United States Patent (19)
`Miller
`
`4,589,137
`Patent Number:
`11
`45) Date of Patent: May 13, 1986
`
`54 ELECTRONIC NOISE-REDUCING SYSTEM
`75) Inventor: Harry B. Miller, Niantic, Conn.
`
`73) Assignee: The United States of America as
`represented by the Secretary of the
`Navy, Washington, D.C.
`
`FOREIGN PATENT DOCUMENTS
`59-64994 4/1984 Japan ..................................... 381/92
`Primary Examiner-Gene Z. Rubinson
`Assistant Examiner-L. C. Schroeder
`Attorney, Agent, or Firm-Robert F. Beers; Arthur A.
`McGill; Prithvi C. Lall
`57
`ABSTRACT
`21 Appl. No.: 688,662
`A method and apparatus for reducing noise from a
`near-field noise source present together with signals
`lar.
`from a far-field source. The method uses an adaptive
`Jan. 3, 1985
`22 Filed:
`shaping filter and a summer, in conjunction with a di
`rectional reference sensor and a primary sensor which
`51) Int. Cl." ............................................. H04R 27/00
`have at least 3 COO sensing element therebetween.
`52 U.S. C. se ease so so a on too as a to e use 381/94; 381/71;
`381/92. The directional reference sensor situated between the
`58 Field of Search ..................... 381/94, 71,92, 111,
`near-field noise source and the far-field signal source,
`381/107, 108
`rejects the broad-band signal but accepts the broad
`band noise and feeds this noise into a reference channel
`of the adaptive filter. The primary sensor accepts both
`the far-field signal and near-field noise with equally
`sensitivity. The primary sensor feeds into the primary
`channel of the adaptive filter. The adaptive filter system
`4,025,721 5/1977 Graupe as a so sae assassau Avon - a wood 381/94
`subtracts the noise in the reference channel from the
`3, SEO
`,3.
`4. 354.059 10/1982 Ishigaki.
`381/92 it. in the play t th E.
`4,417,098 11/1983 Chaplin ...
`38/71
`an output having a greatly improved signal-to-noise
`4,420,655 12/1983 Suzuki ......
`381/94
`ratio.
`4,489,441 12/1984 Chaplin.
`... 381/94
`4,536,887 8/1985 Kaneda ................................. 381/94
`
`56)
`
`References Cited
`U.S. PATENT DOCUMENTS
`
`7 Claims, 15 Drawing Figures
`
`PRIMARY INPU
`
`(S+ n)
`e2
`SYSTEM
`D SGNAL PLUS NOIS () OUTPUT
`
`SIGNAL SOURCE
`
`2.
`
`OUTPUT OF
`PRINARY SENSOR
`("3 ALONE)
`
`INPU
`
`PATTERN
`OF THE
`PATTERN OF THE RECEMER
`ReCEIVER OF
`Y THE
`SS S33
`SOURC
`rNOISE
`U Sii IS n
`OED
`(OUTPUT IS ns)
`
`
`
`ADAPTIVE
`FTER
`
`OUTPUT OF
`SIGNAL-FREE
`
`M2
`2
`NOSE SOURCE
`WA
`yo A.4-21
`
`LGE EXHIBIT NO. 1014
`
`- 1 -
`
`Amazon v. Jawbone
`U.S. Patent 8,467,543
`Amazon Ex. 1014
`
`

`

`U.S. Patent May 13, 1986
`
`Sheet 1 of 5
`
`4,589,137
`
`SGNAL SOURCE
`
`SENSOR
`ELEMENTS
`
`F G.
`
`J4, 2
`b A. -2-5
`É 23, 2 3
`E. sz B
`
`2
`
`33-3
`
`2%
`
`O
`
`-O
`
`9 20-
`
`2
`
`K. AM
`3 3rsial Noise NPuré
`%
`(TO ADAPTIVE
`%
`Ble
`CANCELLER)
`s
`5 NSIGNALt NOSE OUT 3
`-50
`a^*(FROGAE, fivec{NSEER) &
`-605 Z2
`f
`f
`
`5-40
`
`
`
`
`
`F G. 4
`
`- 2 -
`
`

