter may be used as an AD converter for quantizing the detection signal of the microphone 1b. In that case, there is also an advantage that the low-pass filter 3b becomes unnecessary.

Further, a low-pass filter for limiting the band of the output of the DA converter 7 is not required, and the low-pass filter 3
In order to prevent an adverse effect on a high frequency band due to a gentle cutoff characteristic of a, the cutoff frequency of the low-pass filter 3b is set to be equal to or lower than the Nyquist frequency.
Updating of coefficients other than the noise control band may be suppressed.

(Embodiment 2) FIG. 8 is a block diagram showing a silencer according to another embodiment of the present invention. FIG.
Microphones 1a and 1b are noise detectors and error detectors, 2a and 2b are microphone amplifiers, 3a and 3b are low-pass filters, 4a and 4b are sample-hold circuits, 5a and 5b are AD converters, and 6 is An adaptive filter, 7 is a DA converter, 8 is a power amplifier, 9b is a speaker, 10 is an Fx filter as a transfer function corrector, 11 is an LMS calculator as a coefficient calculator, and 12a is a digital IIR filter. .

The operation of the muffler configured as described above will be described below. First, noise is detected by the microphone 1a, and the detection signal is amplified by the microphone amplifier 2a, and then the high-pass component is removed by the low-pass filter 3a. The output of the low-pass filter 3a is
The sample and hold circuit 4a samples at each sampling period, and the output is converted to a digital signal by the AD converter 5a. The output of the AD converter 5a is
Adaptive filter 6 and digital IIR filter 12
a, and the output of the digital IIR filter 12a is input to the Fx filter 10.

The noise signal processed by the adaptive filter 6 is restored to an analog signal by the DA converter 7 and then amplified by the power amplifier 8 and output from the speaker 9b. Here, the digital IIR filter 12a is for correcting the low-frequency transfer function when the number of taps of the Fx filter 10 is reduced.

Next, the microphone 1b installed at the noise control point detects a sound in which the reproduced sound from the speaker 9b interferes with the noise from the noise source. And the microphone 1b
Is amplified by the microphone amplifier 2b,
The high-pass component is removed by the low-pass filter 3b. The output of the low-pass filter 3b is sampled by the sample-and-hold circuit 4b at each sampling cycle, and the output is converted to a digital signal by the AD converter 5b.

On the other hand, the Fx filter 10 receives the AD signal from the DA converter 7 via the speaker 9b and the microphone 1b.
A transfer function up to the converter 5b is obtained in advance as a coefficient, and a signal from the AD converter 5a is processed using the coefficient and input to the LMS calculator 11. LM
The S operator 11 outputs the output from the Fx filter 10 and the AD
Using the output from the converter 5b, the AD converter 5
The coefficient of the adaptive filter 6 is updated by performing an LMS operation (least square method) so that the output signal from b becomes minimum. As a result, noise is attenuated by the control sound from the speaker 9b in the microphone 1b.

In the second embodiment, a digital IIR filter 12a is added to the first embodiment, whereby even if the number of taps of the Fx filter 10 is reduced, the coefficient in the low-frequency component of the Fx filter 10 can be accurately determined. It can be formed.

Since the effects other than the digital IIR filter 12a are the same, the description is omitted.

FIG. 9 is a developed example in which the configuration order of the digital IIR filter 12a and the Fx filter 10 of the above embodiment is exchanged, and has the same effect.

(Embodiment 3) FIG. 10 is a block diagram showing a silencer according to another embodiment of the present invention. In the figure, reference numerals 1a and 1b denote microphones which are a noise detector and an error detector, 2a and 2b denote microphone amplifiers, 3a and 3b denote low-pass filters, 4a and 4b denote sample hold circuits, 5a and 5b denote AD converters, 6 is an adaptive filter, 7 is a DA converter, 8 is a power amplifier, 9b is a speaker, 10 is an Fx filter as a transfer function corrector, 11 is an LMS calculator as a coefficient calculator, and 12b and 12c are the same. This is a digital IIR filter having characteristics.

The operation of the thus configured silencer will be described below. First, noise is detected by the microphone 1a, and the detection signal is transmitted to the microphone amplifier 2
After being amplified by a, the high-pass component is removed by the low-pass filter 3a. The output of the low-pass filter 3a is sampled by the sample-and-hold circuit 4a at each sampling cycle, and the output is converted to a digital signal by the AD converter 5a. The output of the AD converter 5a is supplied to the adaptive filter 6 and the digital IIR filter 12b.
And the output of the digital IIR filter 12b is F
Input to x filter 10. The noise signal processed by the adaptive filter 6 is converted to a DA converter 7
Is restored to an analog signal by the power amplifier 8
And is output from the speaker 9b.

Next, the microphone 1b installed at the noise control point detects a sound in which the reproduced sound from the speaker 9b interferes with the noise from the noise source. And the microphone 1b
Is amplified by the microphone amplifier 2b,
The high-pass component is removed by the low-pass filter 3b. The output of the low-pass filter 3b is sampled at each sampling cycle by the sample-and-hold circuit 4b, the output is converted to a digital signal by the AD converter 5b, and the output of the AD converter 5b is input to the digital IIR filter 12c. On the other hand, the Fx filter 10 receives the AD signal from the DA converter 7 via the speaker 9b and the microphone 1b.
The transfer function up to the converter 5b is obtained in advance as a coefficient, and the signal from the AD converter 5a is processed using this coefficient and input to the LMS calculator 11. LM
The S operator 11 performs an LMS operation (least square method) using the output from the Fx filter 10 and the output from the digital IIR filter 12c so as to minimize the output signal from the digital IIR filter 12c, and performs an adaptive filter. Update coefficient of 6. As a result, noise is attenuated by the control sound from the speaker 9b in the microphone 1b.

