`
`United States Patent
`Elk0
`
`(10) Patent No.:
`(45) Date of Patent:
`
`US 7,171,008 B2
`Jan. 30, 2007
`
`US007 171008B2
`
`(54) REDUCING NOISE IN AUDIOSYSTEMS
`(75) Inventor: Gary W. Elko, Summit, NJ (US)
`
`(73) Assignee: MH Acoustics, LLC, Summit, NJ (US)
`
`( c ) Notice:
`
`Subject tO any disclaimer, the term of this
`patent is extended or adjusted under 35
`U.S.C. 154(b) by 852 days.
`
`(21) Appl. No.: 10/193,825
`
`(22) Filed:
`
`Jul. 12, 2002
`O
`O
`Prior Publication Data
`US 2003/O147538 A1
`Aug. 7, 2003
`
`(65)
`
`Related U.S. Application Data
`(60) Provisional application No. 60/354,650, filed on Feb.
`5, 2002.
`s
`(51) Int. Cl.
`(2006.01)
`H04R 3/00
`(2006.01)
`H04R IM02
`381A92: 381A91: 381/122
`(52) U.S. Cl
`h
`s
`s 381/97
`58 Fi ld f C- - - - - - ficati - - - - - -s
`(58) Field o 38 s s , sgr 1 7318,313.3 2O 32 1.
`74s 1 4 4s. 1-3 - 1
`381 ?o 4.1 3 13
`See application file for complete search histo -
`pp
`p
`ry.
`References Cited
`U.S. PATENT DOCUMENTS
`
`(56)
`
`5/1996 Baumhauer, Jr. et al. ..... 381/92
`5,515,445 A
`2f1997 Kellermann
`5,602,962 A
`5,687.241 A 11/1997 Ludvigsen ................. 381,684
`5,878,146 A
`3, 1999 Andersen ...
`... 381,312
`6,272,229 B1
`8, 2001 Baekgaard
`... 381 313
`6,292,571 B1
`9/2001 Sursen .............
`... 381,312
`6,339,647 B1
`1/2002 Andersen et al. ........... 381,312
`
`FOREIGN PATENT DOCUMENTS
`O6303.689
`10, 1994
`2001 124621
`5, 2001
`WO 95 16259
`6, 1995
`
`JP
`JP
`WO
`
`Primary Examiner Vivian Chin
`Assistant Examiner—Douglas Suthers
`(74) Attorney, Agent, or Firm—Steve Mendelsohn
`
`(57)
`
`ABSTRACT
`
`Two or more microphones receive acoustic signals and
`generate audio signals that are processed to determine what
`portion of the audio signals result from (i) incoherence
`between the audio signals and/or (ii) audio-signal Sources
`having propagation speeds different from the acoustic sig
`nals. The audio signals are filtered to reduce that portion of
`one or more of the audio signals. The present invention can
`be used to reduce turbulent wind-noise resulting from wind
`or other airjets blowing across the microphones. Time
`dependent phase and amplitude differences between the
`microphones can be compensated for based on measure
`ments made in parallel with routine audio system process
`1ng.
`
`5,325,872 A
`
`7, 1994 Westermann ............... 128/897
`
`55 Claims, 13 Drawing Sheets
`
`
`
`
`
`
`12O3
`
`CONTROL
`
`AMPLETUDEA
`PHASE
`FILER
`
`
`
`
`
`12OO
`
`1204
`
`SIGNAL
`PROCESSOR
`
`CONTROL
`
`NOISE
`FILTER
`
`
`
`OUTPUT
`
`1208
`
`12O6
`
`Page 1 of 27
`
`GOOGLE EXHIBIT 1010
`
`
`
`U.S. Patent
`U.S. Patent
`
`Jan. 30, 2007
`Jan. 30, 2007
`
`Sheet 1 of 13
`Sheet 1 of 13
`
`US 7,171,008 B2
`US 7,171,008 B2
`
`Fig. 1
`Fig.
