JP2872545B2 - 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 silencer using active noise control.

[0002]

2. Description of the Related Art In recent years, as can be seen in JP-A-4-282694, JP-A-4-282695, and JP-A-4-282696, noise of an air conditioner having a blower is controlled from a speaker using digital signal processing technology. There has been proposed an active noise control method that outputs sound and silences inside a ventilation duct.

Hereinafter, a conventional silencer will be described with reference to the drawings. FIG. 14 shows a block diagram of a conventional silencer. In FIG. 14, 1a and 1b
Are microphones which are the first and second noise detectors, 5 is a speaker, 9 is a ventilation duct for guiding air from the air conditioner body to each room, 12 is a control device, 13 is an averaging means, and 14 is a comparison unit. Means 15 are stopping means for stopping the operation of the control device 12.

[0004] The operation of the thus configured silencer will be described below. First, noise propagates in the air duct 9 in the direction of the arrow, and is detected by the microphone 1a. The detection signal is input to the control device 12.
At the same time, a noise signal detected by the microphone 1b is also input to the control device 12. Then, the control device 12 reproduces a control sound having the same amplitude and opposite phase as the noise in the microphone 1b from the speaker 5 so that the noise in the microphone 1b is attenuated from the two signals. As a result, noise at the microphone 1b is attenuated. However, since the control device 12 changes its own characteristics while constantly monitoring the signal from the microphone 1b, the transfer function between them may be changed by some factor, for example, aging of the microphone 1a and the microphone 1b. When the correlation between the signals detected by the microphone 1a and the microphone 1b is deteriorated due to the occurrence of a non-linear element, the characteristics may be erroneous, and noise may be added by the microphone 1b. In order to prevent this, a signal with time variation detected by the microphone 1b is time-averaged by the averaging means 13 and its output level is compared with a specified value by the comparing means 14. The operation of the device 12 is stopped.

FIG. 15 is a block diagram showing another conventional silencer. In FIG. 15, reference numerals 1a and 1b denote microphones which are first and second noise detectors, 5 a speaker, 9 a ventilation duct for guiding air from the air conditioner body to each room, 12 a control device, and 13 a Averaging means, 14 is a comparing means, and 15 is a stopping means for stopping the operation of the control device 12.

The operation of the silencer constructed as described above will be described below. The silencing operation is the same as in FIG. FIG. 15 shows the control device 12.
The operation of the control unit 12 itself is controlled by the output signal of the control unit 12. The output signal of the control device 12 with time fluctuation is averaged over time by the averaging unit 13, and the output level is compared with the specified value by the comparison unit 14. If the specified value, stop means 15
The operation of the control device 12 is stopped.

Further, although not shown in the block diagram, a configuration obtained by adding FIGS. 14 and 15 can be considered, but the operation is the same.

[0008]

However, FIG.
As shown in FIG. 15, the microphone 1b or the control device 1
In the conventional configuration in which the time average of the output of No. 2 is taken, noise is added in a certain frequency band but is attenuated in other main control bands. There is a problem that the controller 12 is determined to be operating normally and the control device 12 is operated as it is.

Further, an error such as a delay in the control stop operation occurs due to the accuracy of the averaging means 13, and the stopping means 1
5 also requires the averaging means 13 and the comparing means 14 to operate.

The present invention solves the above-described problem. Even when only a certain frequency band is added with noise without using the averaging means 13, the control characteristics are updated before noise is added. An object of the present invention is to provide a silencer that can be stopped and can perform these stop controls without errors such as delays, and as a result, can safely perform noise control.

[0011]

According to a first aspect of the present invention, there is provided a muffler comprising: a first noise detector for detecting noise from a noise source; an adaptive filter for adaptively controlling a detected noise signal; A digital filter that processes the output of the first noise detector, a speaker that is installed at a position farther from the noise source than the installation position of the first noise detector, and reproduces the output of the adaptive filter; A second noise detector installed at a position distant from the noise source in the vicinity of, and the coefficient of the adaptive filter is calculated in the time domain based on the output of the second noise detector and the output of the digital filter. And a coefficient calculator for updating the coefficient when the magnitude of the first coefficient among the coefficients of the adaptive filter calculated by the coefficient calculator reaches a specified value. Characterized by being constituted by a stop circuit for stopping the coefficient update by number calculator.

According to a fourth aspect of the present invention, there is provided a muffler including a signal processing circuit for adjusting an output frequency characteristic between the adaptive filter and the speaker. According to a fifth aspect of the present invention, there is provided a muffler, wherein a signal processing circuit for adjusting an output frequency characteristic is provided between the second noise detector and the coefficient calculator of the first aspect.

