DE69125601T2 - NOISE REDUCING 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

The invention relates to selectively attenuating noise, whereby tonal sounds are attenuated more than broadband (random) noises. In particular, the tones may be generated by one or more noise sources, each of which produces noise at a fundamental frequency and possibly harmonic frequencies.

The tones reach a range where they constitute a disturbance and reduce a person's ability to hear other sounds that are more random in nature (e.g., speech signals). The invention relates to an active noise control system that provides greater attenuation of tonal noise than random sound in the range and does not require signal chaining to the source or sources of tonal noise. In this context, tonal noise includes narrow-band, random noise, and noise includes vibrations.

Discussion of the state of the art

Headphones or ear defenders which selectively cancel the noise generated by a single rotating machine are known (Chaplin - GB 21 04 754), but require a daisy chain or ultrasonic/infrared transmission to the source of the repetitive noise to generate a trigger signal. For many machines the system requires one link per machine. This link or links can be inconvenient or inconvenient - if they are in the form of cables they restrict the wearer's movement and if using wireless transmission they can be unreliable if the wearer moves into areas where the transmission is distorted. GB 21 04 754 describes the possibility of using a local microphone to obtain a signal which drives a phase-locked loop to generate a trigger signal, but states that this is only effective if the repetitive noise is particularly regular. As described, this method is not able to handle multiple sources of tonal noise.

From IBM Technical Disclosure Bulletin, Volume 31, No. 8 (January 1989), pages 256 - 258 there is also known a system for active noise cancellation which attenuates tonal noise more than random noise by using one or more sensors which generate signals which are processed to produce an output signal feeding a transducer. However, the system is also dependent on the presence of a reference signal or trigger signal which indicates the frequency of a single noise source. An adaptive delay device is used to achieve a delay which is set so that the output of the system is highly related to the tonal noise for a single tone only.

From "Adaptive Signal Processing" by B. Widrow et al, pages 349 and 350, an arrangement is known for receiving an input signal which contains both tonal and broadband noise and which is operable to remove at least part of the broadband component of the signal.

Currently, selective attenuation of tonal noise in a range without the need for trigger signals derived from the source of the tonal noise, which may affect multiple sources, is unknown.

Summary of the invention

According to the invention, an active control system for selectively attenuating tonal noise more than broadband noise affecting a hearing range is provided, comprising

a transducer device for generating sound in the audible range, which interacts with the noise in this range and results in at least a partial cancellation of the noise,

a sensor device having a sensor in the hearing range that provides a first signal related to the residual noise (attenuated noise) in the range and a second signal related at least in part to the noise affecting the hearing range with the exception of selective attenuation, and

a signal processing device for processing the first and second signals (or signals derived therefrom),

characterized in that

both the first and the second signals are representative of tone noise and broadband noise,

the signal processing device comprises an adaptive filter which is fed with a signal which results from at least the second signal and generates signals for driving the converter device, and a circuit arrangement for adjusting the adaptive characteristic of the filter to minimize the correlation between the input to the filter and the first signal,

the signal processing device comprises a device for subtracting a part of the signal which is based on the effect of the converter device from the second signal before introducing the signal into the adaptive filter, and

means for introducing an effective delay into the path of the second signal or a signal derived therefrom through the adaptive filter into the transducer means, the effective delay plus the acoustic delay from the transducer means to the sensor in the audible range being greater than the correlation period of the sound not to be attenuated, the transducer means being operable to produce sound for attenuating the tone noise to a greater extent than the broadband noise.

The sensor device can expediently have a first sensor in the audible range to generate the first signal, and a second sensor, which may or may not be in the audible range, to generate the second signal.

The processing in the processing device may involve a cross-relation of the effectively delayed signal (or signals derived therefrom or from which the signals themselves, and which may themselves be modified by filtering) and the monitoring signal (or signals derived therefrom, or from which the signals themselves are derived, and which may themselves be modified by filtering).

The processing in the processing device may comprise calculating the cross spectrum between the effectively delayed signal (or signals derived therefrom or from which the signals themselves are derived and which themselves may be modified by filtering) and the monitoring signal (or signals derived therefrom or from which the signals themselves are derived and which themselves may be modified by filtering).

The signal, which is at least partially related to the uncontrolled sound in the hearing area where selective attenuation is required, can be obtained from a sensor located near the hearing area where selective attenuation is required.

Alternatively, the signal related at least in part to the uncontrolled sound in the audible range where selective attenuation is required may be obtained by subtracting from the monitor signal the signals driving the transducer device (or signals derived therefrom, or from which the signals themselves are derived and which may themselves be modified by filtering).