`

`U.S. Patent May 13, 1986
`U.S. Patent May 13, 1986
`
`Sheet 2 of 5
`Sheet 2 of 5
`
`4,589,137
`4,589,137
`
`
`
`
`
`
`
`F G. 2
`FIG2
`
`TSkantEYHhieee,
`
`VLEET
`[PYRE
`
`F G. 3
`FIG3
`
`- 3 -
`
`
`
`

`

`U.S. Patent May 13, 1986
`
`Sheet 3 of 5
`
`4,589,137
`
`
`
`SIGNAL SOURCE
`
`PRIMARY INPUT
`
`(S+n.)
`
`62
`
`(S+ n)
`
`REFERENCE
`OUTPUT OF
`PRIMARY SENSOR INPUT:
`("3 ALONE)
`
`
`
`ADAPTIVE
`FILTER
`
`OUTPUT OF
`SSSES5
`(*, *23)
`
`
`
`
`
`NPUT
`
`
`
`|ATTER
`RECEIVER
`EHESSEE
`OSE SOURCE
`SIGNAL SOURCE
`(OUTPUT IS n) OD
`+NOISE SOURCE
`(OUTPUT IS ns)
`N
`WALL- Yo
`
`
`
`-2- NOISE SOURCE
`F G, 6
`
`- 4 -
`
`

`

`U.S. Patent May 13, 1986
`
`Sheet 4 of 5
`
`4,589,137
`
`O- --- -I-I-I I
`
`|
`.
`
`|
`|-
`
`| |
`| |
`|
`|
`l -- - -- |
`
`O
`
`
`
`-
`
`2
`- O
`
`i
`
`>
`H
`2
`H
`(f)
`5. -2O
`
`--
`
`--
`--
`
`L
`2
`
`H s -3OH
`H
`
`f
`SHAKER 72
`70 WALL
`
`FREQUENCY--
`F G. 8
`
`"H
`
`PREAMPS
`MICROPHONE 3
`3>
`LINE ARRAY
`E
`24O
`2-21
`2N-2
`STANDARD a
`ARTIFICAL
`F.
`R>2442
`VOICE
`457
`
`
`
`
`
`FOR PRIMARY CHANNEL
`ASS
`
`R
`FOR REFERENCE LOW PASS
`SEREENESS
`
`
`
`26
`
`03
`as 361
`
`13
`3>
`/
`2>
`D
`
`
`
`
`
`d
`
`OUTPUT 3
`
`OUTPUT
`SUMMER
`
`66
`
`MiSEE
`OUTPU:
`(29
`24 -
`($
`HEADPHONES
`
`64.
`
`
`
`DRECTVTY
`PATTERN
`RECORDER
`
`
`
`CHANNEL B
`CHANNEL. A
`COHERENCE
`NDICATOR
`a
`7.
`
`20
`
`F G. 7
`
`- 5 -
`
`

`

`U.S. Patent May 13, 1986
`
`Sheet 5 of 5
`
`4,589,137
`
`
`
`6 ‘9 | -
`
`
`
`
`
`HELLwd INWO) ?//
`
`
`
`
`
`ENVºld--TvH OHVM (JOH
`
`
`
`
`
`- 6 -
`
`

