United States Patent (19)
`Chang
`
`Patent Number:
`11
`(45) Date of Patent:
`
`4,912,767
`Mar. 27, 1990
`
`(56)
`
`(54) DISTRIBUTED NOISE CANCELLATION
`SYSTEM
`Inventor:
`(75
`73 Assignee:
`
`Robert W. Chang, Oakton, Va.
`International Business Machines
`Corporation, Armonk, N.Y.
`(21) Appl. No.: 167,619
`(22
`Filed:
`Mar. 14, 1988
`51) Int. Cl“................................................ G10 L 5/06
`52 U.S. Cl. ......................................... 381/47; 381/46
`58 Field of Search .................................... 381/41-47,
`381/71, 94; 364/513.5
`References Cited
`U.S. PATENT DOCUMENTS
`4,594,695 6/1986 Garconnat et al. ................... 381/71
`4,625,083 11/1986 Poikela ....................
`... 381/46
`4,630,305 12/1986 Borth et al. ........................... 381/94
`4,649,505 3/1987 Zinser ................................... 381/94
`4,737,976 4/1988 Borth et al. ........................... 381/46
`FOREIGN PATENT DOCUMENTS
`2188763 10/1987 United Kingdom .
`OTHER PUBLICATIONS
`Processing',
`'Voice
`and
`Speech
`Parsons,
`McGraw-Hill, 1976, pp. 345-369.
`Boll, "Suppression of Acoustical Noise in Speech Using
`
`Spectral Subtraction', IEEE Trans ASSP, vol. ASSP
`27, No. 2, 4/79.
`Harrison, "Adaptive Noise Cancellation in a Fighter
`Cockpit Environment', IEEE, 1984, pp. 18A..4.1-1-
`8A.4.4.
`Primary Examiner-Gary V. Harkcom
`Assistant Examiner-John A. Merecki
`Attorney, Agent, or Firm--Douglas M. Clarkson
`57
`ABSTRACT
`A method and system for cancelling noise from sources
`that are distributed over a region, whereby two sensors
`are located so that a first sensor will detect both voice
`signals and noise signals, and a second sensor will detect
`only the noise signals. The voice signals picked up at the
`second sensor are negligible, and the noise signals
`picked up at both sensors are correlated. The signals
`output from each sensor are connected to a predeter
`mined number of narrowband filters in order to divide
`each respective signal into a predetermined number of
`frequencies, such as 15 for example. Thereafter, both
`signals are combined to cancel effectively the noise
`component from the signal output having both voice
`and noise to leave a voice signal that is substantially
`noise free.
`
`7 Claims, 3 Drawing Sheets
`
`
`
`Voice
`Recognition
`
`Adaptive Filters
`Narrowband filters
`
`LGE EXHIBIT NO. 1015
`
`- 1 -
`
`Amazon v. Jawbone
`U.S. Patent 8,467,543
`Amazon Ex. 1015
`
`

`

`U.S. Patent Mar. 27, 1990
`
`Sheet 1 of 3
`
`4,912,767
`
`
`
`
`
`Voice
`Recognition
`
`FG.
`(PRIOR ART)
`
`- 2 -
`
`

`

`U.S. Patent Mar. 27, 1990
`
`Sheet 2 of 3
`
`4,912,767
`
`
`
`Voice
`Recognition
`
`
`
`15-7
`
`Adaptive Filters
`Narrowband filters
`
`FIG 2
`
`- 3 -
`
`