[0013]

According to the structure of the first aspect, the stop circuit monitors only the first coefficient among the coefficients of the adaptive filter calculated by the coefficient calculator, and when the magnitude reaches a specified value, the coefficient calculation is performed. The update of the coefficient by the detector is stopped, and the adaptive filter processes the noise from the first noise detector based on the coefficient and outputs the signal to the speaker 5 to control the noise.

According to the configuration of the fourth aspect, the signal processing circuit provided between the adaptive filter and the speaker adjusts the frequency characteristics of the signal from the adaptive filter and outputs the signal to the speaker.

According to the fifth aspect of the present invention, the signal processing circuit provided between the second noise detector and the coefficient calculator is provided in the second noise detector.
The frequency characteristic of the signal from the noise detector is adjusted and output to the coefficient calculator.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a silencer according to a first embodiment of the present invention. In FIG. 1, reference numerals 1a and 1b denote microphones as first and second noise detectors, 2 an adaptive filter, 3 an FIR filter as a digital filter, and 4 an LMS operation as a coefficient calculator. Container, 5
Is a speaker, 6 is a microprocessor (hereinafter abbreviated as CPU) as a stop circuit, 7 is an air conditioner body, 8
Is a centrifugal blower in the air conditioner main body 7, and 9 is an air conditioner main body 7.
A ventilation duct 10 for guiding air from the air to each room is a digital signal processor (hereinafter referred to as DSP) composed of an adaptive filter 2, an FIR filter 3, and an LMS operation unit 4.
).

The operation of the silencer configured as described above will be described below. First, the centrifugal blower 8
Is emitted from the discharge port into the air duct 9 and detected by the microphone 1a. The detection signal is input to the adaptive filter 2 and the FIR filter 3. The noise signal processed by the adaptive filter 2 is reproduced from the speaker 5. The control sound reproduced by the speaker 5 and the noise radiated from the outlet of the centrifugal blower 8 and transmitted through the ventilation duct 9 are detected by the microphone 1 b and input to the LMS calculator 4. The LMS calculator 4 calculates and updates the coefficient of the adaptive filter 2 based on the detection signal and the output signal from the FIR filter 3.
Here, assuming that the total transfer function from the speaker 5 to the microphone 1b is C, a coefficient c (n) approximating the total transfer function C is set in advance in the FIR filter 3, and the LMS calculator 4 calculates 1, the coefficient W (n) of the adaptive filter 2 is updated.

[0018]

(Equation 1) As a result, the adaptive filter 2 is connected to the microphone 1
At b, it operates to attenuate noise. This way
Filtered-x algorithm (references; B. Widrowand S. St)
earns, "Adaptive Signal Processing" (Prentice-Hal
l, Englewood Cliffs, NJ, 1985)).

Here, the microphones 1a and 1b and the loudspeaker 5 are connected to the coefficients of the adaptive filter 2 (W in Equation 1).
(N)), the coefficient (range from one item of W (n) to one-fourth of the total number of terms) in the vicinity of the forefront is arranged so as to have a peak level. .

This is set based on the following specific description. From microphone 1a to microphone 1b
Assuming that the noise transmission time t1 is t1, the control sound transmission time from the speaker 5 to the microphone 1b is t2, and the signal processing time (delay time) of the device is t3, t3 ≦ t1−t2, and t3 is t1 As the value approaches −t2, the level peak of the coefficient of the adaptive filter 2 approaches one item of the coefficient (0th tap).

As described above, the noise is eliminated by continuously updating the coefficient of the adaptive filter 2, but the noise detected by the microphone 1a or the microphone 1b is now shown in FIG. And has a frequency characteristic as shown in FIG.
When the frequency characteristic of the transfer function up to b is as shown in FIG. 2B, the noise reduction effect in the microphone 1b is as shown in FIG. 2C. In FIG. 2C, FIG.
Although a large and uniform silencing is obtained in the flat characteristic portion of (b), a difference in silencing is observed in a low-frequency part where the reproduction sound pressure level of the speaker 5 is reduced. Has been added. Looking at this with the coefficient of the adaptive filter 2, the time characteristic is that the first coefficient (W ₀ (n) in Equation 1) has the maximum level as shown in FIG. As shown in (b), the low frequency component is larger than the high frequency component. This means that the transfer function from the speaker 5 to the microphone 1b is as shown in FIG.
Since the low-frequency level is smaller than the noise controlled as described above and the coefficient of the FIR filter 3 approximates this, the adaptive filter 2 attempts to compensate for this by increasing the low-frequency component of the coefficient. Will be. However, an unlimited level increase causes an overflow due to the finite word length of the DSP 10 or a nonlinear distortion due to a large input to the speaker 5. As a result, the noise addition shown in FIG. 2 (c) appears first, and when the noise is further worsened, the noise addition also occurs in other bands, and finally diverges to an oscillating state.