`

`O
`
`15
`
`35
`
`1.
`
`ELECTRONIC NOISE-REDUCING SYSTEM
`
`STATEMENT OF GOVERNMENT INTEREST
`The invention described herein may be manufactured
`and used by or for the Government of the United States
`of America for governmental purposes without the
`payment of any royalties thereon or therefor.
`BACKGROUND OF THE INVENTION
`1. Field of the Invention
`Subject invention is related to signal processing and
`more particularly to an adaptive filter for cancelling
`noise without affecting the signal and thereby increas
`ing the signal-to-noise ratio.
`2. Description of the Prior Art
`There are many occasions when a microphone is
`required to pick up sound from a talker or loudspeaker
`situated to the right of the microphone, while simulta
`neously there is intense noise radiating from a noise
`20
`source to the left of the microphone. Noise-cancelling
`or noise-reducing devices based on transmission loss,
`such as, for example, sound absorbers placed between
`the microphone and the noisy wall enclosing a machine
`shop, provide one method of reducing the noise (acous
`25
`tically) before it is picked up by the microphone. How
`ever, the sound-absorbing material often occupies a
`large volume, and when the signal bandwidth is ex
`tended to include the low end of the audio bandwidth,
`this volume can be unacceptably large.
`30
`An alternate and more desirable method is to use an
`electronic noise-cancelling or noise-reducing system to
`reduce the transduced noise (now in electrical form)
`after the microphone has picked it up.
`SUMMARY OF THE INVENTION
`An electronic noise cancelling system according to
`the teachings of subject invention includes a reference
`sensor comprising a short endfire line of electroacoustic
`elements, e.g., microphone elements, situated outside a 40
`noisy wall and positioned perpendicular to the wall.
`This sensor, accepting predominantly wall noise, feeds
`into a small adaptive filter system. A second sensor, the
`primary sensor, accepting signal plus noise, also feeds
`into the adaptive filter system. The adaptive filter sys- 45
`ten comprises an adaptive shaping filter or equalizer of
`both phase and amplitude, and a summer. Ideally, the
`system subtracts the pure wall noise from the combina
`tion of signal plus wall noise, leaving pure signal. It
`should be pointed out that simple subtraction accom
`50
`plishes only little. An adaptive shaping filter must be
`inserted into the system to pre-process the wall noise
`prior to subtraction. The system greatly increases the
`signal/noise ratio. It does this by reducing the response
`to broadband wall noise over a wide frequency band, 55
`without reducing the response to the signal source.
`An object of subject invention is to have a noise can
`celling system which does not require a large volume of
`sound-absorbing material.
`Another object of subject invention is to have a noise 60
`canceling system which reduces the noise over a wide
`frequency bandwidth.
`Still another object of subject invention is to have a
`noise-cancelling or noise-reducing system which
`greatly enhances the signal-to-noise ratio for both male 65
`(low frequencies) and female (high frequencies) talkers.
`Other objects, advantages and novel features of the
`invention may become apparent from the following
`
`4,589,137
`2
`detailed description of the invention when considered in
`conjunction with the accompanying drawings wherein:
`BRIEF DESCRIPTION OF THE DRAWINGS
`FIG. 1 is a schematic representation of a noise cancel
`ling system according to the teachings of subject inven
`tion.
`FIGS. 2 and 3 graphically represent the forward
`directivity patterns of the directional sensor and the
`omnidirectional sensor respectively.
`FIG. 4 shows graphically the improvement of the
`signal-to-noise ratio at the output of the electronic
`noise-cancelling system.
`FIG. 5 shows the preferred modification of the direc
`tivity pattern shown above in FIG. 2.
`FIG. 6 is a block diagram of a noise-cancelling system
`built according to the teachings of subject invention.
`FIG. 7 is a more detailed block diagram of the noise
`cancelling or reducing system.
`FIG. 8 is a graphical representation of the frequency
`responses of both the omnidirectional sensor and the
`directional sensor.
`FIG. 9 diagrammatically shows a variant of the line
`microphone where an area-element replaces each of the
`point-elements of FIG. 7.
`FIG. 10A is a representation of an in-plane circular
`dipole including a central point element and a circular
`ring having eight point elements.
`FIG. 10B is a representation of an in-plane circular
`dipole including a central disc element and an annular
`strip encompassing it.
`FIG. 10C is a representation of an in-plane linear
`dipole parallel or nearly parallel to the wall.
`FIG. 11A is a representation of an in-plane circular
`tripole similar to the dipole of FIG. 10B.
`FIG. 11B shows an almost in-plane tripole of rotation
`wherein ring #3 (the central disc) is pulled out of the
`plane by a small distance.
`FIG. 12 shows one of the possible directivity patterns
`obtainable from the tripole of FIG. 11B.
`DESCRIPTION OF THE PREFERRED
`EMBODIMENT
`The method in subject invention requires that two
`different sensors (a reference sensor and a primary sen
`sor) feed into an adaptive filter system. The reference
`sensor supplies a signal-free running (i.e., continuously
`varying with time) wall noise input. This running wall
`noise input, after both its phase and amplitude have
`been manipulated by the adaptive filter, is then sub
`tracted from the primary sensor's running signal-plus
`noise input. Ideally, only the wall noise is reduced at the
`output. The signal at the primary sensor, being incoher
`ent with the wall noise there, is not reduced. Hence the
`signal/noise ratio can be greatly increased.
`One reason for this improvement lies in the nature of
`the adaptive filter system, which is basically an adaptive
`equalizer plus a summer. The adaptive filter system
`using the so called LMS (Least Mean Squares) algo
`rithm has been used for many years. An important part
`of the operation is that this filter system adaptively
`adjusts the frequency response of the reference sample
`(noise alone) in both phase and amplitude so as to equal
`the frequency response of the primary sample's noise
`component while ignoring the primary sample's signal
`component. This is feasible due to the properties of
`
`- 7 -
`
`

