WO2000001198A1

WO2000001198A1 - System for reducing the effects of acoustically noisy environments on detected sound signals

Info

Publication number: WO2000001198A1
Application number: PCT/US1999/012817
Authority: WO
Inventors: Jon C. Taenzer; Steven H. Puthuff
Original assignee: Resound Corporation
Priority date: 1998-06-30
Filing date: 1999-06-23
Publication date: 2000-01-06
Also published as: AU4820099A

Abstract

The intensity of an incoming sound signal in a pickup device is detected and used to control the signal output by a microphone (20) of the device. In an exemplary configuration, a signal proportional to the microphone output signal is applied to a level detector (26) which senses for example the average amplitude of the signal. The level detector (26) then generates a control signal (fc) representative of the level of the sound detected by the microphone (20). The control signal (fc) is applied to a filter (24) receiving the microphone output signal and operates to dynamically control the frequency of the filter (24) and its effects on the microphone output signal. The filter is thus used to reduce interfering signal frequencies in proportion to intensity of the sensed sound signal.

Description

SYSTEM FOR REDUCING THE EFFECTS OF ACOUSTICALLY NOISY ENVIRONMENTS ON DETECTED SOUND SIGNALS

BACKGROUND OF THE INVENTION

FIELD OF THE INVENTION The invention relates to sound pickup devices, and more particularly, to a system for improving sound pickup device immunity to ambient acoustic noise.

DESCRIPTION OF RELATED ART

It is recognized that the greater energy in acoustically noisy ambient environments tends to be due to low frequency sounds. Examples of such noise include the roar of a fire, the rumble of distant thunder, or the drone of an air conditioner or industrial machinery.

It is also recognized that a characteristic of human speech is that the bulk of information is carried by the consonant sounds rather than the vowel sounds. Consonant sounds generally fall in the higher end of the human speech sound spectrum. Hence most of the information in speech is at the higher frequencies. From the above, a logical conclusion is that most acoustically noisy environments would not interfere with a listener's ability to comprehend speech because the low frequency nature of environmental noise would have little impact on the information-bearing consonant sound component of speech which falls at the high frequency end of the speech spectrum. However, this conclusion relies on an overly simplistic view of the dynamics of human hearing. In actuality, the auditory sensor nerves comprising part of the human hearing mechanism, although tuned to respond to discrete frequencies, can nonetheless be triggered by noise of sufficient intensity at frequencies that are lower than those at which the sensitivity of the nerves is greatest. This effect, referred to as the "upward spread of masking", limits a listener's perception of speech when low frequency noise of sufficient intensity is present, despite the fact that the noise and the information- laden speech component generally occupy different frequency regions. The result is that low frequency noise undesirably interferes with speech intelligibility and a conversation between two people in the presence of high intensity low frequency noise is disrupted. More importantly, in a communication system in which a remote listener is receiving sounds picked up at a talker's location via a sound pickup device, the presence of ambient noise at the talker's location can significantly interfere with the listener's perception of the talker's speech. Again this is primarily due to the upward spread of masking effect of the listener's physiological hearing apparatus, with the high intensity, low frequency noise limiting the sensitivity of the listener's hearing to speech sounds at higher frequencies.

The prior art has attempted to address this problem by filtering out noise signals using for example high pass filters. Fixed filters, or fixed frequency filters with dynamically changing gains in one or more bands into which the noise falls have thus been used to improve intelligibility of speech transmitted through a communication system. One such system, known as Speech Filtering, assigns multiple bands to an incoming sound signal via DSP FFT (Digital Signal Processing Fast Fourier Transforms). The gain of the bands containing non- speech like signals is subsequently reduced before the inverse FFT process is performed, resulting in improved sound quality, but not improved understanding. Also, these systems are complex and costly in terms of power consumption and requirements of computing power and the amount of circuitry involved. SUMMARY OF THE INVENTION The invention overcomes the deficiencies of the prior art by providing a simple signal processing system which reduces the harmful effects of an acoustically noisy ambient environment on the intelligibility of speech detected by a sound pickup device.