However, the CPU 6 in FIG. 1 compares the magnitude of the first coefficient (W ₀ (n) in Equation 1) among the coefficients of the adaptive filter 2 with the value at the time when noise addition starts to occur. As soon as the coefficient (W ₀ (n) in Equation 1) reaches its specified value, the adaptive filter 2
When the DSP 10 is controlled so as to stop the coefficient updating, the noise can be controlled without adding noise. Looking at this with the coefficient of the adaptive filter 2, the time characteristic shows that the first coefficient (W ₀ (n) in Equation 1) has the specified value A as shown in FIG. As shown in FIG. 4B, the low-frequency component is smaller than that in FIG. 3B. As a result, as shown in FIG. 4C, noise control can be performed without generating additional noise.

In this embodiment, the updating of the coefficients of the adaptive filter 2 is stopped by using the CPU 6, but the coefficient may be monitored in the DSP 10 and stopped by itself.

Next, a second embodiment of the present invention will be described with reference to the drawings. FIG. 5 is a block diagram showing a silencer according to a second embodiment of the present invention. FIG.
Wherein 1a and 1b are microphones which are first and second noise detectors, 2 is an adaptive filter, 3 is an FIR filter, 4 is an LMS calculator, 5 is a speaker, and 6 is a CPU which is a stop circuit. , 7 is an air conditioner main body, 8 is a centrifugal blower in the air conditioner main body 7, 9 is an air duct for guiding air from the air conditioner main body 7 to each room, 10
Is adaptive filter 2, FIR filter 3, and LMS
The DSP 11 composed of the arithmetic unit 4 is an equalizer which is a signal processing circuit. Here microphone 1
Assume that a and 1b and the speaker 5 are installed such that the level peak of the coefficient of the adaptive filter 2 is near the forefront of the coefficient.

The operation of the muffler configured as described above will be described below. First, the centrifugal blower 8
Is emitted from the discharge port into the air duct 9 and detected by the microphone 1a. The detection signal is input to the adaptive filter 2 and the FIR filter 3. The noise signal subjected to the signal processing by the adaptive filter 2 has its frequency characteristics adjusted by the equalizer 11 and is reproduced from the speaker 5. The control sound reproduced by the speaker 5 and the air duct 9 radiated from the outlet of the centrifugal blower 8
Transmitted from the microphone 1b is detected by the microphone 1b.
It is input to the MS calculator 4. The LMS calculator 4 calculates and updates the coefficient of the adaptive filter 2 based on the detection signal and the output signal from the FIR filter 3. Where F
The transfer function from the equalizer 11 to the microphone 1b is approximated to the IR filter 3 in advance as a coefficient. As a result, the adaptive filter 2 is connected to the microphone 1b.
Operates to attenuate noise.

In detail, assuming that the equalizer 11 does not exist as shown in FIG. 1, the noise detected by the microphone 1a or the microphone 1b has a frequency characteristic as shown in FIG. When the frequency characteristic of the transfer function up to the microphone 1b is as shown in FIG. 6B, the sound deadening effect in the microphone 1b is as shown in FIG. In FIG. 6 (c), a larger volume is obtained as the frequency becomes lower, but the volume is small or not muted in the high frequency range. Looking at this by the coefficient of the adaptive filter 2, it can be seen that the low-frequency component is larger than the high-frequency component as shown in FIG. 7, and that the low-frequency component is mainly controlled. However, as a practical problem, when it is desired to ensure that the high-frequency silencing is as high as the low-frequency, the equalizer 11 is used as shown in FIG. 5 and the frequency characteristic of the equalizer 11 becomes higher as the frequency becomes higher as shown in FIG. By doing so, the transfer function from the equalizer 11 to the microphone 1b becomes as shown in FIG.
As shown in FIG. 8C, the muffling effect of the microphone 1b is attenuated uniformly in each band as shown in FIG. Looking at this with the coefficient of adaptive filter 2,
As shown in FIG. 9, the control is uniformly performed in each band as compared with FIG.

However, it is sufficient that control can be performed without any problem as described above. However, by making the equalizer 11 have such a characteristic that the level becomes lower as the frequency becomes lower as shown in FIG.
As shown in (d), the same problem as in the case of FIG. 1 appears. The coefficients at this time are as shown in FIG.