`

`FREQUENCY IN Hz
`
`- 4 -
`
`

`

`1.
`
`DSTRIBUTED NOSE CANCELLATION SYSTEM
`
`10
`
`BACKGROUND OF THE INVENTION
`1. Field of the Invention
`The present invention, generally, relates to a method
`and a system for cancelling noise from noise-corrupted
`speech and, more particularly, to an improved method
`and system for rendering speech recognizable in a high
`noise environment, particularly where noise is distrib
`uted.
`One glance into the cockpit of today's commercial
`airliner would give an idea of the hands-busy, eyes-busy
`environment that exists there, and this is more true of 15
`the cockpit in today's military aircraft. The military has
`solved their problem somewhat by the use of voice
`actuated controls for many activities, such as located in
`the cockpit of a fighter aircraft, and this has been ac
`complished through the use of voice recognition sys
`20
`tems.
`It was realized early that, due to the relatively high
`noise in the cockpit of a fighter aircraft, some form of
`noise cancellation was required, and from that need, an
`adaptive filter noise cancellation technique was devel
`25
`oped that has become a standard in the industry. More
`recently, that technique was tried in military helicop
`ters, and it was found to be ineffective.
`2. Description of the Prior Art
`It is understandable that the presence of high levels of 30
`noise in an audio signal will produce a substantial reduc
`tion in the intelligibility of speech, and it has been found
`that the most advanced voice recognition equipment is
`seriously ineffective in recognizing the simplest words
`f 35
`in the high noise levels encountered in the cockpit o
`today's tactical fighter aircraft. A technique that was
`proposed by Bernard Widrow et al. in 1975, known as
`Adaptive Noise Cancellation (or ANC), has been tested
`extensively at the Research Laboratory of Electronics
`at the Massachusetts Institute of Technology.
`The Widrow technique is described in an article that
`is entitled "Adaptive Noise Cancelling: Principles and
`Applications', Proc. IEEE, Vol. 63, No. 12, December,
`1975.
`45
`During the M.I.T. tests, some improvements were
`developed in the Widrow technique, such as placements
`for the two microphones in a fighter cockpit environ
`ment as being one inside the oxygen facemask of the
`pilot and the second microphone outside the facemask.
`50
`The one microphone, called the "primary' micro
`phone, is located to sense, or to detect, the voice of the
`pilot plus the noise.
`The second, or "reference', microphone is located to
`sense, or detect, principally the noise. By locating the
`reference microphone outside the oxygen facemask,
`very little of the pilot's voice is picked up.
`The engineers at M.I.T. learned also that it is better to
`have the signal-to-noise ratio of the primary micro
`phone large compared to the signal-to-noise ratio of the
`60
`reference microphone, so that the adaptive filter can be
`kept as small as possible. Otherwise, the adaptive filter
`must either estimate the delay between the primary and
`reference signals or have a long impulse response in
`order to provide good cancellation of the noise from the
`primary signal.
`A report of the M.I.T. engineers is given in a paper
`entitled "Adaptive Noise Cancellation in a Fighter
`
`4,912,767
`2
`Cockpit Environment' by Harrison, Lim and Singer,
`1984 IEEE, pages 18A..4.1 through 18A..4.4.
`With all of the expertise of these M.I.T. engineers, the
`conclusion was that the Adaptive Noise Cancellation
`technique of Widrow, while effective enough in an
`environment with a localized noise source, degrades in
`performance when there is more than one noise source
`present or when the noise source is distributed over a
`region. Actually, the many sources of noise in a helicop
`ter make the Adaptive Noise Cancellation technique
`virtually ineffective in that high noise environment
`where the noise sources are distributed over a wide
`region. While those experts in the field departed to
`study the use of additional reference microphones in a
`distributed noise environment, the present invention
`proceeds with the development of a unique solution to
`this perplexing problem.
`A review of the prior patent art reveals very little to
`assist in developing a solution such as provided by the
`present invention. For example, U.S. Pat. No. 4,625,083
`to Poikela is concerned with providing a voice operated
`switch that is capable of distinguishing between voice
`and noise. By using one microphone primarily for
`speech and one microphone primarily for ambient noise
`signals, each of these groups of signals have a certain
`sound pressure level, and since it is desired to have the
`sound pressure level of the speech signal always exceed
`that of the noise signal, this is accomplished in two
`ways. One way is by placing the two microphones in
`predetermined locations so that the sound pressure level
`distinctions are realized, and another way is by limiting
`the width of the frequencies, like that customarily used
`in telephone receivers. A typical frequency range is 100
`hertz to 4 kilohertz, but a narrower frequency range of
`250 hertz to 3.5 kilohertz is termed as being satisfactory.
`By connecting both signals to a differential amplifier, an
`output will result when there is speech, and there is no
`output when there is no speech.
`U.S. Pat. No. 4,649,505 to Zinser, Jr. et al. is an exam
`ple of another attempt to improve on the basic adaptive
`filter of Widrow, identified supra, but this effort is for
`the purpose of eliminating crosstalk between speech and
`noise signals. It discloses the use of a speech input, a
`noise input and a reference input with a reference noise
`portion and a crosstalk speech portion to a digital signal
`processing microcontroller, a read-only-memory and a
`random access memory, from which the signals are
`processed digitally. After the inputs are converted first
`from analog to digital signals, they are converted next
`from digital serial signals to digital parallel signals for
`further processing. There is no mention of the problem
`with which the present invention is concerned.
`U.S. Pat. No. 4,658,426 to Chabries et al. discloses
`several different forms of noise suppression devices for
`use where the signal-to-noise ratio is poor at the input
`and where the characteristics of the adaptive filter ad
`just automatically to variations in the input signal.
`These adjustments utilize time and frequency domains
`in making the adaptive filter adjustments in order to
`filter noise, and a mathematical description is given in
`substantial detail for devices constructed to take advan
`tage of such premises. A use for such devices is given as
`one tuned to filter out the normal operating sound of
`machinery as "noise' and to detect the unusual sound of
`a worn or failed component of the machinery. How
`ever, these are illustrations of localized noise, with
`which the adaptive filter type of device is capable of
`
`40
`
`55
`
`65
`
`- 5 -
`
`