In accordance with the invention, the intensity of an incoming sound signal is detected and used to control the signal output by the microphone of a sound pickup device. In an exemplary configuration, a signal proportional to a microphone output signal is applied to a level detector which senses for example the average amplitude of the signal. The level detector then generates a control signal representative of the level of the sound detected by the microphone. The control signal is applied to a filter receiving the microphone output signal and operates to dynamically control the cutoff, or critical frequency of the filter and its effects on the microphone output signal. The filter is thus used to reduce interfering signal frequencies in proportion to intensity of the sensed sound signal. In accordance with the preferred embodiment of the invention, a high pass filter is used in the signal path, with its critical frequency being thus controlled by the control signal. In this manner, as the acoustic ambient noise level increases and the talker naturally responds by elevating his/her voice, the level detector in response increases the critical frequency of the high pass filter. Accordingly, less noise is permitted to pass, rendering the resulting speech signal more intelligible to the listener because the effect of upward spread of masking by his/her physiological hearing apparatus is reduced. As discussed above, since most intelligibility lies in the consonant component of speech, the resultant reduction of vowel intensity has minimal effect on the intelligibility of the speech.

In accordance with a second embodiment of the invention, the signal from the high pass filter is combined with a direct feed from the microphone. This achieves improved signal control. In accordance with another embodiment of the invention, the compression ratios of the sound processing circuit are controlled to achieve the dynamic filtering.

In accordance with yet another embodiment of the invention, the signal fed to the level detector is itself filtered using a lowpass filter, making the level detector primarily responsive to the generally low frequency noise component of the incoming sound signal. The control signal is thus generated in accordance with the noise component of the detected sound and the highpass filter operating on the microphone signal is controlled according to the lowpass filtered signal received by the level detector. The highpass filter can be provided with other filtering shapes depending on the characteristics of the noise being reduced.

BRIEF DESCRIPTION OF THE DRAWINGS

Many advantages of the present invention will be apparent to those skilled in the art with a reading of this specification in conjunction with the attached drawings, wherein like reference numerals are applied to like elements and wherein:

FIG. 1 is a schematic diagram of a circuit in accordance with the preferred embodiment of the invention;

FIG. 2 is a graphical representation of the dynamic filtering characteristics of the circuit of FIG. 1;

FIG. 3 is a graphical representation of typical noise frequency-versus- amplitude characteristics;

FIG. 4 is schematic diagram of a circuit in accordance with a second embodiment of the invention; FIG. 5 is a graphical representation of the dynamic filtering characteristics of the circuit of FIG. 4; FIG. 6 is a graphical representation of the critical frequency shift achieved using a compression ratio greater than one;

FIG. 7 is a graphical representation of the critical frequency shift achieved using compression ratio manipulation in accordance with the invention; FIG. 8. is a schematic diagram of a circuit in accordance with a third embodiment of the invention;

FIG. 9 is a schematic diagram of a circuit in accordance with a fourth embodiment of the invention;

FIG. 10 is a schematic diagram of a circuit in accordance with a fifth embodiment of the invention;

FIG.11 is a graphical representation of a first transfer characteristic of the circuit of FIG. 10;

FIG. 12 is a graphical representation of the noise encountered in certain acoustically noisy environments; FIG. 13 is a schematic diagram of a conventional state-variable filter.

FIG. 14 is a graphical representation of a second transfer characteristic of the circuit of FIG. 10;

FIG. 15 is a graphical representation of a third transfer characteristic of the circuit of FIG. 10; and FIG. 16 is a block diagram of a general noise reduction circuit in accordance with the invention.

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows a circuit diagram of a system in accordance with the invention. As shown, a sound signal is detected by a microphone 20, which can be any known sound transducer such as a directional or omnidirectional microphone or near field microphone forming part of a sound pick up device. The signal from microphone 20 is passed to a pre-amplifier 22. The output of preamplifier 22 is provided to a high pass filter 24 and to a level detector 26. The level detector 26 senses the level of the signal from microphone 20, by for example measuring its average amplitude, RMS amplitude, peak amplitude or intensity (power), and provides as an output an f, control signal to the high pass filter 24. The f_c control signal controls the critical frequency of the high pass filter 24-known as a sliding pole filter-and in operation raises the cutoff frequency as the sound intensity rises. In this manner, as the loudness level of the signal detected by the microphone 20 increases, high pass filter 24 blocks out progressively more lower frequency sounds.