`

`4,589,137
`4.
`3
`dipole) displaying low sensitivity to the signal source
`coherence, and the method works when the primary
`and high sensitivity to the wall noise source.
`noise and the reference noise are highly coherent.
`In explaining the operation of the adaptive filter, we
`A second reason for this improvement lies in our
`taking advantage of the art of close-talking micro
`will consider three scenarios:
`(a) If a narrow band of noise (say Af= 10 Hz) cen
`phones. Consider a dipole consisting of two spaced
`tered around 1000 Hz travels through a medium past
`omnidirectional electro acoustic elements, element #2
`two sensors, first past sensor B and then past sensor A,
`and element #1, having the same sensitivity but a rela
`within the correlation time of 0.1 sec, and if response B"
`tive phase of 180 degrees. This dipole displays a figure-8
`is subtracted from response A (response B' being first
`pattern and a 6 db/oct frequency response toward a
`bulk-delayed and then equalized by the adaptive filter),
`far-field source, but displays an almost omnidirectional
`the resultant noise response will equal approximately
`pattern and an almost flat frequency response toward a
`zero, as is desired.
`near-field source if that source is much closer to element
`(b) If, however, sensor A contains not noise but a
`#2 than to element #1. A similar comment applies to a
`1000 Hz signal of equal power (say, value 1), while
`tripole when the near-field source is much closer to
`sensor B contains only the narrow band of noise just
`element #3 than to element #2 or element #1. (Of
`described, and if the adaptation time of the adaptive
`course, the far-field pattern is now a cardioid rather
`filter is made as long as possible (for example, a full 0.1
`than a figure-8.)
`sec), then subtracting response B" from response A" will
`But it should be noted that there is an important dif
`give a number (i.e., amplitude value), varying from zero
`ference in the way the art of close-talking microphones
`to two. The adaptive filter system will not give a resul
`20
`is used in this inventive concept as opposed to the way
`tant approximating zero. Indeed it might just as well be
`the art of close-talking microphones has been conven
`turned off. The reason is that although the narrowband
`tionally used. In the conventional application of the art,
`noise looks on the oscilloscope, like a pure 1000 Hz
`the dipole or tripole microphone is caused to enhance
`signal, it is actually incoherent with the true 1000 Hz
`the desired signal and reduce the noise. In the present
`signal and therefore the two will not perform destruc
`25
`invention, the close-talking dipole or tripole micro
`tive interference. This is similar to Thomas Young's
`phone is caused to do just the opposite: to enhance the
`demonstration that light from two different candles,
`noise and reduce the desired signal. This reverse appli
`being incoherent with each other, will not form a de
`cation of the art of close-talking microphones is an
`structive and constructive interference pattern when
`essential part of the invention.
`allowed to shine through two slits.
`30
`In subject inventive concept, the primary sensor
`(c) Suppose now that sensor A contains both the
`feeds into a primary channel and the reference sensor
`narrow band of noise and the 1000 Hz signal, while
`feeds into a reference channel of the adaptive filter, as
`sensor B contains only the narrow band of noise. Let us
`shown in FIG. 7. Now in the prior art, the primary
`adaptively equalize sensor B's noise and then subtract it
`sensor and the reference sensor are two independent
`from sensor A's signal-plus-noise. If the adaptation time
`entities, physically separated. For example, the primary
`of the adaptive filter is made as long as possible (for
`sensor would be an omnidirectional or a directional
`example, the full correlation time of 0.1 sec), then the
`microphone pointing toward the signal source, and the
`two noises will cancel to approximately zero, since they
`reference sensor would be an accelerometer rigidly
`are highly coherent with each other; whereas the signal
`attached to the wall. This method suffers from two
`will come through practically undiminished, since it is
`drawbacks: the noise in the reference sensor is not suffi
`incoherent with the noise.
`ciently coherent with the noise in the primary sensor;
`Referring to the figures as briefly described above,
`and the total sound (undesired signal plus noise) in the
`FIG. 1 schematically shows wall 10 and line micro
`reference sensor is not sufficiently signal-free.
`phone 12 comprising three microphone elements, with
`In subject inventive concept, the primary sensor and
`microphone element #3 being very close to wall 10 and
`45
`the reference sensor are not physically separated, the
`the remaining microphone elements #1 and #2 being
`primary sensor being a portion of the reference sensor
`situated as shown. Shaker 14 is rigidly attached to wall
`itself, as shown in FIG. 6 and FIG. 7. That is, at least
`10 and is used to set up vibrations in wall 10, The 3-ele
`one element (e.g., #3) of the reference sensor is used
`ment line microphone 12 is perpendicular to wall 10.
`doubly: in the reference sensor and simultaneously in
`The wall noise travels across the line microphone 12 of
`the primary sensor. As a result, the coherence increases
`length d following the laws of the wave equation, and
`between the two sensors. This coherence can be further
`with a 1/r attenuation.
`increased by placing the reference sensor 12 of FIG. 6
`Off to the right as shown in FIG. 1 there is a far-field
`or FIG. 7 as close as possible to the wall noise source,
`signal source 16 radiating toward wall 10. This signal
`and then additionally increased by letting the primary
`source is often a television news announcer. The signal
`from this source is what we are trying to receive at the
`sensor be the element of reference sensor 12 closest to
`line microphone 12 by pulling the signal out of the
`the wall, viz element #3. Element #3 of reference sen
`sor 12 is then not only the primary sensor but is almost
`wall-noise.
`The 3-element line-microphone is arranged to do two
`the entire reference sensor vs. near-field sound (but not,
`things simultaneously: the complete line microphone
`of course, vs. far-field sound). In this way we have
`60
`12, a tripole, acts as the reference sensor. It supplies a
`greatly increased the coherence of the near-field noise
`signal-free wall noise input to the reference channel of
`between the primary sensor and the reference sensor.
`the adaptive filter system. It accomplishes this by means
`We thus have made use of the art of close-talking
`of a directivity pattern which has a very low sensitivity
`microphones in combination with the art of adaptive
`toward the forward half-plane (facing the far-field sig
`filters.
`Also in subject inventive concept the signal-freeness
`nal source) but a high sensitivity toward the back half
`plane (facing the near-field wall-noise source). A simple
`of the reference sensor is improved by using not an
`example of such a directivity pattern is solid curve 20 as
`accelerometer but a line microphone (e.g., a tripole or a
`
`35
`
`65
`
`O
`
`50
`
`55
`
`- 8 -
`
`