`

`4,912,767
`3
`4.
`functioning quite adequately, according to the M.I.T.
`FIG. 1 is an illustration of a conventional noise can
`reference, supra.
`cellation circuit that has become an industry standard.
`U.S. Pat. No. 4,672,674 to Clough et al. discloses a
`FIG. 2 is an illustration of a noise cancellation system
`system utilizing two specially built microphones that
`that embodies the features of the invention.
`have good near field response and poor far field re
`FIG. 3 is a curve for use in describing the operation
`sponse to produce signals with noise components hav
`of the system of the invention.
`ing high correlation. Like the Poikela U.S. Pat. No.
`DETAILED DESCRIPTION OF THE
`4,625,083 above, the outputs from these microphones
`INVENTION
`are connected to a filter to remove frequencies outside
`the range of 300 Hz to between 5 and 8 kHz. The signals
`In FIG. 1 of the drawings, the conventional, or "stan
`then pass to analog-to-digital converters, to micro
`dard', noise cancellation technique is illustrated in the
`processor circuitry having delay and other capability,
`form it was introduced first by Bernard Widrow et al. in
`to achieve weighted-factor-samples for further process
`1975, and is identified generally by the reference nu
`ing. While this prior patent discloses the use of two
`neral 10. As a system, this technique is considered usu
`microphones, it also suggests that a logical extension of 15
`ally as the input for a voice recognition system. Noise
`this use is to use three or more microphones, one for
`cancellation is performed in a substract circuit 11 be
`speech and the outputs of the other microphones being
`tween one signal received directly from one micro
`used to cancel the noise in the signal from the one mi
`phone 12 and the output from a second microphone 13
`crophone.
`after it is passed through an adaptive filter 14. The out
`On the other hand, the present invention takes a dif
`put from the substract circuit 11 is connected directly to
`20
`ferent approach to providing a solution to the problem
`a voice recognition system 15.
`of cancelling distributed noise from a speech signal,
`The outputs from the two microphones 12 and 13
`because tests show that the Adaptive Noise Cancella
`cover the entire audible voice frequency range; for
`tion technique of the prior art degrades in performance
`example, from 100 to 3,200 Hz. The single adaptive
`when the noise is distributed over a region.
`filter 14 in this standard technique, therefore, must be
`25
`capable of performing effectively over the entire audi
`OBJECTS AND SUMMARY OF THE
`ble voice frequency range.
`INVENTION
`The adaptive filter 14 in the conventional technique
`It is a principal object of the present invention to
`must provide compensating amplitude and phase capa
`provide a system for cancelling distributed noise from a
`bilities that vary greatly from one end of the voice
`signal that contains noise-degraded speech.
`frequency range to the other end. In addition, such an
`An important object of the invention is to provide a
`adaptive filter 14 would require a large number of ad
`method for cancelling distributed noise from a voice
`justable elements; for example, 100 tap coefficient ad
`signal.
`justments, or just "taps', all of which leads to problems,
`Another object of the present invention is to provide
`such as:
`-
`a new and improved method and means for cancelling
`(1) The adjustment of a large number of control ele
`distributed noise from a voice signal.
`ments (using the conventional gradient method, or the
`Yet another object of the invention is to provide a
`like) is a very slow process.
`noise cancellation method and system that is effective in
`(2) Efforts to speed up the process of working with a
`a high distributed noise environment.
`large number of control elements can produce other
`40
`Still another object of the invention is to provide an
`problems, such as numerical instability due to trunca
`effective noise cancellation method and system for use
`tion errors, rounding errors, statistical averaging errors,
`with a speech (or voice) recognition system.
`etc.
`A further object of the present invention is to provide
`Noise that is detected by the microphones 12 and 13
`a noise cancellation method and system that will func
`from a single, localized source will be the same "noise'
`45
`tion effectively with standard speech (or voice) sensing
`at each microphone; that is, it will be the same fre
`pickups.
`quency or frequencies, but it will be displaced in time
`A still further object of the present invention is to
`due to differencies in length of the paths it must travel.
`provide a noise cancellation method and system that
`This is the meaning of the term "correlated' as applied
`will function effectively with a standard speech (or
`to the two noise frequencies.
`voice) recognition system in a helicopter environment.
`It is an important function that is performed by the
`Briefly, a method and system that is constructed and
`adaptive filter 14, therefore, when it compensates for
`arranged in accordance with the present invention in
`the differences in time between the two noise frequen
`cludes two sensors, or microphones, located so that a
`cies. It is this compensation between the two signals
`first sensor will detect both voice and noise and a sec
`that results in an effective cancellation when they are
`55
`ond sensor will detect principally only the noise. The
`combined in the substract circuit 11.
`voice picked up at the second sensor is negligible, and
`When the circuit illustrated in FIG. 1, therefore, was
`the noise that is picked up at both sensors is correlated.
`tried with noise that was distributed over a region, it
`The signal output from each sensor is connected to
`was immediately apparent that its performance was
`means to divide each respective signal output into a
`degraded seriously relative to its performance with a
`60
`predetermined number of frequencies. Then, both signal
`single localized noise source. Although much effort has
`outputs are connected to a circuit to cancel effectively
`been devoted to solving this problem in recent years,
`the noise component from the signal output with both
`none has been effective until the present invention.
`voice and noise.
`In FIG. 2 of the drawings there is illustrated a circuit
`arrangement to solve the problem of effectively cancel
`BRIEF DESCRIPTION OF THE DRAWINGS
`ling the noise from voice, or similar information signals,
`The present invention will be described with refer
`sufficiently for a voice recognition system to be useful
`ence to the accompanying drawings, in which:
`reliably. A noise cancellation system in accordance
`
`30
`
`65
`
`O
`
`35
`
`50
`
`- 6 -
`
`