The significance of the scheme discussed above can be more fully appreciated with reference to FIGS. 2 and 3, showing plots of signal amplitude versus frequency. FIG. 3 shows a typical type of noise spectrum in which the higher noise intensities are at the low frequencies but decrease at the higher frequencies. Specifically, as noise increases, the shape of the noise often does not change significantly-for example, if the talker simply moves closer to a noisy blower or an airplane taxis toward a ground crew member talking into a pickup device, the general character of the noise remains similar.

FIG. 2 shows the transfer characteristic of the circuit of FIG. 1. The arrow in FIG. 2 shows the direction of critical frequency shift implemented by the circuit of FIG. 1 with increasing sound intensity. As can be seen from FIG. 2, this shift has little effect on the consonants of speech, which carry the majority of speech information, since the information-bearing consonant component of speech typically lies in the higher end of the sound spectrum. By contrast, the f, shift exhibited by the circuit of *FIG. 1 has a significant effect on blockage of unwanted noise, plotted in FIG. 3, whose shown increased intensity is anticipated and corrected by the changing pass band of the dynamically controlled filter 24. A comparison of FIGS. 2 and 3 shows how the inverse frequency characteristic of the circuit in accordance with the invention automatically corrects for the change in noise amplitude. It is also possible to feed the microphone signal, with a certain amount of attenuation, straight through to the output and then to add to it a high pass filtered portion of the same signal. Such a configuration is shown in FIG. 4, in which like numerals are used for like elements, and in which an adder is used for adding the signal from microphone 20, via pre-amplifier 22 and attenuator 25, to the signal from the high pass filter 24. The signal transfer characteristic achieved by this circuit is shown in FIG. 5, with filter critical frequency ξ again exhibiting the varying reduction of bass frequencies for improved intelligibility. As can be seen from FIG. 5, the lower flat part of the curve is the attenuated signal (at low frequency), while the upper flat part is due to the sum of the two signals. The FIG. 5 response achieved by the circuit of FIG. 4 would be especially beneficial for products used by for instance by maintenance workers in noisy environments since the equipment responsible has noise characteristics more accurately represented in FIG. 12, wherein the noise has a umform intensity up to a particular frequency, beyond which the intensity drops. The circuit of FIG. 4 also finds particular application in providing an improved product because it still exhibits some bass response, although attenuated. The amount of attenuation in this design is selected at the time of product design based on the intended application. The circuits of FIGS. 1 and 4 are designed to provide a linear signal transfer. If compression is desired it can be achieved using subsequent processing circuitry relying on for example one or more conventional analog compression circuits or digitally programmable gain control devices (not shown). A typical environment where this type of circuit can be found is in programmable hearing aid devices, and more specifically, in proprietary ARP™ (Advanced Resound™ Processor) integrated circuits whose electronic configuration can be modified in accordance with the present invention to realize the improved sound pickup performance contemplated. Such systems are disclosed in for example U.S. Patent No. 4,882,761 entitled "Low Voltage Programmable Compressor" to Waldhauer, U.S. Patent No. 4,882,762 entitled "Multi-Band Programmable Compressed System" to Waldhauer, U.S. Patent No. 5,278,912 entitled "Multi-band Programmable Compression System" to Walhauer, and U.S. Patent No. 5,488,668 to Waldhauer entitled "Muti-band Programmable Compression System," all of which are assigned to the same assignee of record. It is also contemplated that a compression ratio greater than one can be beneficially applied. For example, in the circuit of FIG. 9 in accordance with the invention, the microphone signal is fed to a bandsplit filter 34 having two outputs, HP and LP. Output HP contains frequencies above a bandsplit critical frequency f_c, while output LP contains frequencies below f_c. A standard filter possessing this property is referred to as a state variable filter and is illustrated in more detail in FIG. 13. By processing the LP signal through a compressor (32) before recombining the two signals in adder 28, the overall transfer characteristic of FIG. 6 is achieved. The critical frequency shift of such a circuit can be effected by manipulating the bandsplit filter block 34. Thus ratiometric compression can be effected where the signal from the microphone is bandsplit into two bands (with, for example, LP being compressed and HP either being not compressed or expanded), compression being appropriately selected to effect the bass reduction shifting characteristics shown in FIG. 7. As an example, at low noise, compressor 32 has gain = 1. As signal increases, by for example 10 dB, a 2:1 compression ratio would give a 5 dB increase to bass in the low-pass region while treble goes up the full 10 dB in the high-pass region. When noise increases by 40 dB, then there is 20 dB attenuation of bass while is also increased due to f^ control of the bandsplit filter. By changing compression ratio and rate of ξ change, almost any beneficial characteristic can be created. Those skilled in the art will recognize that numerous transfer characteristics can be achieved with appropriate combinations of bandsplitting, gain control and compression and/ or expansion, with the particular combinations being dictated by the noise reduction application. The circuit of FIG. 9 can be implemented using circuit elements contained within the ARP™ integrated circuit since the ARP™ contains all of the elements shown in FIG. 9, except for the microphone 20. Advantages of the ARP circuit include its very small size, low power consumption, operation from a single hearing aid battery, excellent dynamic range and very low audio self noise, making it particularly well-suited for use in communication headsets, earphones, hearing aids and the like.