`

`10
`
`15
`
`4,589,137
`5
`6
`shown in FIG. 2. We will call this a "backfire cardioid
`dB to 10 dB over the entire 180 forward half-plane.
`pattern' having a single null 22 facing the far-field sig
`Curve 26 is called a perturbed backfire cardioid pattern.
`nal source. The back response is not shown but is essen
`The essentially omnidirectional response of the primary
`tially uniform and of high sensitivity over the back
`sensor, curve 24', is repeated here to show the compara
`half-plane. The back response picks up all the near-field
`tive forward patterns and sensitivities of the two sen
`noise emanating from wall 10. Curve 20 of FIG. 2 is
`sors. The sensitivity in the back half-plane for both
`created by feeding each of the three omnidirectional
`sensors is essentially the same.
`microphone elements 1, 2 and 3 of line microphone 12,
`It should be pointed out that as long as the reference
`after amplification, into its own phase shifter and its
`channel's residual source-signal (undesired) is at least 6
`own attenuator, adjusting magnitude and phase, and
`dB lower than the primary channel's source-signal (de
`then summing in a summer to create a cardioid pattern.
`sired), there is the possibility of increasing the signal/-
`The line microphone 12 is then called a tripole.
`noise ratio by 20 dB or more. That is, there is a nonlin
`Simultaneously a portion of the tripole 12 acts as the
`ear relationship inherent in the functioning of the adapt
`primary sensor. One of the three microphone elements,
`ive filter, which allows a S/N improvement far greater
`i.e., electroacoustic elements (having, of course, a free
`than is possible from a directional sensor without an
`field omnidirectional pattern) feeds signal-plus-noise
`adaptive filter.
`directly into the primary channel of the adaptive filter
`However, a major limitation to increasing the sig
`system. Note that this microphone element is contribut
`nal/noise ratio is the imperfect coherence between the
`ing simultaneously to both the reference channel and
`noise at the reference channel input and the noise at the
`the primary channel. The forward half-plane directional
`primary channel input. A coherence of 90 percent is
`20
`response of the primary sensor is shown as curve 24 in
`generally required to achieve a 10 dB increase in sig
`nal/noise ratio. A coherence of 99 percent is generally
`FIG. 3. This curve is also shown as dotted curve 24' in
`FIG. 2. The response is nearly uniform and of high
`required to achieve a 20 dB increase in signal/noise
`sensitivity over most of the forward half-plane. The
`ratio. Furthermore, since every piece of information in
`back response is not shown here but is essentially uni
`the reference channel that is coherent with information
`25
`form and of high sensitivity over the back half-plane,
`in the primary channel will be subtracted, any residual
`and nearly identical with the back response of the back
`source-signal in the reference channel will also be sub
`fire cardioid pattern of FIG. 2, thus allowing a direct
`tracted from the source-signal in the primary channel.
`comparison between the reference sensor response
`This subtraction will therefore reduce the expected
`(solid curve 20) and the primary sensor response (dotted
`improvement in signal/noise ratio to less than the 10 dB
`30
`curve 24). In the angular sector 330' to 30' of FIG. 2
`and 20 dB values mentioned. Hence, the residual
`the reference sensor could be considered signal-free
`source-signal in the "signal-free' reference channel
`because its sensitivity is at least 8 dB lower than the
`should be at least 6 dB lower than the source-signal in
`primary sensor's sensitivity.
