JPH07500677A - noise reduction system - Google Patents

Abstract

An active control system for attenuating tonal noise in a defined region is described. In its most basic form the system includes sensors (1, 8) for generating signals indicative of the residual noise in the region after attenuation and the uncontrolled sound affecting the region, signal processing circuits (10, 26) for processing the generated signals differently depending on the tonal content thereof, an adaptive filter (5) supplied with at least one of the generated signals whose characteristic is controlled by the processing circuitry (10), a transducer (6) for producing tonal-noise-attenuating disturbance in the region and delay means (4) for delaying signals relating to the uncontrolled noise before or after or during the adaptive filtering. The system finds direct application in a personal headset or ear defender.

Description

[Detailed description of the invention] This invention proposes the selection of signal noise or attenuated noise over broadband (random) noise. Concerning selective attenuation in general. That is, the signal tones are each at their fundamental frequency and perhaps is produced by one or more noise sources that generate noise at harmonic frequencies. It's okay.

Signal tones are more disturbing and generally more random in nature (such as (e.g. audio signals) reaches the point where it reduces the human ability to listen to pond sounds. This invention An active noise control system that provides attenuation of signal tones more than attenuation of random tones in the region of system and require a signal link to one or more sources of signal tone noise. Not required. In this sense, signal tone noise includes narrowband random noise, And noise includes vibration.

Discussion of selected prior art Hetsutoseno 1- selectively eliminates noise produced by a single rotating machine is known (Chaplin (Chaph mouth) - GB2104754), but these Repetitive noise from cables or ultrasonic/infrared transmission to generate signals Requires a link to the source. For many machines, this system Requires one link per page.

This link or links can be inconvenient - if each cable restrict the wearer's movements if they are transmission-based; It can become unreliable when moving into areas where user or transmission is uncertain.

GB2104754 manually generates a signal, drives a phase-locked loop, and generates a trigger signal. Explain the "female" nature of using a local microphone to It is acknowledged that repetitive noise may only be effective at particularly regular intervals. explanation His method cannot handle multiple sources of signal tones and noise, as Will.

Therefore, from the source of the signal g7 noise, which is currently operable against multiple sources, signal in an area without requiring one heliger of the signal to be derived. Selective attenuation of sound noise appears to be unknown.

Wj cheap of invention In accordance with one aspect of the invention, a ring l placed in a region is less likely to cause noise than noise. signal? ’) An active control system arm is provided to attenuate the noise, and this system Temu does not require a link to the number of characters of Shinyumi & No.Is or to the source of NVi - Generate gates that interfere with the signal 7,000 noise in the area, and reduce the cancellation of the signal tone noise. 7, thereby making it more reliable than random noise in that region. 1 means for attenuating the signal noise, and - the remainder of the area. The monitor signal related to the remaining (quiet) noise can be determined by the number of or with multiple sensors - for selective attenuation of uncontrolled sound in areas where it is needed; first circuit means for effectively delaying, if not partially, associated signals; , - an effectively delayed signal (or multiple beliefs derived therefrom, i.e. it The signal itself is derived from the (which itself may be modified) and monitor signals (or which may be a plurality of signals, -) and from which the signals themselves are derived, and the filter is a second circuit for processing (which may itself be modified by filtering) The result of its processing for noise is generally , unlike those for noise, − j − withered effectively delayed The adaptive filter which has a signal and an adaptive attenuation characteristic of 4T is controlled by a second circuit means. outputting a signal 71 for driving the transducer means; The filter operates in such a way that it selectively attenuates signal sounds over random sounds. Lt, : To be done.

The processing in the second circuit means effectively delays the borrowing (or then the bow one output). a plurality of signals derived from it, that is, the signals themselves or derived therefrom; (which may itself be modified by filtering) and monitor beliefs (which may themselves be modified by filtering). or multiple signals derived from it, i.e. the signal itself or derived from it. and may itself be changed by filtering. may also include cross-correlation with

Processing in the second circuit means effectively derives the delayed signal (or a plurality of signals, each of which is derived from the signal itself, and (which itself may be modified by filtering) and monitor signals; (or multiple signals derived from it, i.e. from which said signal itself or and which may themselves be modified by filtering. ).

Selective attenuation is at least partially related to uncontrolled sound in the required area. The signal to be first placed near the area where selective attenuation is required! Nsa? It may be obtained from .

Alternatively, selective attenuation or at least a portion of the uncontrolled signal in the required area The signal associated with the transducer means (or A plurality of signals derived from the (which may itself be modified by filtering) You can also reduce manpower first by subtracting from the number.