Another circuit for effectively achieving a critical frequency shift is shown in FIG. 10. As FIG. 10 shows, the ξ control signal from level detector 26 and the microphone signal from pre-amplifier 22 are provided to controlable bandsplit filter 34. Compression circuits 32 and 36 operate respectively on outputs LP and HP of bandsplit filter 34, and the outputs of the compressor circuits are added in adder 28. It should be noted that while circuits 32 and 36 are referred to as "compression" circuits, the compression may be implemented using known "compander" circuits which, depending on the application, may either compress or expand the signal. For example, with reference to FIG. 11, compressor circuit 36 exhibits signal expansion. The resultant transfer characteristic shown in FIG. 11 is especially desirable in that it keeps the loudness of the output relatively constant as the input ambient noise increases. This advantage is created by a simultaneous reduction in bass response for a noise reduction accompanied by an increase of the intensity of the consonant portion of speech. Other desirable transfer characteristics, such as those illustrated in FIGS 13 and 14, may be readily achieved by the circuit of FIG. 10, depending on the particular application.

Another embodiment in accordance with the invention is shown in FIG. 8, wherein the output signal from the microphone 20 is filtered before being provided to the level detector 26. The output signal is applied to level detector 26 through low pass filter 30. In this manner, the level detector 26 becomes primarily responsive to the generally low frequency noise component of the incoming sound signal. The f_c control signal is thus generated in accordance with the noise component of the detected sound, and the high pass filter 24 operating on the signal from microphone 20 is controlled according to the lowpass filtered signal received by the level detector 26. Of course it is to be understood that the filters 24 and 30 can be provided with other filtering shapes depending on the characteristics of the noise being reduced.

The above circuits are designed with a view to provide noise reduction filtering immediately at the microphone in order to minimize the noise load carried through the rest of the system. However, it will be appreciated that the same dynamic filtering expedient could be utilized anywhere in the signal chain in a system in which electrical signals represent speech in noise, with the noise having a characteristic similar to that of ambient environmental noise. Thus, as shown in FIG. 15, the noise reduction system 40 in accordance with the invention can be designed to operate on signals from any combined speech-noise source 42 such as a tape recorder, compact disk player, etc., or in real-time devices such as two-way radios in which the noise reduction can be implemented at the remote receiving end. Noise reduction system 40 represents any of the above-described noise reduction schemes.

The above are exemplary modes of carrying out the invention and are not intended to be limiting. It will be apparent to those skilled in the art that modifications thereto can be made without departure from the spirit and scope of the invention as set forth in the following claims.