`the primary channel. A greater improvement will take
`The reference channel's adaptively adjusted noise is
`place if the residual source-signal is lower by 8 dB or 10
`subtracted from the primary channel's signal-plus-noise,
`dB.
`leaving a signal having an improved S/N ratio. This is
`FIG. 6 shows the essential components needed for a
`shown in FIG. 4 for a single frequency, where the S/N
`wall-noise-cancelling system. The reference sensor or
`ratio at the output of the adaptive filter is 17 dB higher
`line microphone 12 in the figure is a 3-element sensor, or
`than that at the input. Note that the adaptive filter sys
`tripole, situated perpendicular to the wall. It is also
`40
`possible to use a 2-element sensor, or dipole, situated
`tem has reduced the noise over a broadbandwidth.
`The upper curve 30 of FIG. 4 shows the spectral
`perpendicular to the wall. Also, it is possible to situate
`response from wall 10 driven by random noise from
`the tripole or the dipole nearly parallel to the wall, the
`shaker 14. Superimposed on curve 30 is the spectrum of
`trade-off being a less bulky mechanical arrangement
`a single-frequency signal from a far-field source 16 hav
`versus a reduced improvement in signal/noise ratio.
`45
`ing a spectral level 36 about the same as the noise spec
`As can be seen in FIGS. 6 and 7, the reference sensor
`tral level 33. The S/N ratio is thus about zero dB. The
`12 must always use more than one omnidirectional mi
`sum of these two spectra provides the input to the pri
`crophone element, whereas the primary sensor need use
`mary channel of the adaptive filter system.
`only one, e.g., #3. However, the system also works well
`The lower curve 32 of FIG. 4 shows the spectral
`if the primary sensor is #2 alone or #1 alone or even a
`50
`response output from the adaptive filter system. The
`combination of #1 plus #2 plus #3 if the phases and
`noise spectral response has been reduced over a broad
`amplitudes are such that the forward pattern 24 is essen
`bandwidth, whereas the signal spectral response comes
`tially omnidirectional. Each of the microphone or elec
`through the system practically untouched as spectral
`troacoustic elements #1, #2 and #3 of line microphone
`level 36. At the signal frequency, the S/N ratio is in
`12 feeds into its respective preamp 40, 42 or 44 of FIG.
`55
`creased by 17 dB (note reduced noise spectral level 38).
`7 and thence into its respective phase shifter 46, 48 or 50
`If now we replace the single-frequency signal with a
`and buffer amplifier 52, 54 or 56.
`broadband speech signal, and retain the broadband
`It is highly advantageous to let the reference sensor
`noise, a signal-to-noise improvement will occur over
`12 and the primary sensor have at least one microphone
`the whole speech band. The average S/N improvement
`element in common. Thus, in FIGS. 6 and 7, element #3
`over this band will of course be less than that for the
`is used twice, i.e., it is the common element. This en
`single frequency case of FIG. 4.
`sures high coherence between the noise input in the
`FIG. 5 shows a more sophisticated backfire cardioid
`reference channel and the noise input in the primary
`pattern, curve 26, than that of curve 20 of FIG. 2
`channel.
`(which had only a single null and was signal-free over
`FIG. 7 shows also a more detailed layout of the com
`65
`only about a 60° angle out of the entire 180° of the
`ponents used, including monitoring devices. Observe
`forward half-plane). In FIG. 5, curve 26, there are two
`that #3 microphone element or electroacoustic element
`nulls, 28 and 29, and an overall attenuation of about 8
`is used simultaneously in the reference channel 60 and in
`
`35
`
`- 9 -
`
`