`1
`
`cD
`
`Page 2 of 27
`
`Page 2 of 27
`
`
`
`U.S. Patent
`
`Jan. 30, 2007
`
`Sheet 2 of 13
`
`US 7,171,008 B2
`
`Fig. 2
`
`
`
`f Rig C ue as t' (H )
`
`Page 3 of 27
`
`
`
`U.S. Patent
`
`Jan. 30, 2007
`
`Sheet 3 of 13
`
`US 7,171,008 B2
`
`Fig. 3
`
`s
`
`
`
`frequeact (H2)
`
`Page 4 of 27
`
`
`
`U.S. Patent
`U.S. Patent
`
`Jan. 30, 2007
`Jan. 30, 2007
`
`Sheet 4 of 13
`Sheet 4 of 13
`
`US 7,171,008 B2
`US 7,171,008 B2
`
`Fig. 4
`Fig.
`
`4
`
`
`
`Page 5 of 27
`
`Page 5 of 27
`
`
`
`U.S. Patent
`U.S. Patent
`
`Jan. 30, 2007
`Jan. 30, 2007
`
`Sheet S of 13
`Sheet 5 of 13
`
`US 7,171,008 B2
`US 7,171,008 B2
`
`Fig. 5
`Fig. 5
`
`u(n)+ u(n)
`(A+ vQ(N)
`
`
`
`
`
`Page 6 of 27
`
`Page 6 of 27
`
`
`
`U.S. Patent
`
`Jan. 30, 2007
`
`Sheet 6 of 13
`
`US 7,171,008 B2
`
`Fig. 6
`
`0
`
`-5
`
`-
`
`1N
`(Y) - 15
`-r
`Nu
`2 -20
`
`C)
`1.
`C
`W -25
`AC
`CA
`Q
`5 -3)
`u/
`
`-35
`
`u
`
`-40
`0.
`
`0.
`
`0.2
`0.3
`(),4
`0,5
`0.
`(),
`MA(n) ?upe-SQw Ago (d. He ReA)ce
`
`0.8
`
`0.9
`
`Page 7 of 27
`
`
`
`U.S. Patent
`
`Jan. 30, 2007
`
`Sheet 7 of 13
`
`US 7,171,008 B2
`
`Fig. 7
`
`
`
`s
`
`The (Sec)
`
`Page 8 of 27
`
`
`
`U.S. Patent
`
`Jan. 30, 2007
`
`Sheet 8 of 13
`
`US 7,171,008 B2
`
`Fig. 8
`
`
`
`- - - -e- is
`
`y
`
`rr
`
`vs tra to
`
`a sers w
`
`re r
`
`Fragaray (Hz
`
`Page 9 of 27
`
`
`
`U.S. Patent
`
`Jan. 30, 2007
`
`Sheet 9 of 13
`
`US 7,171,008 B2
`
`FIG. 9
`
`904
`
`9 OO
`
`
`
`SIGNAL
`PROCESSOR
`
`
`
`
`
`CONTROL
`
`
`
`NOISE
`FILTER
`
`OUTPUT
`
`908
`
`906
`
`Page 10 of 27
`
`
`
`U.S. Patent
`
`Jan. 30, 2007
`
`Sheet 10 of 13
`
`US 7,171,008 B2
`
`690-
`
`Fig. 10
`
`ld62.
`
`
`
`
`
`
`
`AID and FFT
`
`
`
`A/D and FFT
`
`dO2
`
`869th
`
`Compute sum and
`difference powers
`of input signals
`
`Weight desired
`signals
`to attenuate high
`WaVettaber
`
`
`
`Equalize, IFFT
`overlap-add
`and A
`
`Page 11 of 27
`
`
`
`U.S. Patent
`
`Jan. 30, 2007
`
`Sheet 11 of 13
`
`US 7,171,008 B2
`
`Fig. 11
`
`1703
`differential
`microphone
`
`) Oy
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Compute powers
`of input
`signals
`
`Weight desired
`signals
`to attenuate high
`Wavenumber
`signals
`
`
`
`Equalize, FFT
`overlap-add
`and DIA
`
`pressure
`microphone
`O 2.