The first circuit means either by actually delaying the received signal. Tezoreka reception It is possible to effectively delay the signal being transmitted. Advantageously, this delay is The first suitable tL is itself responsive to the spectrum or autocorrelation of human power to the road means: regulated by the unit means and thus provides a monitor signal from the transducer. The sum of this delay plus any delays in the transmission process to the sensor is attenuated. Correlation of poor noise that is larger than the correlation time of poor noise and is attenuated shorter than time. This is due to the short duration cross-correlation (i.e. broadband It is not possible to attenuate the noise that has a random signal), but it is difficult to attenuate the This means that it is possible to attenuate noise that has a cross-correlation of durations. .

In an alternative method of increasing the effective delay of the first circuit means by 14, the first circuit means contains a narrowband filter of 1- to 1- whose output is sent to an adaptive filter. But that's fine. These narrowband filters have a center frequency or the frequency of the signal tone being canceled. the bandwidth is adjusted to accommodate the number of signals and whose bandwidth includes these tones and noise. It can be fixed or tunable so that it can be adjusted accordingly. The narrowband filter is the first The first application is itself responsive to the spectrum or autocorrelation of the input to the circuit means. adjustment unit means.

The adaptive filter minimizes the cross-correlation between each of its inputs and the monitor signal. one or more sequences having a first property determined at least in part by It is possible to include a column adaptive filter section. The first circuit means or the narrowband filter Where filters are included, each filter output is a parallel adaptive filter of adaptive filters. one to each section. Each adaptive filter section has a coefficient or gradient The lowest impulse response of the order of 1 is adjusted by the descending algorithm l. It's okay.

Where a separate sensor is used to generate a signal for the first circuit means--human power. advantageously, the sensor is insensitive to the noise produced by the transducer. L: is or is the influence of 1 to 7 on the signal it produces Positioned to be reduced by E child subtraction.

This system is incorporated into a heno1~seso1~ or ear protection device111. one system is used, or alternatively two systems are used, resulting in The quieted area generally includes one or both ears.

When installed in a headset or ear protection device, the monitor signal The sensor is placed close to the outer ear to infer unstable pressure points. transducers, such as microphones placed near the outer ear. It may be an icrophone. A set for applying a signal by human power to the first circuit means. The microphone is placed on or near the headset. or alternatively near the underside of the outer skin of one or both earpieces. and in a further alternative the sensor is specific to generate this signal. is not required, the signal is derived from the monitor signal as described herein.

A transducer is any device capable of generating unstable LL force fluctuations, such as The target may be a loudspeaker. L. lanceunyusa around one or both ears May be worn close to the ear to affect unstable pressure in the area of For example, a transducer may be attached to the outer skin of one or both earpieces. stomach.

head set It may also be one with an opening at the back. Head Sera 1~ is the desired amount or expansion. It may also be part of a communication set played via a voice organ. Additionally, In general, the effectively delayed adaptive filters described here are Microphone channels are used to attenuate background signal noise picked up by may be incorporated into the audio receiving channel.

The invention will now be explained by way of example with reference to the accompanying drawings, in which: FIG.

drawing FIG. 1 is a block schematic diagram of an embodiment of the present invention.

FIG. 2 is a block schematic diagram of an alternative embodiment of the invention.

FIG. 3 shows an example of the invention in which the sensor signal is derived from the sensor in the controlled area. FIG. 2 is a block schematic diagram of an example.

FIG. 4 shows an alternative embodiment in which the sensor signal is derived from a sensor in the controlled area. It is a block schematic diagram.

Figure 5 shows example automatic correlations for random noise and for signal tone noise. FIG.

FIG. 6 is a diagram showing an embodiment of the present invention incorporated into the Yatsul-Shichisoto.

FIG. 7 is a diagram showing an embodiment of a pond of the present invention incorporated in a heno1-seso.

FIG. 8 shows a further embodiment of the invention incorporated into a headset.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS Figure 1 shows a sensor that produces a signal representing the sound in area 2! shows. This signal 3 is the first circuit means 4 and the resulting signal is traversed via an adaptive filter 5. is sent to user 6, which generates a sound to interfere with area 2's 7-( . The coefficients of the JI8i-adaptive filter 5 are determined according to the adaptive algorithm described below in JJ. The sensor 8 of the diaphragm area 2 is regulated by the second circuit means 26, which is controlled by the second circuit means 26. using the monitor signal 7 from and the input 9 to the adaptive filter 5. first circuit hand Stage 4 comprises first adaptive unit means according to the spectrum or autocorrelation of signal 3. 10. The delay is the correlation time of the crackle that is left unattenuated. Adjusted to be louder and smaller than the correlation time of the attenuated first sound. . This delay is either frequency sensitive or frequency dependent where required. It may be done as follows.