Claims

Claims:

1. A sound pickup system comprising: a microphone for generating an electrical signal representative of a detected sound signal; a level detector for generating a control signal in proportion to the electrical signal; and a controllable filter for selectively filtering the electrical signal from the microphone in accordance with the control signal to thereby generate an output signal.

2. The sound pickup system of Claim 1 , wherein the controllable filter is a high pass filter and wherein the control signal controls the critical frequency of the high pass filter.

3. The sound pickup system of Claim 1 , wherein the controllable filter is a bandsplit filter adapted to divide the electrical signal into first and second frequency band components differentiated by a critical frequency, the control signal controlling said critical frequency.

4. The sound pickup system of Claim 1, further comprising a low pass filter, the electrical signal being provided to the level detector through the low pass filter.

5. The sound pickup system of Claim 1 , further comprising a preamplifier for amplifying the electrical signal prior to detection by the level detector.

6. The sound pickup system of Claim 1 , wherein the level detector comprises an average amplitude detector.

7. The sound pickup system of Claim 1, wherein the level detector comprises a signal peak detector.

8. The sound pickup system of Claim 1, wherein the level detector comprises a signal intensity detector.

9. The sound pickup system of Claim 1, further comprising an attenuator for attenuating the electrical signal to thereby generate an attenuated signal and an adder for adding the attenuated signal and the output signal.

10. The sound pickup system of Claim 1, wherein the output signal is applied to a compression circuit.

11. A method for dynamically filtering an electrical signal representative of detected sound, the method comprising: generating the electrical signal using a sound transducer; detecting a characteristic level of the electrical signal; and selectively filtering the electrical signal in accordance with the characteristic level.

12. The method of Claim 11, wherein the step of selectively filtering comprises: applying the electrical signal to a controllable high pass filter to thereby generate an output signal; and controlling a critical frequency of the high pass filter in accordance with the characteristic level.

13. The method of Claim 12, wherein the characteristic level is signal amplitude.

14. The method of Claim 12, wherein the characteristic level is signal peak.

15. The method of Claim 12, wherein the characteristic level is signal intensity.

16. The method of Claim 12, further comprising the step of applying the electrical signal to a low pass filter prior to detecting the amplitude of the electrical signal.

17. The method of Claim 12, further comprising the steps of: attenuating the electrical signal to thereby generate an attenuated signal; and adding the attenuated signal to the output signal.

18. The method of Claim 11, wherein the step of selectively filtering is effected using ratiometric compression.

19. The method of Claim 11, wherein the ratiometric compression comprises the steps of: dividing the electrical signal into first and second frequency band components differentiated by a critical frequency; applying signal compression to at least one of said first and second frequency bands; and adding the first and second frequency bands.

20. The method of Claim 19, wherein the signal compression comprises signal expansion.

21. The method of Claim 19, wherein the characteristic level is signal amplitude.

22. The method of Claim 19, wherein the characteristic level is signal peak.

23. The method of Claim 19, wherein the characteristic level is signal intensity.

24. The method of Claim 19, wherein the characteristic level is signal RMS value.

25. A dynamic filtering system comprising: a signal source for providing a source signal having a speech component and an acoustic ambient noise component; a level detector for generating a control signal in proportion to the source signal; and

a controllable high pass filter for selectively filtering the source signal in accordance with the control signal to thereby generate an output signal.

26. The sound pickup system of Claim 25, wherein the control signal controls the critical frequency of the controllable high pass filter.

27. The sound pickup system of Claim 26, further comprising a low pass filter, the source signal being provided to the level detector through the low pass filter.

28. The sound pickup system of Claim 27, wherein the level detector comprises an average amplitude detector.

29. The sound pickup system of Claim 25, wherein the level detector comprises a signal peak detector.

30. The sound pickup system of Claim 25, wherein the level detector comprises a signal intensity detector.

31. The sound pickup system of Claim 25, wherein the level detector comprises a signal RMS value detector.

32. The sound pickup system of Claim 25, further comprising an attenuator for attenuating the source electrical signal to thereby generate an attenuated signal and an adder for adding the attenuated signal and the output signal.

33. The sound pickup system of Claim 25, wherein the output signal is applied to a compression circuit.