`
`19 t
`
`Page 12 of 27
`
`
`
`U.S. Patent
`
`Jan. 30, 2007
`
`Sheet 12 of 13
`
`US 7,171,008 B2
`
`FIG. 2
`
`12OO
`
`
`
`
`
`AMPLITUDE/
`PHASE
`FILTER
`
`
`
`
`
`
`
`
`
`CONTROL
`
`
`
`
`
`12O4
`
`SIGNAL
`PROCESSOR
`
`CONTROL
`
`
`
`NOISE
`FILTER
`
`OUTPUT
`
`12O8
`
`12O6
`
`Page 13 of 27
`
`
`
`U.S. Patent
`
`Jan. 30, 2007
`
`Sheet 13 of 13
`
`US 7,171,008 B2
`
`Fig. 13
`
`
`
`ift
`
`
`
`Sarticle
`
`i in 1302.
`
`claires
`
`3O.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`channel data
`
`Route plaf - l3) O
`suf a
`
`(ifference towers
`
`and ccherence
`
`13 la
`
`
`
`Cempute front ard
`rear power ?atos
`using fixed or
`acastle
`tea forming
`
`13th
`
`
`
`
`
`
`
`Corpuste aptude
`and prases
`calibatic
`
`
`
`s
`
`Tubulent
`wind-use
`
`yes
`
`Suppress tutulent
`Wrd-nose using
`soard
`suppression
`
`Syntesia using
`overlapiard,
`equalizeard
`apply gas
`
`
`
`also
`
`Page 14 of 27
`
`
`
`1.
`REDUCING NOISE IN AUDIO SYSTEMS
`
`US 7,171,008 B2
`
`CROSS-REFERENCE TO RELATED
`APPLICATIONS
`
`This application claims the benefit of the filing date of
`U.S. provisional application No. 60/354,650, filed on Feb. 2,
`2002.
`
`BACKGROUND OF THE INVENTION
`
`10
`
`1. Field of the Invention
`The present invention relates to acoustics, and, in par
`ticular, to techniques for reducing noise, Such as wind noise,
`generated by turbulent airflow over microphones.
`2. Description of the Related Art
`For many years, wind-noise sensitivity of microphones
`has been a major problem for outdoor recordings. A related
`problem is the susceptibility of microphones to the speech
`jet, i.e., the flow of air from the talker's mouth. Recording
`studios typically rely on special windscreen socks that either
`cover the microphone or are placed between the mouth and
`the microphone. For outdoor recording situations where
`wind noise is an issue, microphones are typically shielded by
`acoustically transparent foam or thick fuZZy materials. The
`purpose of these windscreens is to reduce—or even elimi
`nate the airflow over the active microphone element to
`reduce—or even eliminate—noise associated with that air
`flow that would otherwise appear in the audio signal gen
`erated by the microphone, while allowing the desired acous
`tic signal to pass without significant modification to the
`microphone.
`
`SUMMARY OF THE INVENTION
`
`15
`
`25
`
`30
`
`35
`
`The present invention is related to signal processing
`techniques that attenuate noise, Such as turbulent wind
`noise, in audio signals without necessarily relying on the
`mechanical windscreens of the prior art. In particular,
`according to certain embodiments of the present invention,
`two or more microphones generate audio signals that are
`used to determine the portion of pickup signal that is due to
`wind-induced noise. These embodiments exploit the notion
`that wind-noise signals are caused by convective airflow
`whose speed of propagation is much less than that of the
`desired acoustic signals. As a result, the difference in the
`output powers of Summed and Subtracted signals of closely
`spaced microphones can be used to estimate the ratio of
`turbulent convective wind-noise propagation relative to
`acoustic propagation. Since convective turbulence coher
`ence diminishes quickly with distance, Subtracted signals
`between microphones are of similar power to Summed
`signals. However, signals propagating at acoustic speeds
`will result in relatively large difference in the summed and
`Subtracted signal powers. This property is utilized to drive a
`time-varying Suppression filter that is tailored to reduce
`signals that have much lower propagation speeds and/or a
`rapid loss in signal coherence as a function of distance, e.g.,
`noise resulting from relatively slow airflow.