`

`10
`
`15
`
`20
`
`4,912,767
`6
`5
`amplitude and phase differences in the signal detected
`with the invention is sufficiently effective to be useful in
`by the sensor 18. By this means, when each of the di
`every known environment where noise-degraded
`vided signals is combined in each circuit in a group that
`speech renders a voice recognition system ineffective;
`is identified by the reference numeral 22, the noise sig
`such as, for example, in a factory, on a manufacturing
`nal from the sensor 18 is subtracted from the voice-plus
`floor, in large office areas, at airports, etc., etc.
`noise signal from the sensor 17 to provide the substan
`Referring now to FIG. 2, a system that is constructed
`tially noise-free voice signal.
`and arranged in accordance with the principles of the
`While each circuit in the group 22 is indicated as
`invention is identified generally by the reference nu
`being a “subtract' circuit, it will be apparent to one
`meral 16. Two standard sensors 17 and 18, that are
`readily available commercially, such as, for example,
`skilled in the art that other procedures are available for
`obtaining a "difference' action, such as, the signals
`microphones, are located so that the sensor 17 detects
`from the adaptive filters 21 can readily be inverted and
`both voice and noise. It is contemplated that the sensor
`then "added' to the voice-plus-noise signal from the
`17 will be located so that it will detect as much voice as
`possible, even though that signal is degraded by noise.
`narrowband filters 19. Other ways of obtaining a differ
`ence action also will give a similar result.
`The sensor 18, however, is located so that it will
`detect principally noise and very little of the voice.
`The output from each of the individual subtract cir
`cuits in the group 22, as illustrated in FIG. 2 of the
`When used in a pilot's environment, the sensor 17 is
`drawings, is connected to a voice recognition system 23.
`located inside of the pilot's oxygen facemask and the
`With a system 16 constructed and arranged in accor
`sensor 18 is located outside the oxygen facemask. In
`dance with the present invention, the voice recognition
`other environments, where a wire-like headset is used,
`system 23 has no difficulty responding to spoken com
`the sensor 17 is located close to the mouth of a speaker,
`mands in noisy environments and even with noises that
`and the sensor 18 is located also on the headset but as far
`as possible from the mouth of the speaker and is pointed
`are distributed over a wide region.
`in such a way that it detects principally noise. It is im
`FIG. 3 of the drawings illustrates a waveform to
`show this division of the signal from either sensor 17 or
`portant to note, however, that the distance between the
`25
`18 into individual component frequencies. For example,
`two sensors 17 and 18 is quite small, a matter of inches,
`so that the two sensors pick up effectively the same
`the entire curve in FIG. 3 can be an illustration of the
`output signal from either one of the sensors 17 or 18.
`noise but displaced relative to each other a small
`The number '1', identified also by the reference nu
`amount.
`meral 24, is illustrative of a signal that is divided by the
`The signals detected by each of the sensors 17 and 18
`30
`narrowband filter in either group 19 or 20.
`are connected to a suitable device to divide them into a
`number of frequencies. For example, each signal is di
`Similarly, the reference numeral 25 in FIG. 3 identi
`vided into a predetermined number of frequency signals
`fies the number "2" that corresponds to the narrowband
`filter "2" in either the group 19 or 20, in FIG. 2, and the
