JP3544999B2 - Silencer - Google Patents

Description

[0001]
[Industrial applications]
The present invention relates to a noise reduction device that cancels noise that enters a room through a duct such as an air conditioner with an additional sound having a phase opposite to that of the noise.
[0002]
[Prior art]
In general, as one of noise reduction methods, there is a passive noise control method in which noise is controlled using a sound absorbing material or a vibration isolating material. In this method, noise is controlled by installing a sound absorbing material or vibration damping material in the drive source, its surroundings, or the noise propagation path.It is necessary to change the quality, size and shape of the material depending on the frequency to be controlled. There is a disadvantage that the system becomes large to control the frequency noise.
[0003]
On the other hand, in recent years, an active noise control system using digital signal processing has been proposed as a noise control method, and some of them have been commercialized. FIG. 5 is a block diagram showing an example of this noise control system, wherein 101 is a noise source (for example, a ventilation fan), 102 is a noise detection microphone, 103 is a muffling speaker, 104 is an error detection microphone, and 105 is the noise detection A variable coefficient FIR (Fast Impulse Response) filter 106 for generating a reverse-phase additional sound for canceling this noise from a detection signal of the microphone for use 102, and a variable coefficient calculation unit 106 for calculating a coefficient of the filter 105.
[0004]
In such a conventional control method, a noise signal generated by the noise source 101 is detected by the noise detection microphone 102, and based on this signal, at the control point position where the detection microphone 104 is installed, the same amplitude and When generating a signal of the opposite phase, the noise signal is filtered by the variable coefficient filter 105, and the speaker 103 is driven by the processed signal, and the speaker 103 additionally generates an erasing sound. The generated cancellation sound and noise interfere with each other at the control point, and as a result, noise is reduced. The interference sound is further detected by the error detection microphone 104, and the detected signal is used as an error signal to calculate a new filter coefficient by the coefficient updating algorithm in the variable coefficient calculation unit 106 so that the signal is minimized.
[0005]
Generally, Filtered-XLMS is often used as the coefficient updating algorithm here. This is to update the filter coefficient every moment based on the following equation (3).
[0006]
[Equation 3]
[0007]
[Problems to be solved by the invention]
In the above-described conventional configuration, if it is desired to control the noise at a frequency over a wide band, or if it is desired to control the noise with high accuracy, the order of the variable coefficient filter is increased, or the frequency range of the noise to be controlled to some extent is limited. There is a disadvantage that it is necessary to perform
[0008]
In addition, when processing is performed by dividing the signal processing unit into a low-band part and a high-band part and performing processing, a plurality of adaptive processing units are required, and accordingly, the processing amount increases and the scale of hardware increases. There was a problem.
[0009]
SUMMARY OF THE INVENTION The present invention has been made in view of the problems of the related art, and has as its object to provide a noise reduction device that can accurately control broadband noise with a small amount of arithmetic processing.
[0010]
[Means for Solving the Problems]
The present invention provides a noise detection microphone, an error detection microphone, and a detection signal from the error detection microphone multiplied by a weight coefficient calculated based on the coherence between the detection signal and the detection signal from the noise detection microphone. Weighting filter for adjusting a detection signal from the error detection microphone, a variable coefficient filter for generating a signal for canceling noise based on the detection signal from the noise detection microphone, a detection signal from the noise detection microphone, and the weighting filter a variable coefficient calculating unit for calculating the coefficients of the variable coefficient filter by the output from, the speaker issuing sound to cancel the noise on the basis of the output signal from the variable coefficient filter consists, wherein the weighting filter, the speaker stops From the measurement results at the time It is composed of a filter having the same characteristics as the control effect R (omega) which satisfies the following equation 5 obtained.
(Equation 4)
(Equation 5)
[0011]
[Action]
It is generally known that the effect in active noise control is closely related to the coherence between the noise signal and the noise signal at the control point. Here, the coherence is a quantity indicating the strength of the relevance of the frequency components included in the two signals, and is obtained from the above equation (4).
[0013]
According to Equation 4, if the coherence between the noise detection signal and the noise signal at the control point is large, it can be said that the relevance of the frequency components of both signals is strong and the noise control effect is large. Therefore, an ideal control effect can be derived from the above equation (5) using this coherence.
[0015]
The configuration of the present invention is intended to increase the control effect based on this theory. The coherence is measured in advance, and the ideal control effect R (ω) is calculated from the measurement result. Then, a filter having the same characteristic as that of R (ω) is designed, and the filter is weighted as a weighting filter provided near the error detection microphone provided at the control point. In this way, operations such as reducing the weight are performed on the signals in the frequency band where the coherence is small and the control effect is small, and by performing the weighting, the signals unnecessary for the control are reduced and the active control is efficiently performed, and the noise in the wide band is reduced. Can be accurately controlled.
[0016]
【Example】