The delay device can effectively delay the signal being received by delaying the received signal. Advantageously, this delay is adjusted by the adaptive filter, which itself is responsive to the spectrum or to an auto-correlation of the input to the delay device, so that the sum of this delay plus a delay in the transmission process from the transducer to the sensor giving the monitoring signal is greater than the correlation period of the noises to be attenuated and shorter than the correlation period of the noises to be attenuated. This means that the system is not in the Capable of attenuating sounds with short-term cross-relationships (ie, broadband random signals), and yet capable of attenuating sounds with long-term cross-relationships.

In an alternative approach, the effective delay is achieved by using narrowband filters in series with the adaptive filter, the narrowband filters being able to reject signals at frequencies other than the tonal frequencies to be attenuated. The narrowband filters may be fixed or tunable so that the center frequencies can be set to correspond to the frequencies of the tones to be rejected and their bandwidth can be set to include these tonal noises, the narrowband filters being set by the adaptive filter which is itself responsive to the spectrum or auto-correlation of the input to the delay device.

The adaptive filter may comprise one or more parallel adaptive filter sections, each having a characteristic determined at least in part by minimizing the cross-relationship between its input and the monitor signals. If the delay device comprises narrowband filters, the individual filter outputs are fed to one of each of the parallel adaptive filter sections of the adaptive filter. Each adaptive filter section may be a j-th order finite impulse response filter, with the coefficients adjusted using a gradient descent algorithm.

If a separate sensor is used to generate the signal for input to the delay device, this sensor is advantageously positioned so that it is insensitive to noise generated by the transducers or the influence of the transducer on the signal it generates is reduced by electronic subtraction.

The system can be built into an ear protector or a headset. Either one system is used or alternatively two systems are used, so that the quiet zone generally covers one or both ears.

When the system is incorporated in a headset or earmuff, the sensors providing the monitoring signal may be transducers designed to sense the unsteady pressure at a location positioned near each ear, for example a microphone located in close proximity to one of the ears. The sensors providing the signal as input to the first switching device may be a microphone located at or near the top of the earmuff, or alternatively a microphone located near the back of the housing of one or both earpieces; as a further alternative, no sensor is required to specifically generate this signal, the signal being derived from the monitoring signal as described. The transducer may be any device capable of generating unsteady pressure fluctuations, for example a loudspeaker. The transducers may be mounted in close proximity to the ear so that they affect the unsteady pressure in the area around one or both ears, for example the transducers may be mounted in the housing of one or both ear protectors.

The earmuff or earphone may be open at the back to allow the desired speech signals to enter easily; it may be part of a news headset in which desired sounds are reproduced through the loudspeakers. In addition, an effective delayed and adaptive filter in the form described above may be incorporated into the microphone channel picking up the speech to attenuate the background sound noise picked up by the microphone.

The invention is explained below in conjunction with the drawing using exemplary embodiments. It shows:

Fig. 1 is a block diagram of an embodiment of the invention,

Fig. 2 is a block diagram of another embodiment of the invention,

Fig. 3 is a block diagram of an embodiment of the invention in which the sensor signal is obtained from a sensor in the area to be controlled,

Fig. 4 is a block diagram of a modified embodiment of the invention in which the sensor signal is obtained from a sensor in the area to be controlled,

Fig. 5 an example of auto-correlations for a random noise and for tone noise,

Fig. 6 an embodiment of the invention incorporated into a headset or ear protector,

Fig. 7 shows another embodiment of the invention incorporated into a headset or ear protector, and

Fig. 8 shows another embodiment of the invention incorporated into a headset or head protector.

Fig. 1 shows a sensor 1 which generates a signal representative of the sound in the area 2. This signal 3 is delayed in a first circuit 4 and the resulting signal is introduced via an adaptive filter 5 into a transducer 6 which generates sound which interferes with the sound in the area 2. The coefficients of the adaptive filter 5 are adjusted by a second circuit 26 in accordance with an adaptive algorithm, described below, which uses a monitoring signal 7 from a sensor 8 in the area 2 to be controlled and the input 9 to the adaptive filter 5. The first circuit 4 is adjusted by the first adjustment device 10 in accordance with the spectrum or auto-correlation of the signal 3. The delay is adjusted so that that it is longer than the correlation period of the sound that is to be left undamped and yet smaller than the correlation period of the sound to be dampened. This delay can be made dependent on the frequency if a certain frequency selectivity is required.