The adaptive algorithm of the second circuit means 26 determines the distance between the input of the filter and the monitor signal. Adjust the adaptive filter to minimize the correlation of any! l! i response al It is a algorithm. These are frequency domain including JJ calculation of H1 alternating spectrum. It may be of the type or of the time-domain type, including cross-correlation. stomach. Many algorithms of this type are & 5TEARNS) Norr adaptive signal processing (adapjive signal processing) J. One time-domain algorithm will be explained below.

The human signal 9 is denoted by u(t) and the output of the adaptive filter section 5 is denoted by Y(t). Normally the signal is the delay unit n5i or later The sampled signal is converted by an analog-to-digital converter into It is in digital format. The sampled versions of input and output are u ( k) and y (k), where k represents time. The signal is adaptive After filter 5 it will be converted back to analog form. The output y (k) is where b(i) represents the first coefficient of the filter.

The output of sensor 8, that is, the monitor signal is w (k), and this consists of two components v(k), including uncontrolled noise and components due to transuser 6. nothing.

Here, C represents the effect of transuser characteristics. Separate area noise controlled is minimized by adapting the coefficients b(i) using a gradient descent algorithm. (For its derivation, see Prentice Hall. Whitlow & Stearns's “Suit 1. Coni” published in 1985 by Please see Signal Processing J).

1) +m (i) = b + (i) -μr (k-i) w (k), And l)+41(1) is the filter coefficient S, the next number of (i)! It's new.

The formula r (ki) w (k) is the 11th phase Q4 correlation between the two signals l and W. 41. Interpreted as a -lead -1°le estimate.　j'、4ru.

)/S other J5 similar value or i + J function, and in this, I〈 is any number and i FJru 3゜ One particularly novel feature of this invention (1 circuit f-stage 11 wideband Shin-J? There is a +IL car C that reduces the λ)1' error. , Kameno )Explain the characteristics using ko and =. Dojo's fit toshifu (Ruta L, /7, -) 7shi is fi By adjusting the coefficient of ’l leap relation (positive time piece’ (this, +4-ru) is driven by 1, and its result is Elimination to eliminate random noise that contributes to the correlation between the two. Noise will be generated. By introducing a delay into the human power system, The correlation between the two for broadband noise with a correlation time shorter than the delay ( for a positive time shift) will no longer be any, and therefore the adaptive filter Tasection will do nothing. On the other hand, the correlation time is significantly longer than the delay. The system still generates a cross-correlation between the delayed human power and the monitor signal, and Therefore, adaptive filters tend to eliminate these correlations to eliminate them. Probably. Figure 58 shows that the sensors are close to each other and receive broadband noise. 2 shows a typical cross-correlation of the input to the first circuit means with the monitor signal when

Figure 5b shows typical / Shows S cross-correlation. The introduction of the delay shifts the cross-correlation, so that its origin is This is the point marked D. The adaptive filter has an embarrassing level of mutual correlation on the right side of the origin. Since there is no other way to control the signal that connects the 41- It will be controlled. If this system receives a combination of many lights, If the cross-correlation exceeds the intersection by 4+, then eliminate the A signal of no result? The brush will be divided by jJl:.

47 No. 3 is received by Sen → No. 1 from 1. It is hoped that it will not be contaminated. This is a transducer for sensor j. by making it directional so that it is insensitive to sound from the user, or by making it insensitive to sound from the user. It is sensitive. or by positioning it as an input transformer the output of the transducer for the monitor signal. Sensor signals in a manner similar to that described to eliminate the effect of extraneous noise is used to reduce the effect of the transducer from (: J additive filter) This can be achieved by getting a job.

Alternative implementation FIG. 2 shows a layerer 1- similar to FIG. Has a layout. The signal 3 from the sensor l is connected to one or more of the first circuit means 12. The output 16 is sent to a narrow band filter, and the output 16 is suitable for 1, 1≦parallel suitable L1 of filter 14. ．． : One is sent to each of the filter sections. All outputs of parallel sections are combined to form the drive to transducer 6. The adaptive filter 14 The application of the individual parallel adaptive filter sections is to This is accomplished by second circuit means 15 in response to each of the filter outputs 16. Second The circuit means uses two series of adaptive algorithms, in which algorithm 1, What are the coefficients of each parallel adaptive filter section on the corresponding input and monitor signals? )

The narrowband filter of the first circuit means 12 may be fixed or tunable. So If they are fixed, a sufficient number of closely spaced fields in frequency , so that the tonal signals at frequencies within the range of interest are One of them will pass. This practice allows the selection process to occur. The novel features of the embodiment will now be described.