`

`SS.
`
`-
`
`4,589,137
`8
`7
`oid pattern, e.g., by using six omnidirectional micro
`the primary channel 62 of adaptive filter 64. When two
`phone elements in a line instead of the three electro
`sets of phase shifters and two summing networks are
`acoustic elements of line microphone 12. This reduces
`used, it is even possible to create a 3-element backfire
`the response of the backfire cardioid lobes by an even
`cardioid sensor for the reference channel, and simulta
`greater amount than the 8 dB to 10 dB shown in curve
`neously a 3-element forward cardioid sensor for the
`26 of FIG. 5. A decision to use higher-order patterns is
`primary channel, using the same set of three elements.
`based on a tradeoff of financial cost versus signal-free
`The noise-coherence between the two channels is high
`because the same noise excites the same three elements
`Returning to the discussion of flat frequency response
`for both inputs (reference and primary). However, it is
`and 6 dB/octave slopes, we see in FIG. 8, curve 74, the
`sometimes considered undesirable to use a forward
`relatively flat frequency response of a single omni
`cardioid pattern for the primary input (which deter
`directional microphone element located close to the
`mines the system output 66) because the frequency re
`sponse which goes with any cardioid pattern has a 6
`wall.
`The non-flat far-field frequency response of the back
`dB/ octave slope. This means that at low frequencies,
`fire cardioid sensor is shown in curve 76 of FIG. 8. At
`e.g., where d=W/16, even the maximum pattern sensi
`the chosen signal frequency, for which the cardioid
`tivity is very low (down from its highest value by 14
`pattern was optimized, a directional null exists in the
`dB) and that therefore the far-field signal response will
`pattern. The relative orientation of sensor 12 and wall
`be much weaker than is desirable. Hence, it is then
`preferable to use for the primary input only a single
`10 was such as to let the directional null face the stan
`dard artificial voice 58 of FIG. 7. With a fixed setting of
`microphone element, having an omnidirectional pat
`20
`the three phase shifters of FIG. 7, and a fixed angular
`tern. This single microphone will have a relatively flat
`orientation of sensor and wall, there is only a single,
`frequency response over the whole frequency band
`rather sharp, null region in the frequency response
`width.
`(curve 76 of FIG. 8.) The useful bandwidth of the null
`The backfire cardioid pattern used for the reference
`region is about a half-octave. This is the region over
`input will inherently also have a far-field frequency
`25
`which the response is down at least 8 dB compared to
`response whose envelope has a 6 dB/octave slope. This
`is shown in FIG.8. This means that at low frequencies
`the omnidirectional curve 74.
`At frequencies above and below the null frequency,
`where d=A/16, the far-field maximum pattern sensitiv
`the frequency response somewhat resembles that of a
`ity of the cardioid (pointing now toward the back half
`normal forward-looking cardioid system. The reason is
`plane) is down 14 dB from its highest value. However,
`30
`that the fixed phase angles selected to form the backfire
`since we are in a near-field situation, the -14 dB value
`cardioid pattern are optimum only over about a half
`does not hold. And in fact, because of the characteris
`octave. Beyond this null region a new setting of phase
`tics of close-talking microphones, the reduction in sensi
`angles is required. Thus if a bandwidth of, say, a decade
`tivity is approximately zero. Thus a backfire cardioid
`or about 3 octaves is to be covered, the necessary