The adaptive algorithm of the second circuitry 26 is an adaptive algorithm that adjusts the adaptive filter to minimize the correlation between the input of the filter and the monitor signal. These may be of the frequency domain type, involving the calculation of cross-spectrums, or of the time domain type, involving cross-correlation. Many algorithms of this type are described in the reference WIDROW and STEARNS "Adaptive Signal Processing". One such time domain algorithm is explained below.

The input signal 9 is denoted by the function u(t), and the output of the adaptive filter section 5 by the function y(t). Normally the signals are in sampled, digital form and have been converted by an analogue/digital converter either before or after the delay unit. The sampled version of the input and the output are represented as u(k) and y(k), where k represents the time constant. The signals are converted back to analogue form after the adaptive filter 5. The output y(k) is related to the input according to the equation

, where b(i) represents the i-th coefficient of the filter.

The output of the sensor 8, the monitor signal, is w(k) and this consists of two components v(k), the uncontrolled noise, and the component due to the transducer 6.

where c is the effect of the transducer characteristic. The noise in the region to be controlled is minimized by adapting the coefficients b (i) using a gradient descent algorithm (for the derivation of which reference is made to the reference by WIDROW & STEARNS "Adaptive Signal Processing", published in 1985 by Prentice Hall). where

and bj+1 (i) is the next update of the filter coefficient b₁ (i).

The expression r(ki) w(k) can be interpreted as a single sample estimate of the cross-correlation between the two signals r and w. Other approximations are possible, e.g.

where N can be any integer.

A particular novel feature of the present invention is the ability of the circuitry 11 to reduce the tonal noise to a greater extent than the wideband random noise. This feature is explained below. The adaptive filter section described above drives the correlation (for positive time delays) between the input and the monitor signal to zero by adjusting the coefficients of the filter to produce a cancellation noise to eliminate any noise contributing to the correlation between the two. By introducing a delay into the input signal, there is no longer any correlation between the two (for positive time delays) for wideband noises with correlation times shorter than the delay, so that the adaptive filter section does nothing. On the other hand, sound with a long correlation time, which is significantly is greater than the delay, a cross-correlation will occur between the delayed input and the monitor signal, and thus the adaptive filter will tend to cancel this to eliminate the correlation. Fig. 5a shows a typical cross-correlation of the input to the first circuit and the monitor signal when its sensors are close together and picking up a broadband sound. Fig. 5b shows the typical cross-correlation when its sensors are picking up a narrow-band (tonal) sound. The introduction of a delay D shifts the cross-correlation so that the origin is at the point marked D. Since the adaptive filter is only able to control sound with a significant level of cross-correlation to the right of the origin, only the narrow-band signal will be controlled. If the system is picking up a combination of many sounds, it will eliminate all those with a cross-correlation beyond the point D, thus eliminating all tones.

It is desirable that the signal 3 not be mixed with noise picked up by the sensor 1 from the transducer 6. This can be achieved by making the sensor 1 directional so that it is insensitive to sound from the transducer, or by positioning it so that it becomes insensitive, or by arranging an additional filter which takes the transducer signal as an input and whose output is used to subtract the influence of the transducer from the sensor signal in a manner similar to that described for eliminating the effect of the noise generated by the transducer on the monitor signal.

Other versions

Fig. 2 shows an arrangement similar to Fig. 1, but with a different internal design of the circuit arrangement 11. The signal 3 from the sensor is fed to one or more narrow band filters in a first circuit arrangement 12, the outputs 16 of which are fed, one each to each of the parallel adaptive filter sections of the adaptive filter 14. The output of all parallel sections is combined so that the drive for the converter 6 is formed. The adaptation of the individual parallel adaptive filter sections of the adaptive filter 14 is achieved by the second circuit arrangement 15 in dependence on the monitor signal 7 and each of the narrowband filter outputs 16. The second circuit arrangement uses the adaptive algorithm described above, the coefficients of each of the parallel adaptive filter sections being adapted in dependence on the corresponding input and the monitor signal.

The narrow band filters in a first circuit arrangement 12 are either fixed frequency or tunable. When fixed frequency they are provided in sufficient number and closely spaced frequency so that an audio signal of any frequency in the range of interest passes through one of these filters. The new feature of this embodiment which enables the selection process is described below.

When a tone noise is fed to the first circuitry with a bandwidth smaller than the bandwidth of the narrowband filters, an output is obtained from one of the filters which is a delayed (and hence phase-shifted) version of the input. Because the cross-correlation of the sensor and monitor signals has a long correlation period despite the delay introduced by the filter, this output has a correlation (for positive time delays) with the monitor signal and thus the corresponding parallel adaptive filter section of the adaptive filter 14 is adapted to produce an output signal for attenuating the tone noise.

