JP3316259B2 - Active silencer - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to noise in a propagation path such as a duct and the like, in which a reverse sound having the same phase and the same amplitude acts on the noise to muffle the noise. The present invention relates to an active silencer for silencing.

[0002]

2. Description of the Related Art It has been known that a noise in a duct is subjected to a reverse sound of an active amplitude in a phase opposite to that of the noise, and the noise is canceled by mutual interference between the noise and the reverse sound. .

In particular, in recent years, with the progress of electronic devices and signal processing techniques, inversion sound that is faithful to noise is generated, and as inversion sound that responds quickly to unsteady high-band noise,
An active silencer that silences even fluctuating noise has been proposed.

[0004] For example, an electronic noise reduction system disclosed in Japanese Patent Application Laid-Open No. 61-296392 is one of such systems.

However, the above-described active silencer is applied to silence one-dimensional noise having a limited sound field, such as noise in a duct, and responds to changes over time such as temperature, air volume, and flow velocity. This is to mute the sound, and the inverted sound signal for muffling is generated almost faithfully with respect to the noise, but there is a limit to the amount of noise that can be reduced in reality, and sufficient muffling cannot be performed. There was a problem.

To solve this problem, Japanese Patent Laid-Open Publication No.
Japanese Patent Application Laid-Open No. 8 (1996) -1995 discloses a technique in which a transfer function from a speaker in a duct to a monitor microphone is equalized between a FIR (Finite Impulse Response) filter and a speaker, and an inverse filter for superimposing a compensation signal on the inverted sound signal is provided. Was proposed.

[0007]

In the prior art described above, the input signal for updating the coefficient of the FIR filter is not affected by the transfer characteristics between the monitor microphone and the speaker, and is over a wide frequency band. Although a large volume reduction can be obtained, if the transfer characteristics between the speaker and the monitor microphone change due to factors such as temperature changes, the system diverges, abnormal noise is generated, and as a countermeasure, the transfer characteristics are automatically re-measured. Is performed when there is nobody around or when there is no problem to make a sound to generate the artificially created sound from the speaker for measurement. There was a problem that it was not possible to maintain.

The present invention has been made in view of the above-mentioned problems of the prior art, and measures a transfer characteristic between a speaker and a monitor microphone without generating an abnormal sound from the speaker, so that an optimum state is always obtained. An object of the present invention is to provide an active silencer capable of maintaining the above characteristics and performing optimal control without diverging even if the characteristics of the space change.

[0009]

According to the present invention, there is provided a noise detection microphone installed in a duct for detecting noise, and a noise detection microphone having the same amplitude in opposite phase to the noise based on a noise signal detected by the noise detection microphone. A control sound generation unit that generates a control sound, a speaker that emits the control sound of the control sound generation unit into the duct, a silence detection microphone that detects an interference result between the noise and the control sound, and a normal silence During operation, a buffer that constantly stores a detection signal detected by the silence detection microphone in a fixed amount, and wherein the control sound generation unit is configured to connect the speaker and the silence detection microphone during normal silence operation. A fixed filter that has a filter coefficient using a transfer characteristic between the filter and the noise signal, and an output signal of the fixed filter and the silence detection microphone. A coefficient update unit that generates a filter coefficient from the output detection signal, and an adaptive FIR filter that generates a control sound having the same amplitude in the opposite phase to the noise based on the filter coefficient generated by the coefficient update unit, And supplying a detection signal stored in the buffer at the time of the mute operation to the speaker, the adaptive FIR filter, and the coefficient updating unit according to an input of an external control signal. The filter updating unit updates the filter coefficient of the adaptive FIR filter by the coefficient updating unit so that the difference between the detection signal detected by the silence detection microphone and the output of the adaptive FIR filter is minimized. The filter coefficient of the adaptive FIR filter is updated and set as the filter coefficient of the fixed filter.

[0010]

According to the above arrangement, it is possible to generate a sound having the same level and the same timbre as the actual noise, and to update the coefficient of the fixed filter used for the control based on this signal, thereby giving a feeling of abnormality or discomfort to a person. , The filter can be updated at any time, and optimal noise control is always possible.

[0011]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a perspective view of an active silencer according to the present invention.