`having limited bandwidths, and in FIG. 2, the number
`reference numeral 26 identifies the number “15” that
`that is illustrated is 15.
`35
`corresponds to the narrowband filter “15” shown in
`In FIG. 2, the signal output from each of the sensors
`either group 19 or 20, also in FIG. 2. Therefore, in
`17 and 18 is connected to 15 respective narrowband
`accordance with the present invention, the noise cancel
`filters. It is important that the same number used for one
`lation system 16, FIG. 2, divides the total signal that is
`sensor be used for the other. The narrowband filters
`detected by each of the sensors 17 and 18 into a plurality
`that are connected to receive the signal output from the
`40
`of narrow band frequencies each of which covers only
`sensor 17 are in a group that is identified generally by
`a small fraction of the total signal frequency.
`the reference numeral 19, and the narrowband filters
`Of course, this dividing of the total signal into a plu
`that are connected to receive the signal output from the
`rality of smaller frequencies may be accomplished
`sensor 18 are in a group that is identified generally by
`through a variety of hardware component parts. For
`the reference numeral 20.
`45
`example, it is always acceptable to use a plurality of
`Since the usual frequency range for the voice signals
`spans approximately 3 kHz (or 3000 Hertz), by dividing
`individual narrowband filters, but the presently pre
`ferred way the division is accomplished is by means of
`this range into 15 different bandwidths, each one of the
`a computer, because a computer permits the number of
`narrowband filters in the two groups 19 and 20 will be
`the divided frequencies to be changed readily and
`approximately 200 Hertz wide in this example. In tests
`50
`quickly.
`that have been made on this technique, the voice fre
`Tests that have been performed on the invention
`quency has been divided into as many as 25 different
`show that it is possible to obtain a substantially noise
`narrowband frequencies with exceptional results, a
`free signal by dividing the total signal into a predeter
`good range for the number of narrowband filters being
`mined number of individual frequencies before the can
`about 10 to about 25. This range covers most instances
`cellation is attempted. By dividing the noise signal into
`of their use.
`Any particular number of narrowband filters 19 and
`a plurality of narrow bands, then there is less noise in
`20 may be used, or to be more accurate, the signal out
`each narrow band. Now, it has been discovered that it is
`much easier to cancel the noise by this division tech
`put from each sensor 17 and 18 can be divided into any
`nique.
`number of signals. It is important, however, that the
`60
`A system arranged in accordance with the invention
`number of the divisions be the same for the signals from
`has the following unique advantage. Since each individ
`the two sensors 17 and 18, because one of these group of
`divided signals is subtracted from the other to provide a
`ual adaptive filter in the group 21, FIG. 2, must com
`substantially noise-free voice signal.
`pensate for only the frequency in its own narrow band,
`each of the adaptive filters in the group 21 of the inven
`Each of the narrowband filters in the group 20 is
`tion needs only a small number of adjustable elements;
`connected to an adaptive filter in a group that is identi
`fied by the reference numeral 21. Each of the adaptive
`such as, 4 tap coefficients, for example. Now, it will be
`more readily apparent that such an adaptive filter as
`filters in the group 21 functions to compensate for the
`
`55
`
`65
`
`- 7 -
`
`