If a wideband signal is fed to the first circuitry having a bandwidth substantially greater than the bandwidth of the narrowband filter, an output will be produced at each of the filters in the bandwidth of the original signal, but the outputs will be delayed versions of the input (delayed by a time equal to the reciprocal of the bandwidth of the narrowband filters), and because of the short correlation period of the original wideband signal, this (effectively) delayed signal will have a low cross-correlation with the monitor signal, and thus the parallel adaptive filter sections will produce a low output and the noise will be unattenuated.

The wideband filters may be made tunable to minimize the complexity of the system by reducing the number of narrowband filters. If only a few tones are to be attenuated, it may be convenient to use only a few narrowband filters, one for each tone. This can be achieved by monitoring the spectrum of the signal at the input to the first circuitry to identify the number and frequency of tones in the signal and adjusting the narrowband filters to correspond to those frequencies. The narrowband filters are continuously adjusted by the first adaptation unit 13 to maintain their center frequency close to the frequency of the corresponding tone. Their bandwidth is adjusted to ensure that it is wider than the tone to be attenuated, but not too wide so that wideband signals are passed with insufficient delay. This processing is automatic.

Fig. 3 shows how the sensor signal can be derived from the sensor 8 in the area 2 to be controlled. The output 17 from the circuit arrangement 11, which in this figure is similar to that of Fig. 1, is fed to a filter 18. The characteristic of this filter is set to correspond to the transfer function between signal 17 and signal 7. This is identical to the filter c(i) used in the first adaptation unit 6. The output of the filter 18 is subtracted from the monitor signal 7 to obtain an equivalent signal 3. The characteristic of the filter 18 can be updated to maintain it as an exact representation of the transfer function.

Fig. 4 shows the equivalent circuit where the input to the first circuitry and monitor signals from the same sensor are derived for the circuitry 11 which uses narrow band filters.

Fig. 6 shows the system incorporated in an earmuff or headset. There are two systems, one for each ear, in a portable container 20 which contains the battery 21 for power supply. The earmuff may be an open protector so that the desired sound can reach the ear unhindered. The sensor 1 which generates the input to the first circuit arrangement is a microphone 19 mounted on the housing of the earpiece (or pinna), and the sensor 8 which generates the monitor signals is a microphone housed in the earpiece near the ear. The transducer 6 is a loudspeaker built into the housing of the earpiece.

Fig. 7 shows a system in which a microphone 22 is used to provide the input to the first switching arrangement for both circuits 11.

Fig. 8 shows an embodiment in which no separate sensor is present to obtain an input to the first circuit arrangement (the input is derived from the monitor signals in the manner described above).

An additional input may be given to one or each of the loudspeakers carrying desired communication signals for the wearer, and these signals are converted into sound audible to the wearer. This sound is unaffected by the circuitry 11. When the ear protector forms part of a communication headset, a microphone is attached to the headset to receive the wearer's speech. In this case, it may be desirable to insert an effectively delayed and adaptive filter section into the voice communication channel to eliminate the background audio noise picked up by the speech microphone.

The effective delay may be achieved by introducing a delay into the signal either before, during or after the adaptive filtering, or a combination thereof. Thus, according to Figures 1 - 3, part or all of the delay introduced by the device 4 may be incorporated into the filter 5, or by a discrete device inserted between the filter 5 and converter 6. The second circuit 76 must of course be adjusted accordingly.

In the above description and the claims, reference is made to a cross-correlation of the two signals. This expression is defined as follows: The cross-correlation (c) of two digitally sampled signals r(i) and w(i) (where the offset between the two cross-correlation signals is n) is given by:

where N can be any number and, as stated above, n is the number of sampled locations by which one signal is offset from the other.

Summary

An active control system for attenuating audio noise in a defined area is described. In its basic form, the system comprises sensors (1, 8) for generating signals indicative of the residual noise in the area after attenuation and the uncontrolled sound affecting the area, signal processing circuits (10, 26) for processing the generated signals in different ways depending on the audio content, an adaptive filter (5) fed with at least one of the generated signals, the characteristics of which are controlled by the processing circuit (10), a transducer (6) for generating an audio noise attenuating perturbation in the area, and a delay device (4) for delaying signals relating to the uncontrolled noise before or after or during the adaptive filtering. The system is directly applied in a personal headset or ear protector.