However, as in the case of FIG. 1, the CPU 6 in FIG.
The magnitude of the first coefficient (W ₀ (n) in Equation 1) is compared with the value at the time when noise addition starts, and the first coefficient (W ₀ (n) in Equation 1) is compared.
_When the magnitude of ₀ (n)) reaches the specified value, the DSP 10 stops updating the coefficient of the adaptive filter 2 immediately.
, Noise control can be performed without adding noise. Looking at this with the coefficient of the adaptive filter 2, the time characteristic shows that the first coefficient (W ₀ (n) in Equation 1) has a specified value A as shown in FIG. As shown in FIG. 11B, the low-frequency component is smaller than that in FIG. As a result, as shown in FIG. 11C, noise control can be performed without generating noise.

In this embodiment, the updating of the coefficients of the adaptive filter 2 is stopped by using the CPU 6. However, the configuration may be such that the coefficients are monitored in the DSP 10 and stopped by itself. In addition, although the equalizer 11 is used to adjust the output of the adaptive filter 2, a device using an attenuation characteristic such as an active filter may be used.

Finally, a third embodiment of the present invention will be described with reference to the drawings. FIG. 12 is a block diagram showing a silencer according to a third embodiment of the present invention.
In FIG. 12, microphones 1a and 1b are first and second noise detectors, 2 is an adaptive filter, 3 is an FIR filter, 4 is an LMS calculator,
5 is a speaker, 6 is a CPU which is a stop circuit, 7
Is an air conditioner main body, 8 is a centrifugal blower in the air conditioner main body 7,
9 is a ventilation duct for guiding air from the air conditioner body 7 to each room, 10 is an adaptive filter 2 and an FIR filter 3
A DSP 11 composed of the LMS calculator 4 and an equalizer 11 is a signal processing circuit. Here, it is assumed that the microphones 1a and 1b and the speaker 5 are installed such that the level peak of the coefficient of the adaptive filter 2 is near the forefront of the coefficient.

The operation of the thus configured silencer will be described below. First, the centrifugal blower 8
Is emitted from the discharge port into the air duct 9 and detected by the microphone 1a. The detection signal is input to the adaptive filter 2 and the FIR filter 3. The noise signal processed by the adaptive filter 2 is reproduced from the speaker 5. The control sound reproduced by the speaker 5 and the noise radiated from the outlet of the centrifugal blower 8 and transmitted through the ventilation duct 9 are detected by the microphone 1 b, the frequency characteristics are adjusted by the equalizer 11, and input to the LMS calculator 4. Is done. The LMS calculator 4 calculates and updates the coefficients of the adaptive filter 2 based on the output signal from the equalizer 11 and the output signal from the FIR filter 3. The FIR filter 3 has a speaker 5 in advance.
The transfer function from to is equalized as a coefficient. As a result, the adaptive filter 2 operates so as to attenuate noise in the microphone 1b.

Looking at this in detail, the microphone 1a
Alternatively, the noise detected by the microphone 1b is
When having the frequency characteristic as shown in FIG. 13A, the equalizer 11 is used as shown in FIG.
If the level is increased in the higher frequency range as shown in (b), the noise detected by the microphone 1 b will be equalized by the equalizer 11.
13 (a) to FIG. 13 (c), the sound deadening effect of the microphone 1b is attenuated uniformly in each band as shown in FIG. 13 (d).

However, as long as the control can be performed without any problem, the equalizer 11 has such a characteristic that the level becomes lower as the frequency becomes lower as shown in FIG.
The same problem as in the case of comes out.

However, as in the case of FIG. 5, the CPU 6 in FIG. 12 determines the magnitude of the first coefficient (W ₀ (n) in the equation 1) among the coefficients of the adaptive filter 2 when the noise is added. Compared to the value, the DSP 1 stops updating the coefficient of the adaptive filter 2 as soon as the magnitude of the first coefficient (W ₀ (n) in Equation 1) reaches its specified value.
When 0 is controlled, noise control can be performed without adding noise.

In the present embodiment, the CPU 6 is used to stop the updating of the coefficients of the adaptive filter 2, but it is also possible to monitor the coefficients in the DSP 10 and stop it by itself. In addition, although the equalizer 11 is used to adjust the output of the adaptive filter 2, a device using an attenuation characteristic such as an active filter may be used.