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a silencer according to the present invention will be described in detail with reference to the drawings based on one embodiment.
[0017]
FIG. 1 shows a system configuration diagram of a silencer, wherein 1 is a duct for an air conditioner, 2 is a fan as a noise generating source, 3 is a noise detection microphone provided near the fan 2 to detect noise 21, Reference numeral 4 denotes a speaker which is provided near a control point in the duct 1 and emits an additional sound 41 for canceling noise. Reference numeral 5 denotes an error detection microphone which detects a sound after the additional sound 41 is superimposed on the noise 21. Reference numeral 6 denotes a microphone. A variable coefficient filter that generates a drive signal for driving the speaker 4 by performing a filter operation on the detection signal of the noise detection microphone, 7 is a weighting filter that weights a signal obtained by the error detection microphone 5, and 8 is a weighting filter. The coefficient of the variable coefficient filter 6 is adjusted based on the detection signal from the noise detection microphone 3 and the output from the weighting filter. It is a variation coefficient calculation unit.
[0018]
In such a configuration, noise generated by the rotation of the fan 3 is transmitted through the duct 1 to the opening 11 of the duct 1.
[0019]
First, the coefficients of the weighting filter 7 are calculated. The coherence is measured by using the detection signal of the noise detection microphone 3 near the fan 3 and the detection signal of the error detection microphone 5 near the opening 11 of the duct 1. Then, the ideal silencing volume is calculated by the above equation (5).
[0020]
FIG. 2 is a diagram showing a change in the ideal silencing volume calculated from the coherence measured in the present embodiment depending on the frequency. From this figure, it can be seen that in the present embodiment, according to the conventional method, the value of the ideal silencing volume is small in the low and high frequency ranges, and the control effect by silencing is weak.
[0021]
FIG. 3 is a characteristic diagram showing an impulse response of the weighting filter 7 in which data correction by an appropriate window function is performed using the ideal silencing data shown in FIG. 2 and a weighting coefficient is set.
[0022]
The actual active control operation was performed after setting the weighting coefficient of the weighting filter 7 in advance as described above. That is, the noise generated by the fan 2 is detected by the noise detection microphone 3, the detection signal is input to the variable coefficient filter 6 for processing, and the speaker 4 is driven by the control signal from the filter 6.
[0023]
On the other hand, the error signal detected by the error detecting macrophone 5 is weighted by the weighting filter 7 having a filter coefficient derived from the coherence obtained in advance by measurement as described above, and the weighted error signal and the noise signal Thus, the coefficient of the variable coefficient filter 6 is updated based on an LMS (Least Mean Square) algorithm or the like, which is one of adaptive control algorithms, so that the error signal in the variable coefficient calculation unit 8 is minimized.
[0024]
As described above, at a certain point of the error detection microphone 5, the noise signal and the control signal having the same amplitude and opposite phase interfere with each other, and as a result, noise is reduced. FIG. 5 is a gain-frequency characteristic diagram for explaining the effect of silencing according to the present embodiment. In the figure, a solid line indicates raw noise without performing silencing control at all, and a dotted line indicates a case where silencing control according to the above embodiment is performed. The broken lines indicate the cases where the silencing control is performed by the conventional method. As is clear from this figure, in the present embodiment, it is understood that good characteristics due to noise reduction are obtained uniformly from the low frequency band to the high frequency band.
[0025]
【The invention's effect】
As described above, according to the present invention, by weighting the error signal based on the coherence obtained in advance, the weight of the signal in the frequency band having a large control effect is increased, and the coherence is reduced. It is possible to reduce the weight of the signal in the frequency band where the effect is small, and it can be expected that the noise can be efficiently silenced in a wide range from the low frequency range to the high frequency range.
[Brief description of the drawings]
FIG. 1 is a basic block diagram illustrating a configuration of an embodiment of a muffler according to the present invention.
FIG. 2 is a frequency characteristic diagram of an ideal control effect.
FIG. 3 is a diagram illustrating an impulse response of a weighting filter.
FIG. 4 is a characteristic diagram for comparing the silencing effects of the conventional and the silencing devices of the present invention.
FIG. 5 is a basic block diagram of a conventional silencer.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Duct 2 Fan 3 Noise detection microphone 4 Noise reduction speaker 5 Error detection microphone 6 Variable coefficient filter 7 Weighting filter 8 Variable coefficient calculation unit

Claims (1)

A noise detection microphone, an error detection microphone, and a signal detected by the error detection microphone multiplied by a weight coefficient calculated based on the coherence between the detection signal and the detection signal from the noise detection microphone. A weighting filter that adjusts a detection signal from the microphone, a variable coefficient filter that creates a signal that cancels noise based on the detection signal from the noise detection microphone, a detection signal from the noise detection microphone, and an output from the weighting filter. A variable coefficient calculation unit that calculates the coefficient of the variable coefficient filter, and a speaker that emits a sound that cancels noise based on an output signal from the variable coefficient filter,
The muffling device according to claim 1, wherein the weighting filter is a filter having the same characteristic as a control effect R (ω) satisfying the following relational expression obtained in advance from the measurement result when the speaker is stopped .