FIG. 3 is a system block diagram for explaining a silencing operation of a basic active silencer. In the figure, the Noise <br/> sound generated from a noise source 1 such as a fan in the duct 9 is detected by the noise detection microphone 2, and the detected noisy
At the control point in the duct 9 based on the sound signal, a control signal having the opposite phase and the same amplitude as the noise is created by the control sound creating unit 4,
This control signal is emitted from the muffling speaker 5 to reduce noise due to sound interference.

[0013] In such a system, as an algorithm for creating a control signal Ri by Noise signals, although various methods have been proposed and used here to Filtered x LMS algorithm is most common and practical use. This algorithm uses an adaptive digital filter that follows even when the acoustic characteristics change due to a change in temperature or the like during the operation of the system. The principle of this algorithm will be described below.

The system is composed of a microphone 2 for detecting noise, a muffling speaker 5, a muffling detection microphone 3 as a monitor microphone installed at a control point, and a control sound creating section (controller) 4 for creating a control sound. ing.

In the control sound generator 4, the transfer characteristic between the muffling speaker 5 and the muffling detection microphone 3 is measured in advance by any method, and a fixed coefficient FIR filter 6 having a filter coefficient using this characteristic is provided. Coefficient variable digital filter 7 created from a noise signal that has detected a control sound, and said coefficient variable digital filter 7 based on a noise signal and silencing volume
And a coefficient updating unit 8 for calculating the coefficient of

The coefficient update is calculated as shown in Expression 1, and the coefficient of the coefficient variable digital filter 7 is updated so that the signal of the microphone 3 is minimized.

[0017]

(Equation 1)

In general, the coefficient of the coefficient variable digital filter 7 operates adaptively and sufficiently responds to changes in acoustic characteristics during operation, but the FIR filter 6 used for updating the coefficient of the filter 7 is used. Is fixed. Therefore, an error occurs in the coefficient H _'2 of the transfer characteristic H ₂ and FIR filter 6 between the temperature change and speaker actual silencing amount detecting microphone and mute the speaker 5 due to aging or the like of the microphone 3, a large the error The system diverges, and an abnormal sound is generated from the muffling speaker 5.

Normally, the coefficient (H ' ₂ ) of the FIR filter 6 is determined by using a system as shown in FIG. 4 (similar to Japanese Patent Application Laid-Open No. 3-40598 described later) at the time of designing the system. The operation principle is described below.

[0020] The microphone 2, 3, and the speaker 5, as shown in FIG. 3
To the same as those used for the control of noise indicated, it can radiate artificial sounds created by the noise generator 9 from the speaker 5, the consumption volume detection microphone 3 signal and a coefficient variable digital filter 7 by the difference between the output of updating the coefficients in the coefficient updating unit 8 so as to minimize the coefficient of resulting in variable coefficient digital filter (H _'2), the speaker 5
And the transfer characteristic (H ₂ ) between the microphone and the microphone 3.

However, this series of processes for measuring the spatial characteristics and obtaining the filter coefficients requires the emission of artificial sound from the noise generator 9, and in actual use, when there are no people around or noises are generated. Perform it when there is no problem,
It is practically impossible to do it often. Therefore, it is difficult to always maintain an optimum filter coefficient, and a spatial characteristic and a filter coefficient may greatly differ due to temperature or other factors, and the system may diverge.

Therefore, the present invention is configured as shown in FIGS. Hereinafter, these figures will be described.

The measuring Initial Configuration transfer characteristics between the silencing amount detection microphone 3 and mute speaker 5 at block shown in FIG. 4 when installed the air conditioner system, namely the filter coefficients H _'2. At this time, the coefficient H ′ ₂ becomes almost equal to H ₂ .

[Mute Operation] FIG. 1 shows a system block diagram of a mute operation. A signal input from the noise detection microphone 2 is processed by the coefficient variable digital filter 7, and a sound having the opposite phase and the same amplitude as the noise is output from the speaker 5 by the noise reduction detection microphone 3 installed at the control point.
Generate more.

Further, the input signal of the silence detection microphone 3 is constantly stored in the buffer 10 by a fixed amount.

[Update of Fixed (FIR) Filter Coefficients] As shown in FIG. 1, when a control signal 12 is externally input to the control unit 11 for monitoring the entire system, the system changes to a fixed filter update process shown in FIG. .