Here, the digital IIR filter 12b,
Reference numeral 12c removes the high frequency component, suppresses the growth of the high frequency component of the adaptive filter coefficient other than the noise control band, and stably grows the adaptive filter coefficient. Further, by making the digital IIR filters 12b and 12c have the same characteristics, the Fx is identified by
b, which can be identified without affecting the transfer characteristics to the AD converter 5b via the microphone 1b.

FIG. 11 is a development example of the above embodiment.
The digital IIR filter 12b and the Fx filter 10 are interchanged in configuration order, and have the same purpose and effect.

[0050]

As described above, according to the present invention, since the sound pressure frequency characteristic sufficiently reduces the sound pressure in a band higher than a half frequency (Nyquist frequency) of the sampling frequency, especially the error detector By using a speaker whose sound pressure level in the band above the Nyquist frequency is equal to or lower than the sound pressure level in the frequency band where noise is reduced at the position, a low-pass filter that limits the band of the output of the DA converter is omitted. Thus, the delay time of the entire silencer can be reduced.

[Brief description of the drawings]

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

FIG. 2 is a sound pressure frequency characteristic diagram of the speaker.

FIG. 3 is a sound pressure frequency characteristic diagram of the microphone 1a or the microphone 1b, which is the main part, in a case where the sound pressure level in a band higher than the Nyquist frequency is low.

FIG. 4 is a sound pressure frequency characteristic diagram in the case of noise having a high sound pressure level in a band higher than the Nyquist frequency in the microphone 1a or the microphone 1b.

FIG. 5 is an amplitude-frequency characteristic diagram of a low-pass filter which is the main part

FIG. 6 is a timing chart of the AD conversion and the DA conversion.

FIG. 7 is a timing chart of AD conversion and DA conversion when the noise control band is higher than the Nyquist frequency and the level is low.

FIG. 8 is a block diagram showing another embodiment.

FIG. 9 is a block diagram of the development example.

FIG. 10 is a block diagram showing another embodiment.

FIG. 11 is a block diagram of the development example.

FIG. 12 is a block diagram of a conventional silencer.

FIG. 13 is a diagram showing the processing delay time;

FIG. 14 is a diagram showing a sound pressure frequency characteristic of a speaker as the main part.

[Explanation of symbols]

 1a, 1b Microphone 2a, 2b Microphone amplifier 3a, 3b, 3c Low pass filter 4a, 4b Sample hold circuit 5a, 5b AD converter 6 Adaptive filter 7 DA converter 8 Power amplifier 9, 9b Speaker 10 Fx filter 11 LMS calculator 12a, 12b , 12c Digital IIR filter

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Kazuhiro Yamamoto 1006 Kazuma Kadoma, Osaka Prefecture Matsushita Electric Industrial Co., Ltd.

Claims (8)

[Claims] 1. A noise detector for detecting noise from a noise source, a first low-pass filter for removing a high-frequency component of the noise from the noise detector, and an analog output of the first low-pass filter. A first A / D conversion circuit for performing quantization on the digital signal, an adaptive filter for adaptively controlling the digital signal from the first A / D conversion circuit, and a digitally converted signal from the first A / D conversion circuit. A transfer function corrector that processes a noise signal, a DA converter that converts the output of the adaptive filter to an analog signal, a speaker that reproduces the output of the DA converter, control sound from the speaker, and noise from a noise source. And an error detector installed at a control position for reducing noise due to interference, and removing a high-frequency component of an error signal detected by the error detector. A second low-pass filter, and a second AD converter that samples an analog output of the second low-pass filter and quantizes the digital signal,
A noise reduction device comprising: a coefficient calculator that calculates and updates a coefficient of the adaptive filter from an output of the transfer function corrector and an output of the second AD conversion circuit;
The transfer function corrector previously identifies the transfer function from the speaker to the error detector as a coefficient, and the sound pressure frequency characteristic of the speaker is sufficiently high in a band higher than half the sampling frequency (Nyquist frequency). Silencer with reduced sound pressure. 2. The muffler according to claim 1, wherein a digital IIR filter for performing low-frequency transfer function correction is added to an input side or an output side of the transfer function corrector. 3. A first digital IIR filter for removing a high-frequency component on an input side or an output side of a transfer function corrector and preventing a coefficient calculator from updating a high-frequency adaptive filter coefficient outside a desired noise control frequency band. And a second digital IIR filter that removes high-frequency components from the output of the second AD converter and prevents the coefficient calculator from updating the high-frequency adaptive filter coefficients outside the desired noise control frequency band. 2. The muffler according to claim 1, wherein the first digital IIR filter and the second digital IIR filter have substantially the same characteristics. 4. The sound pressure frequency characteristic of the loudspeaker is such that the sound pressure level in a band above the Nyquist frequency is equal to or lower than the sound pressure level in a frequency band in which noise is reduced at the position of the error detector. The muffler according to any one of claims 1 to 3. 5. The first and second AD converters each convert a conversion from an analog signal to a digital signal and input the converted signal to the adaptive filter and the transfer function corrector in a sampling period (= reciprocal of the sampling frequency). The noise reduction device according to any one of claims 1 to 3, which is not more than half of (1). 6. The muffler according to claim 1, wherein each of the first and second AD converters is a successive approximation type. 7. The DA converter has a sampling period (= reciprocal of the sampling frequency) for converting the digital signal from the adaptive filter into an analog signal.
The noise reduction device according to any one of claims 1 to 3, which is not more than half. 8. The muffler according to claim 7, wherein the DA converter is of a ladder type.