`

`4,912,767
`7
`8
`the output of each respective sensor into substantially
`needed in a system of the invention can be adjusted
`easily, rapidly and much more accurately.
`the same number of individual frequency bands.
`3. A noise cancellation system as defined in claim 2
`The system of the present invention, therefore, offers
`wherein said two groups of narrowband filters divide
`a solution to a problem that has been heretofore impos
`the output of each respective sensor into approximately
`sible technically. Moreover, published statements by
`fifteen individual frequency bands.
`researchers in this field indicate that they are consider
`4. A noise cancellation system as defined in claim 2
`ing other and materially different arrangements to solve
`wherein said two groups of narrowband filters divide
`the problem of cancelling noise from distributed
`the output of each respective sensor into a frequency
`SOTCS.
`range of approximately 200 Hertz for each narrowband
`Having described the invention completely with ref. O
`filter.
`erence to the presently preferred embodiment, it will be
`5. A noise cancellation system as defined in claim 2
`apparent to those skilled in this art that modifications
`wherein said two groups of narrowband filters divide
`and changes can be made, but it is understood that all
`the output of each respective sensor into approximately
`such modifications and changes that come within the
`10 to 25 equal frequency bands.
`spirit and scope of the claims appended hereto are
`6. A method for cancelling noise from a noise
`degraded voice signal in a distributed noise environ
`within the present invention.
`ment, comprising:
`What is claimed is:
`detecting a noise-degraded voice signal by placing a
`1. A noise cancellation system for increasing the ef
`first sensor so that a voice component of said noise
`20
`fectiveness of a voice recognition device in a distributed
`degraded voice signal will be dominant over a
`noise environment, comprising:
`noise component;
`two sensors to detect voice and noise frequencies, one
`locating a second sensor to detect a signal that is
`of said sensors being located to detect voice plus
`predominately said noise component, and so that a
`noise frequencies, and the other of said two sensors
`difference in phase displacement between said
`25
`being located to detect principally noise frequen
`noise component in said noise-degraded voice sig
`cies;
`nal and said noise component detected by said
`two groups of narrowband filters, one group of nar
`second sensor is small;
`rowband filters being connected to said one of said
`dividing each of said detected signals into a plurality
`sensors, and the other group of said two groups of 30
`of narrowband frequency bands;
`connecting said predominately noise component sig
`narrowband filters being connected to said other of
`nal, after it is divided into said plurality of narrow
`said sensors;
`band frequency bands, to a plurality of adaptive
`a plurality of adaptive filters, one adaptive filter being
`filters equal in number to the number of narrow
`connected to receive an output from each narrow
`band frequency bands of said predominately noise
`35
`band filter in said other group of narrowband fil
`component signal; and
`ters;
`processing said divided narrowband frequency bands
`means including subtract circuit means connected to
`of said first sensor with corresponding predomi
`combine corresponding outputs from said plurality
`nately noise component signals from said adaptive
`of adaptive filters and said one group of narrow
`filters to obtain a voice signal that is substantially
`band filters; and
`noise free.
`voice recognition means connected to receive the
`7. A method of cancelling noise from a noise
`combined corresponding outputs from said means
`degraded voice signal as defined in claim 6 wherein the
`including subtract circuit means to function princi
`step of dividing said detected signals into a plurality of
`pally on voice frequencies.
`narrowband frequency bands includes the step of divid
`45
`2. A noise cancellation system as defined in claim 1
`ing them into approximately 200 Hertz each.
`wherein said two groups of narrowband filters divide
`
`15
`
`50
`
`55
`
`60
`
`65
`
`a
`
`k
`
`k
`
`:
`
`sk
`
`- 8 -
`
`