This first signal has a bandwidth smaller than the bandwidth of the narrowband filter. When sent to circuit means, the first (and therefore phase shifted) There will be an output from one of the filters that is positive. Sensor and monitor signals Since the cross-correlation with the signal has a long correlation time, the delay introduced by the filter Regardless of the delay, this output is correlated with the monitor signal (for positive time shifts). ), and therefore the corresponding parallel adaptive filter section of the at* filter 14 The section is adapted to produce an output signal and attenuate the li-sound noise. Will.

This first signal has a bandwidth much larger than that of the narrowband filter. , the output of each of the filters is the bandwidth of the original signal. or the output is a delayed version of the input (the bandwidth of a narrowband filter ), and the short phase of the original wideband signal is delayed by a time corresponding to the inverse of Because of the connection time, this (effectively) delayed signal has little interaction with the monitor signal. uncorrelated and therefore parallel adaptive filter sections produce little output. and the noise will not be attenuated.

Narrowband filters reduce system complexity by reducing the number of narrowband filters. may be tunable to minimize the Attenuated signal tone or slight If there are only a few, use a few narrowband filters, one for each signal tone. That may be Zi Dingyi. This means that if the signal at the input to the first circuit means The spectrum of the signal or the number of tones and frequency of the signal are monitored and And what can be achieved if a narrowband filter is tuned to accommodate these frequencies. It is possible. A narrowband filter sets its center frequency to the frequency of the corresponding signal tone. Continuously adjusted by the 1st al, Schnilla l-13 to maintain close Will. Their bandwidth is greater than that or the signal tone being attenuated, but However, by passing a wideband signal too wide and with insufficient delay, Adjustments are made to ensure that there are no failures. This process is automatic It will be held in

FIG. 3 shows that the sensor signal is extracted from the sensor 8 as in the area 2 to be controlled. FIG. In this figure, the times identical to those shown in Figure 1 are shown. The output 17 from the path means II is sent to a filter 18. The characteristics of this filter are reliable. is adjusted to correspond to the transfer function between signal 17 and signal 7. This is the first suit It is the same as filter c(i) used in response unit 6. Output of filter 18 The force is subtracted from monitor signal 7 to give an equivalent signal 3. The characteristics of the filter 18 are It can be updated to maintain it as an accurate representation of the transfer function.

Figure 4 shows an equivalent circuit in which the human power and monitor signals to the first circuit means are narrow band The circuit means 11 using a filter is derived from the same sensor.

Figure 6 shows the system incorporated into the ear protection device. Two shells, one for each outer ear. There is a stem, it has a portapu box 20 including 2 batteries for power Built-in. Ear protection devices allow the desired sound to reach the ear unhindered. It may also be of the type with an opening at the back to make it more functional. Input to the first circuit means The sensor 1 that generates the noise is a microphone mounted on the outer skin of the earpiece 19. The sensor 8 that generates the monitor signal is a microphone built into the earpiece near the ear. It's Crophon. Transducer 6 is a loudspeaker built into the outer skin of the earpiece. It is.

FIG. 7 shows that one microphone 22 transfers power to the first circuit means to both circuits 11. Shows the system used to give.

FIG. 8 shows an embodiment of a separate sensor for providing input to the first circuit means. (and its input is derived from the monitor signal as described above).

Additional human power is provided by one of the loudspeakers or These signals are then converted into sounds for the wearer to listen to. will be done. These hemorrhoids will not be affected by circuit means +1.

The microphone may be worn if it forms part of a headset or communications headset. will be attached to the headset to receive the voice of the person. In this environment to eliminate background signal noise raised by the a-voice microphone. In order to effectively integrate delayed adaptive filter sections into the voice communication channel, It may be desirable to include

An effective delay is appropriate; either before, during or after filtering. or any combination thereof that can be achieved by introducing delays into the signal. It must be understood. In this way, in FIGS. 1 to 3, the device 4 Some or all of the delays introduced by the filter 5. can be incorporated into or by a separate device placed between the filter 5 and the transuser 6. can be achieved. The second circuit 76 must of course be adjusted accordingly.

Reference is made in the foregoing specification and in the following claims to the cross-correlation of two signals. , and this representation is defined as follows, i.e. two digitally sampled The cross-correlation (C) of the ringed signals j; r(i and w(i) (where the cross-correlation The deviation between the two signals to be determined is n) is given by, where N is may be any number, and n is the number of sampling points as shown above. , thereby shifting one signal from the other.