[0036]

According to the configuration of the first aspect, the stop circuit includes:
The first coefficient among the coefficients of the adaptive filter calculated by the coefficient calculator is monitored, and when the magnitude reaches a specified value, the operation of stopping the coefficient update by the coefficient calculator is stopped. Even if the noise detected by the noise detector 2 has a flat frequency characteristic in any frequency band and only a certain frequency band has noise added, the coefficient arithmetic unit operates before the noise is added. Coefficient update can be stopped. Further, since the noise addition can be determined by checking the magnitude of only the first coefficient among the coefficients of the adaptive filter, the coefficient update by the coefficient calculator can be immediately stopped without causing an error such as a delay. As a result, safe noise control without divergence can be performed.

According to the fourth aspect of the present invention, since the signal processing circuit capable of adjusting the frequency characteristic of the signal is provided between the adaptive filter and the speaker, the frequency characteristic of the output of the adaptive filter can be adjusted. Can be converted into a signal of a uniform level in each frequency band and input to the coefficient calculator, and the same effect as the first aspect can be obtained.

According to the fifth aspect of the present invention, the signal processing circuit capable of adjusting the frequency characteristic of the signal is provided between the second noise detector and the coefficient calculator, so that the output of the second noise detector can be adjusted. Can be adjusted, so that even a noise whose frequency characteristic is not flat can be converted into a signal of a uniform level in each frequency band and input to the coefficient calculator, and the same effect as the first aspect can be obtained.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a silencer according to a first embodiment of the present invention.

FIG. 2 is a characteristic diagram illustrating noise addition in the embodiment.

FIG. 3 is a characteristic diagram of coefficients of an adaptive filter when noise is added in the embodiment.

FIG. 4 is a control characteristic diagram of the muffler of the embodiment.

FIG. 5 is a configuration diagram of a muffler according to a second embodiment of the present invention.

FIG. 6 is a characteristic diagram for explaining the embodiment when the equalizer is not used.

FIG. 7 is a characteristic diagram of coefficients of an adaptive filter when an equalizer is not used in the embodiment.

FIG. 8 is a characteristic diagram illustrating noise addition when an equalizer is used in the embodiment.

FIG. 9 is a characteristic diagram of coefficients of an adaptive filter when an equalizer is used in the embodiment.

FIG. 10 is a characteristic diagram of coefficients of an adaptive filter when noise is added in the embodiment.

FIG. 11 is a control characteristic diagram of the muffler of the embodiment.

FIG. 12 is a configuration diagram of a muffler according to a third embodiment of the present invention.

FIG. 13 is a characteristic diagram illustrating control of the muffler of the embodiment.

FIG. 14 is a configuration diagram of a conventional silencer.

FIG. 15 is a configuration diagram of another conventional silencer.

[Explanation of symbols]

 1a, 1b microphone 2 adaptive filter 3 FIR filter 4 LMS calculator 5 speaker 6 CPU (microprocessor) 10 DSP (digital signal processor) 11 equalizer

Continuation of the front page (72) Inventor Tadashi Tamura 1006 Kazuma Kadoma, Kadoma, Osaka Prefecture Inside Matsushita Electric Industrial Co., Ltd. (56) References JP-A-5-40486 (JP, A) JP-A-4-282696 (JP) (A) JP-A-3-266516 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) G10K 11/178 F24F 13/02

Claims (7)

(57) [Claims] 1. A first noise detector for detecting noise from a noise source, an adaptive filter for adaptively controlling a detected noise signal, and a digital filter for signal-processing an output of the first noise detector. A speaker installed at a position farther from the noise source than the installation position of the first noise detector and reproducing the output of the adaptive filter; and a second speaker installed at a position far from the noise source near the speaker. A noise detector, a coefficient calculator for calculating and updating a coefficient of the adaptive filter in a time domain based on an output of the second noise detector and an output of the digital filter, and a calculation performed by the coefficient calculator. A stop circuit that stops updating the coefficient by the coefficient calculator when the magnitude of the first coefficient among the coefficients of the adaptive filter reaches a specified value; A silencer composed of: 2. The apparatus according to claim 1, wherein the first noise detector, the second noise detector, and the loudspeaker are arranged such that the level of a coefficient near the forefront of the coefficients of the adaptive filter has a peak. Silencer. 3. The muffler according to claim 1, wherein the adaptive filter, the coefficient calculator and the stop circuit are constituted by a digital signal processor. 4. The muffler according to claim 1, further comprising a signal processing circuit for adjusting an output frequency characteristic between the adaptive filter and the speaker. 5. A signal processing circuit for adjusting an output frequency characteristic between a second noise detector and a coefficient calculator.
The sound deadening device according to claim 1. 6. The muffler according to claim 4, wherein the signal processing circuit is configured to attenuate low frequency components. 7. The muffler according to claim 4, wherein the signal processing circuit is configured to be adjusted to an arbitrary frequency characteristic.