`According to one embodiment, the present invention is a
`method and an audio system for processing audio signals
`generated by two or more microphones receiving acoustic
`signals. A signal processor determines a portion of the audio
`signals resulting from one or more of (i) incoherence
`between the audio signals and (ii) one or more audio-signal
`Sources having propagation speeds different from the acous
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`2
`tic signals. A filter filters at least one of the audio signals to
`reduce the determined portion.
`According to another embodiment, the present invention
`is a consumer device comprising (a) two or more micro
`phones configured to receive acoustic signals and to gener
`ate audio signals; (b) a signal processor configured to
`determine a portion of the audio signals resulting from one
`or more of (i) incoherence between the audio signals and (ii)
`one or more audio-signal sources having propagation speeds
`different from the acoustic signals; and (c) a filter configured
`to filter at least one of the audio signals to reduce the
`determined portion.
`According to yet another embodiment, the present inven
`tion is a method and an audio system for processing audio
`signals generated in response to a Sound field by at least two
`microphones of an audio system. A filter filters the audio
`signals to compensate for a phase difference between the at
`least two microphones. A signal processor (1) generates a
`revised phase difference between the at least two micro
`phones based on the audio signals and (2) updates, based on
`the revised phase difference, at least one calibration param
`eter used by the filter.
`In yet another embodiment, the present invention is a
`consumer device comprising (a) at least two microphones;
`(b) a filter configured to filter audio signals generated in
`response to a sound field by the at least two microphones to
`compensate for a phase difference between the at least two
`microphones; and (c) a signal processor configured to (1)
`generate a revised phase difference between the at least two
`microphones based on the audio signals; and (2) update,
`based on the revised phase difference, at least one calibration
`parameter used by the filter.
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`Other aspects, features, and advantages of the present
`invention will become more fully apparent from the follow
`ing detailed description, the appended claims, and the
`accompanying drawings in which like reference numerals
`identify similar or identical elements.
`FIG. 1 shows a diagram of a first-order microphone
`composed of two Zero-order microphones;
`FIG. 2 shows a graph of Corcos model coherence as a
`function of frequency for 2-cm microphone spacing and a
`convective speed of 5 m/s:
`FIG. 3 shows a graph of the difference-to-sum power
`ratios for acoustic and turbulent signals as a function of
`frequency for 2-cm microphone spacing and a convective
`speed of 5 m/s;
`FIG. 4 illustrates noise Suppression using a single-channel
`Wiener filter;
`FIG. 5 illustrates a single-input/single-output noise Sup
`pression system that is essentially equivalent to a system
`having an array with two closely spaced omnidirectional
`microphones;
`FIG. 6 shows the amount of noise suppression that is
`applied by the system of FIG. 5 as a function of coherence
`between the two microphone signals;
`FIG. 7 shows a graph of the output signal for a single
`microphone before and after processing to reject turbulence
`using propagating acoustic gain settings:
`FIG.8 shows a graph of the spatial coherence function for
`a diffuse propagating acoustic field for 2-cm spaced micro
`phones, shown compared with the Corcos model coherence
`of FIG. 2 and for a single planewave;
`FIG. 9 shows a block diagram of an audio system,
`according to one embodiment of the present invention;
`
`Page 15 of 27
`
`
`
`3
`FIG. 10 shows a block diagram of turbulent wind-noise
`attenuation processing using two closely spaced, pressure
`(omnidirectional) microphones, according to one implemen
`tation of the audio system of FIG. 9;
`FIG. 11 shows a block diagram of turbulent wind-noise
`attenuation processing using a directional microphone and a
`pressure (omnidirectional) microphone, according to an
`alternative implementation of the audio system of FIG. 9;
`FIG. 12 shows a block diagram of an audio system having
`two omnidirectional microphones, according to an alterna
`tive embodiment of the present invention; and
`FIG. 13 shows a flowchart of the processing of the audio
`system of FIG. 12, according to one embodiment of the
`present invention.