~) Do\ To 2 q〕 Submission of amendment (7 (translation) submission (Patent Law Article 184-8) August 1992) 9 []

Claims (1)

[Claims] 1. Signal tone content rather than random content of noise affecting areas (quiet areas) An activation control system for further attenuating - a sound that interferes with the signal tone noise and produces at least a partial cancellation of the signal tone noise; occurs in a quiet region, making it more believable than random noise in that region. transducer means for further attenuating the signal noise; - the remainder of the area ( at least one of the quiet regions providing a first signal associated with the noise (attenuated); a first sensor of - at least a portion of the noise that would otherwise affect the area without selective attenuation; - at least one second sensor providing a second signal that is related to the first and second signals; and a second signal (or a signal derived therefrom), the processing being a signal tone. Signal processing for noise and random noise are generally different. means and - a signal derived from at least a second signal is provided and by the processing means; having adaptive characteristics to be controlled and a signal for driving the transducer means; The adaptive filter that generates and further - Introducing an effective delay in the signal path between the second sensor and the transducer an activation control system comprising means for; 2. 2. A method according to claim 1, wherein the delay is applied to the signal to be fed to the adaptive filter. Activation control system. 3. The activation control system of claim 1, wherein the delay is incorporated into an adaptive filter. 4. The delay is applied to the signal to be applied to drive the transducer means. The activation control system according to claim 1. 5. The processing means apply an adaptive filter to the first signal or to a signal derived therefrom. (or a signal derived from it, i.e., from which it is fed to an adaptive filter. The obtained signal is extracted and itself modified by filtering. Any one of claims 1 to 4, which performs the function of creating a cross-correlation with The activity control system described in Crab. 6. The processing means applies an adaptive filter to the first signal or a signal derived therefrom. (or a signal derived from it, i.e., from which it is fed to an adaptive filter. The obtained signal is derived and itself modified by filtering. Any one of claims 1 to 4, which performs the function of creating a cross-correlation with The activation control system according to any one of the above. 7. 6. The method according to any of the preceding claims, wherein the second sensor is in the quiet area. activation control system. 8. 7. According to any of claims 1 to 6, the first and second sensors are the same. activation control system. 9. Effective delay plus transducer to first sensor The acoustic delay to the sensor is greater than the correlation time of the noise that should not be attenuated. The activation control system according to any one of the claims. 10. Effective delay cancels signals at frequencies other than the tone frequency to be attenuated The embodiment of the preceding claim is realized by the use of a narrowband filter adapted for The activation control system according to any one of the above. 11.Activity control according to claim 10, wherein a set of parallel narrowband filters is used. system. 12. The adaptive filter combines the signal supplied to the filter with the signal from the first sensor. as claimed in any of the preceding claims, arranged so as to minimize the cross-correlation between activation control system. 13. The second sensor is placed in an area unaffected by the sound from the transducer. An activation control system according to any of the preceding claims, wherein: 14. The signal applied to the adaptive filter is the second signal (or signal fed to (or derived from) a transducer i.e. from which it is derived) or the first signal (or from which it is derived) insensitivity to the transducer by using a combination of An activation control system according to any of the preceding claims, adapted to the above claims. 15. Fitted into and part of a headset, headphones or ear protection device An activation control system according to any of the preceding claims, when forming a part. 16. 16. A separate control system as claimed in claim 15, wherein a separate control system is provided for each of the ear covering devices. Combinations listed. 17. The sensor is located on the headband of the headset or in the transducer housing. 17. The microphone according to claim 15 or 16, wherein the microphone is positioned outside the microphone. A combination of 18. Claims 15, 16 or 16, wherein for each ear the first and second sensors are the same. or the combination described in 17. 19. The transducer is placed close to the wearer's ear when in use. 19. The combination according to any one of claims 15 to 18. 20. Each transducer is a closed cavity formed by the outer skin that surrounds the ear. 19. A combination according to any one of claims 15 to 18, wherein the combination is mounted on. 21. The material properties of the skin are selected to enhance sound transmission in certain frequency ranges. 21. The combination according to claim 20. 22. Cross-correlations can be performed across a single sample or across many samples. The activation control according to any one of claims 5 to 21, which is calculated directly or recursively. control system. 23. As described herein with reference to and illustrated in the accompanying drawings Adapted activation control that is controlled and arranged to operate substantially system. 24. As set forth herein with reference to and illustrated in the accompanying drawings. Ear protection gear adapted to operate in a controlled manner to substantially value.