`sensor can pick up a strong wall-noise sample to feed
`modifications can be accomplished in any of several
`into the reference channel. In addition, the sample will
`ways. One way is to divide the frequency bandwidth
`- be quite signal-free since the forward sensitivity of the
`shown in FIG. 8 into, say, seven frequency bins (using
`sensor is very low.
`contiguous half-octave bandpass filters), all in parallel.
`It should be noted that for dish/16 the backfire car
`Each bin contains a phase shifter and amplifier which
`idioid pattern (from a tripole or dipole perpendicular to
`provide the optimum phase value and amplitude value
`the wall) can be replaced with a simple figure-8 pattern
`to form a backfire cardioid for that frequency region.
`(from a dipole perpendicular to the wall), since the 14
`When the contents of the seven bins are summed and
`dB or more drop in far-field sensitivity and the 0 dB
`fed into the reference channel of the adaptive filter, the
`drop in near-field sensitivity together assure an accept
`resulting frequency response is the same as if from a
`able signal-free reference sensor.
`45
`broad band-elimination filter, with the null covering a
`It should also be noted that all the distinctive features
`complete decade.
`of the response of the reference channel's sensor, such
`FIGS. 1, 6 and 7 depict the three microphone or
`as, e.g., a frequency response with a 6 dB/octave slope,
`electroacoustic elements as three point-sensors. Some
`are irrelevant to the system output 66 (FIGS. 6 and 7)
`times it is desirable to use area microphone elements in
`because the reference channel acts merely as a tempo
`50
`place of the point microphone elements. FIG. 9 shows
`rary scaffolding. The channel that determines the input
`a variant 80 of the line microphone 12 where area mi
`to our ultimate receiving device, the headphone pair 74,
`crophone elements 1', 2',3' replace the point micro
`is the primary channel. That is, the information that
`phone elements 1, 2, and 3 of FIG. 7.
`goes to the headphones 74 comes from the system out
`Instead of three microphone elements positioned
`put, which itself is determined only by the primary
`55
`perpendicular to the wall (a volumetric sensor) for cres
`channel. And if the primary channel's sensor is a single
`ating the reference sensor, it is sometimes desirable to
`omnidirectional element, then the system output fre
`use a planar sensor as shown in FIG, 10A. An in-plane
`quency response will be relatively flat.
`dipole-of-rotation may be approximated, using a ring 90
`FIG, 7 also shows that the cardioid patterns can be
`of acoustically sensitive material surrounding a central
`examined with the help of a pattern recorder 70 inserted
`point-element 92. Ring 90 can consist either of discrete
`ahead of the adaptive filter 64. The coherence between
`elements such as 94, 96, 98, 100, 102,104,106 and 108 as
`the two channels can be monitored by a coherence
`shown in FIG. 10A, or of a continuous strip, 110, as
`indicator 72. The system output going to the head
`shown in FIG. 10B. The basic free-field pattern in each
`phones 74 can be examined with the help of a spectrum
`case is a toroid, parallel to the wall. An in-plane linear
`analyzer 68.
`dipole 112, may also be used, as shown in FIG. 10C.
`It should be noted here that the signal-freeness of the
`The basic free-field pattern is a dumbbell, nearly paral
`reference sensor, as shown by curve 26 of FIG. 5, can
`lel to the wall. An in-plane tripole of rotation 114 can
`be improved by creating a higher-order backfire cardi
`
`35
`
`O
`
`15
`
`65
`
`- 10 -
`
`

`

`4,589,137
`10
`9
`