`DETAILED DESCRIPTION
`
`10
`
`15
`
`Differential Microphone Arrays
`A differential microphone array is a configuration of two
`or more audio transducers or sensors (e.g., microphones)
`whose audio output signals are combined to provide one or
`more array output signals. As used in this specification, the
`term “first-order applies to any microphone array whose
`sensitivity is proportional to the first spatial derivative of the
`acoustic pressure field. The term “n'-order” is used for
`microphone arrays that have a response that is proportional
`to a linear combination of the spatial derivatives up to and
`including n. Typically, differential microphone arrays com
`bine the outputs of closely spaced transducers in an alter
`nating sign fashion.
`Although realizable differential arrays only approximate
`the true acoustic pressure differentials, the equations for the
`general-order spatial differentials provide significant insight
`into the operation of these systems. To begin, the case for an
`acoustic planewave propagating with wave vector k is
`examined. The acoustic pressure field for the planewave
`case can be written according to Equation (1) as follows:
`
`25
`
`30
`
`35
`
`p(k, r, t) = Peek)
`
`(1)
`
`40
`
`where P is the planewave amplitude, k is the acoustic wave
`vector, r is the position vector relative to the selected origin,
`and () is the angular frequency of the planewave. Dropping
`the time dependence and taking the n'-order spatial deriva
`tive yields Equation (2) as follows:
`
`45
`
`50
`
`where 0 is the angle between the wavevector k and the
`position vector r, r-r, and k-k-21/W, where w is the
`acoustic wavelength. The planewave solution is valid for the
`response to sources that are “far from the microphone array,
`where “far means distances that are many times the square
`of the relevant source dimension divided by the acoustic
`wavelength. The frequency response of a differential micro
`phone is a high-pass system with a slope of 6n dB per
`octave. In general, to realize an array that is sensitive to the
`n" derivative of the incident acoustic pressure field, m
`p'-order transducers are required, where, m+p-1=n. For
`example, a first-order differential microphone requires two
`Zero-order sensors (e.g., two pressure-sensing micro
`phones).
`
`55
`
`60
`
`65
`
`US 7,171,008 B2
`
`4
`For a planewave with amplitude Po and wavenumber k
`incident on a two-element differential array, as shown in
`FIG. 1, the output can be written according to Equation (3)
`as follows:
`T(ke)=P(1-edia cose)
`(3)
`where d is the inter-element spacing and the Subscript
`indicates a first-order differential array. If it is now assumed
`that the spacing d is much smaller than the acoustic wave
`length, Equation (3) can be rewritten as Equation (4) as
`follows:
`
`(4)
`|T(k,0) is Pkd cos0
`The case where a delay is introduced between these two
`Zero-order sensors is now examined. For a planewave inci
`dent on this new array, the output can be written according
`to Equation (5) as follows:
`T(o,0)=P(1-e jet cose)
`(5)
`where T is equal to the delay applied to the signal from one
`sensor, and the Substitution k (Dfc has been made, where c
`is the speed of Sound. If a small spacing is again assumed
`(kd.<<It and (ot-at), then Equation (5) can be written as
`Equation (6) as follows:
`|T(a),0)-spo(t+d/c cos0)
`
`(6)
`
`One thing to notice about Equation (6) is that the first-order
`array has first-order high-pass frequency dependence. The
`term in the parentheses in Equation (6) contains the array
`directional response.
`Since n'-order differential transducers have responses
`that are proportional to the n' power of the wavenumber,
`these transducers are very sensitive to high wavenumber
`acoustic propagation. One acoustic field that has high
`wavenumber acoustic propagation is in turbulent fluid flow
`where the convective velocity is much less than the speed of
`Sound. As a result, prior-art differential microphones have
`typically required careful shielding to minimize the hyper
`sensitivity to wind turbulence.
`Turbulent Wind-Noise Models
`The subject of modeling turbulent fluid flow has been an
`active area of research for many decades. Most of the
`research has been in underwater acoustics for military
`applications. With the rapid growth of commercial airline
`carriers, there has been a great amount of work related to
`turbulent flow excitation of aircraft fuselage components.