The external control signal can be set to be generated when the air conditioner is started / stopped.

In the process of updating the coefficient of the fixed filter 6 shown in FIG. 2, the transfer characteristic between the speaker 5 and the silence detection microphone 3 provided at the control point based on the data stored in the buffer 10 during the silencing operation is described. The measurement is performed using the coefficient variable digital filter 7 as in the initial setting.

The signal output from the buffer 10 is
The signal is output through the filter 13 having the inverse characteristic of the coefficient H′2 used at the time of the above-described silencing operation. At this time, the inverse characteristic filter 13
Is the speaker 5 according to the coefficient updated by the control sound creation unit 4.
Transfer function H when the sound from
This is provided to prevent unnecessary noise from being increased by the output of the buffer 10 due to the influence of (2).

As a result, the loudness and tone of the sound emitted from the speaker 5 to the space in the coefficient updating process of the fixed filter are the same as those at the time of the muffling operation, so that the surrounding persons do not feel abnormal or uncomfortable.

When a series of fixed filter coefficient updating processes is completed in this way, the silencer performs the silencing operation shown in FIG. 1 again, and reduces noise using an optimal fixed (FIR) filter.

[0032]

As described above, the present invention always stores the input signal from the microphone for detecting the sound-extinguishing volume at the time of the sound-extinguishing operation. By filtering the transfer characteristic between the microphone and the transfer characteristic between the speaker and the silence detection microphone based on this signal,
It is possible to measure and update the characteristics that are the control parameters of the system without giving a sense of abnormality or discomfort to the people around them. By constantly updating the system parameters (coefficients) optimally, the It is possible to expect an effect that enables an optimal and stable silencing operation even for changes in characteristics due to changes in acoustic characteristics or changes over time in speakers or microphones.

[Brief description of the drawings]

FIG. 1 is a system block diagram at the time of a silencing operation of an active silencer of the present invention.

FIG. 2 is a system block diagram of the apparatus of FIG. 1 at the time of updating a fixed filter coefficient.

FIG. 3 is a system block diagram of a general active silencer corresponding to FIG. 1 during a silencing operation.

FIG. 4 is a system block diagram of the general active silencer corresponding to FIG. 2 at the time of updating a fixed filter coefficient.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 noise source 2 noise detection microphone 3 silence detection microphone 4 control sound creation unit 5 speaker 6 fixed (FIR) filter 7 coefficient variable digital filter 8 coefficient update unit 9 duct 10 buffer 13 inverse characteristic filter

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ identification code FI H03H 21/00 G10K 11/16 B (58) Fields investigated (Int.Cl. ⁷ , DB name) G10K 11/178 F01N 1 / 00 F24F 13/02 G10K 11/16 H03H 17/02 601 H03H 21/00

Claims (1)

(57) [Claims] 1. A noise detection microphone installed in a duct for detecting noise, and a control sound generation for generating a control sound having the same amplitude in the opposite phase to the noise based on the noise signal detected by the noise detection microphone. Unit, a speaker that emits the control sound of the control sound generation unit into the duct, a silence detection microphone that detects a result of interference between the noise and the control sound, and a silence detection during a normal silencing operation. A buffer for constantly storing a detection signal detected by a microphone in a fixed amount, wherein the control sound generation unit uses a transfer characteristic between the speaker and the sound volume detection microphone during a normal sound reduction operation. A fixed filter having a coefficient and allowing the noise signal to pass therethrough; and a filter based on an output signal of the fixed filter and a detection signal detected by the silence detection microphone. An active FIR filter having a coefficient updating unit for generating a filter coefficient, and an adaptive FIR filter for generating a control sound having the same amplitude in the opposite phase to the noise based on the filter coefficient generated by the coefficient updating unit. A silencing device for supplying a detection signal stored in the buffer during the silencing operation to the speaker, the adaptive FIR filter, and the coefficient updating unit in accordance with an input of an external control signal; as the difference between the output of the detection signal detected by the microphone and the adaptive FIR filter is minimized, the adaptive FIR Fi by the coefficient updating unit
Filter coefficients are updated, and the updated adaptive
The filter coefficient of the fixed FIR filter is
An active silencer, which is updated and set as a filter coefficient of data.