Claims (20)

1. Active control system for selectively attenuating sound noise more than broadband noise affecting a hearing range (2), with a transducer device (6) for generating sound in the hearing range (2) which interacts with the noise in this range and results in at least a partial cancellation of the noise, a sensor device with a sensor (8) in the hearing area (2) which provides a first signal (7) related to the residual noise (attenuated noise) in the area, and a second signal (8) related at least in part to the noise that affects the hearing area (2) with the exception of selective attenuation, a signal processing device (11) for processing the first and second signals (or signals derived therefrom), characterized in that both the first and the second signals are representative of tone noise and broadband noise, the signal processing device comprises an adaptive filter (5) which is fed with a signal which results from at least the second signal (3) and generates signals for driving the converter device (6), and a circuit arrangement (26) for adjusting the adaptive characteristics of the filter for minimizing the correlation between the input (9) to the filter and the first signal (7), the signal processing device (11) comprises a device for subtracting the part of the signal (3) which is based on the effect of the converter device (6) from the second signal (3) before introducing the signal into the adaptive filter (5), and a device (4) for introducing an effective delay into the path of the second signal (3) or a signal derived therefrom via the adaptive filter (5) into the transducer device (6), the effective delay plus the acoustic delay from the transducer device to the sensor in the audible range being greater than the correlation period of the sound which is not to be attenuated, the transducer device being operable to generate sound for attenuating the tonal noise more than the broadband noise. 2. Active control system according to claim 1, characterized in that the sensor device comprises the sensor (8) for generating the first signal (7) and a second sensor (1) for generating the second signal (3). 3. Active control system according to claim 2, characterized in that the second sensor (1) is also effective in the hearing range. 4. Active control system according to claim 2, characterized in that the second sensor (1) is arranged in an area which is unaffected by sound from the transducer. 5. Active control system according to one of claims 1 - 4, characterized in that the delay is applied to the signal (7) which is to be applied to the adaptive filter (5). 6. Active control system according to one of claims 1 - 4, characterized in that the delay is built into the adaptive filter (5). 7. Active control system according to one of claims 1 - 4, characterized in that the delay is applied to the signal (3) used to drive the converter device. 8. Active control system according to one of claims 1 - 7, characterized in that the processing device (26) serves to cross-correlate the first signal (7) or a signal derived therefrom, and the signal (9) applied to the adaptive filter (or a signal derived therefrom, or from which the signal applied to the adaptive filter is derived, and which is itself modified by the filtering process). 9. Active control system according to one of claims 1 - 8, characterized in that the processing device (26) serves to calculate cross-spectra of the first signal (7) or a signal derived therefrom, and of the signal (9) which is applied to the adaptive filter (or a signal which is derived therefrom, or from which the signal applied to the adaptive filter is derived and which is itself modified by the filtering process). 10. Active control system according to one of the preceding claims, characterized in that the effective delay is realized by using narrowband filters (12) in series with the adaptive filter, the narrowband filters being able to reject signals at frequencies different from the audio frequencies to be attenuated. 11. Active control system according to claim 10, characterized in that a set of parallel narrow band filters (12) is used. 12. Active control system according to one of the preceding claims, characterized in that the signal applied to the adaptive filter (5) is made insensitive to the converter device (6) by using a combination of the second signal (3) (or a signal derived therefrom) and the signal applied to the converter device (6) (or a signal derived therefrom or from which the signal is derived) or the first signal (or a signal derived therefrom). 13. Active hearing device, headphones or defensive hearing protection with an active control system according to one of claims 1 - 12, characterized in that the Sound noise at the listener's ear is reduced more than ambient noise. 14. Active hearing device, headphones or defensive hearing protection according to claim 13, characterized in that a separate control system is connected to each of the ear covering devices. 15. Active hearing device, headphones or defensive hearing protection according to claim 13 or 14, characterized in that the sensor or sensors (1, 8) is or are a microphone which is or are arranged on the headband of the headphones or on the outside of the housing of the transducer device. 16. Active hearing device, headphones or defensive hearing protection according to claim 13, 14 or 15 in conjunction with claim 4, characterized in that a single sensor (8) is used for each ear. 17. Active hearing device, headphones or defensive hearing protection according to one of claims 13 - 16, characterized in that the transducer device (6) is mounted so that it is arranged close to the ears of the wearer during operation. 18. Active hearing device, headphones or defensive hearing protection according to one of claims 13 - 16, characterized in that each transducer device (6) is mounted in a closed cavity which is formed by a shell (19) enclosing the ear. 19. Active hearing device, headphones or defensive hearing protection according to claim 18, characterized in that the material properties of the shell (19) are selected so that they enhance the transmission of sound in a selected frequency range. 20. Active hearing device, headphones or defensive hearing protection according to one of claims 13 - 19, in conjunction with claim 8, characterized in that the cross-correlation is calculated directly or recursively over a single sample or over a number of samples.