`Due to the complexity of the equations of motion describing
`turbulent fluid flow, only rough approximations and rela
`tively simple statistical models have been Suggested to
`describe this complex chaotic fluid flow. One model that
`describes the coherence of the pressure fluctuations in a
`turbulent boundary layer along the plane of flow is described
`in G. M. Corcos, The structure of the turbulent pressure field
`in boundary layer flows, J. Fluid Mech., 18: pp. 353–378,
`1964, the teachings of which are incorporated herein by
`reference. Although this model was developed for turbulent
`pressure fluctuation over a rigid half-plane, the simple
`Corcos model can be used to express the amount of spatial
`filtering of the turbulent jet from a talker. Thus, this model
`is used to predict the spatial coherence of the pressure
`fluctuation turbulence for both speech jets as well as free
`space turbulence.
`The spatial characteristics of the pressure fluctuations can
`be expressed by the space-frequency cross-spectrum func
`tion G according to Equation (7) as follows:
`
`Page 16 of 27
`
`
`
`US 7,171,008 B2
`
`Gpp (th, (0) = | Ries(e. t)et dr
`
`7
`(7)
`
`where R is the spatial cross-correlation function between the
`two microphone signals, () is the angular frequency, and
`is the general displacement variable which is directly related
`to the distance between measurement points. The coherence
`function Y is defined as the normalized cross-spectrum by the
`auto power-spectrum of the two channels according to
`Equation (8) as follows:
`
`10
`
`Gpp
`(a)12
`p2p2
`
`G
`plp2
`
`15
`
`(8)
`
`It is known that large-scale components of the acoustic
`pressure field lose coherence slowly during the convection
`with free-stream velocity U, while the small-scale compo
`nents lose coherence in distances proportional to their wave
`lengths. Corcos assumed that the stream-wise coherence
`decays spatially as a function of the similarity variable
`(or/U, where U is the convective speed and is typically
`related to the free-stream velocity U as U-0.8U. The
`Corcos model can be mathematically stated by Equation (9)
`as follows:
`
`(9)
`
`6
`limiting approximation is important to the present inven
`tion's detection and resulting Suppression of signals that are
`identified as turbulent.
`Single-Channel Wiener Filter
`It was shown in the previous section that one way to
`detect turbulent energy flow over a pair of closely-spaced
`microphones is to compare the scalar sum and difference
`signal power levels. In this section, it is shown how to use
`the measured power ratio to Suppress the undesired wind
`noise energy.
`One common technique used in noise reduction for single
`input systems is the well-known technique of spectral Sub
`traction. See, e.g., S. F. Boll, Suppression of acoustic noise
`in speech using spectral subtraction, IEEE Trans. Acoust.
`Signal Proc., vol. ASSP-27, Apr. 1979, the teachings of
`which are incorporated herein by reference. The basic
`premise of the spectral Subtraction algorithm is to parametri
`cally estimate the optimal Wiener filter for the desired
`speech signal. The problem can be formulated by defining a
`noise-corrupted speech signal y(n) according to Equation
`(10) as follows:
`
`25
`
`30
`
`35
`
`where s(n) is the desired signal and Vn) is the noise signal.
`FIG. 4 illustrates noise Suppression using a single-channel
`Wiener filter. The optimal filter is a filter that, when con
`volved with the noisy signal y(n), yields the closest (in the
`mean-square sense) approximation to the desired signal s(n).
`This can be represented in equation form according to
`Equation (11) as follows:
`s(n)-hy(n)
`
`(11)
`
`66:
`
`denotes convolution. The optimal filter that
`where
`minimizes the mean-square difference between s(n) and S(n)
`is the Wiener filter. In the frequency domain, the result is
`given by Equation (12) as follows:
`
`where C. is an experimentally determined decay constant
`(e.g., C.-0.125), and r is the displacement (distance) vari
`able. A plot of this function is shown in FIG. 2. The rapid
`decay of spatial coherence results in the difference in powers
`between the Sums and differences of closely-spaced pressure
`(Zero-order) microphones to be much smaller than for an
`acoustic planewave propagating along the microphone array
`axis. As a result, it is possible to detect whether the acoustic
`signals transduced by the microphones are turbulent-like or
`propagating acoustic signals by comparing the Sum and
`difference signal powers. FIG. 3 shows the difference-to
`Sum power ratios (i.e., the ratio of the difference signal
`power to the Sum signal power) for acoustic and turbulent
`signals for a pair of omnidirectional microphones spaced at
`2 cm in a convective fluid flow propagating at 5 m/s. It is
`clearly seen in this figure that there is a relatively wide
`difference between the desired acoustic and turbulent dif
`ference-to-sum power ratios. The ratio difference becomes
`more pronounced at low frequencies since the differential
`microphone output for desired acoustic signals rolls off at-6
`dB/octave, while the predicted, undesired turbulent compo
`nent rolls off at a much slower rate.
`If sound arrives from off-axis from the microphone array,
`the difference-to-sum power ratio becomes even smaller. (It
`has been assumed that the coherence decay is similar in
`directions that are normal to the flow). The closest the sum
`and difference powers come to each other is for acoustic
`signals propagating along the microphone axis (e.g., when
`0–0 in FIG. 1). Therefore, the power ratio for acoustic
`signals will be less than or equal to the power ratio for
`acoustic signals arriving along the microphone axis. This
`
`40
`
`45
`
`Gs (a))
`Hopf (co) =
`Gy (co)
`
`12
`(12)
`
`where G(a)) is the cross-spectrum between the signals s(n)
`and y(n), and G(co) is the auto power-spectrum of the
`signaly(n). Since the noise and desired signals are assumed
`to be uncorrelated, the result can be rewritten according to
`Equation (13) as follows:
`
`50
`
`Gss (co)
`Hopf (co) =
`Gss (co) + Gy(CO)
`
`13
`(13)
`
`55
`
`Rewriting Equation (11) into the frequency domain and
`Substituting terms yields Equation (14) as follows:
`
`60
`
`65
`
`S(a))
`
`Gyy (co) - Gy (co)
`Gyy (co)
`
`(14)
`
`This result is the basic equation that is used in most spectral
`Subtraction schemes. The variations in spectral Subtraction/
`spectral Suppression algorithms are mostly based on how the
`estimates of the auto power-spectrums of the signal and
`noise are made.
`
`Page 17 of 27
`
`
`
`7
`When speech is the desired signal, the standard approach
`is to use the transient nature of speech and assume a
`stationary (or quasi-stationary) noise background. Typical
`implementations use short-time Fourier analysis-and-syn
`thesis techniques to implement the Wiener filter. See, e.g., E.
`J. Diethorn, “Subband Noise Reduction Methods.” Acoustic
`Signal Processing for Telecommunication, S. L. Gay and J.
`Benesty, eds., Kluwer Academic Publishers, Chapter 9, pp.
`155-178. Mar. 2000, the teachings of which are incorpo
`rated herein by reference. Since both speech and turbulent
`noise excitation are non-stationary processes, one would
`have to implement Suppression schemes that are capable of
`tracking time-varying signals. As such, time-varying filters
`should be implemented. In the frequency domain, this can be
`accomplished by using short-time Fourier analysis and Syn
`thesis or filter-bank structures.
`
`Multi-Channel Wiener Filter
`The previous section discussed the implementation of the
`single-channel Wiener filter. However, the use of micro
`phone arrays allows for the possibility of having multiple
`channels. A relatively simple case is a first-order differential
`microphone that utilizes two closely-space omnidirectional
`microphones. This arrangement can be seen to be essentially
`equivalent to a single-input/single-output system as shown
`in FIG. 5, where the desired “noise-free” signal is shown as
`Z(n). It is assumed that the noise signals at both microphones
`are uncorrelated, and thus the two noises can be added
`equivalently as a single noise source. If the added noise
`signal is defined as V(n)-V (n)+V (n), then the output from
`the second microphone can be written according to Equation
`(15) as follows:
`
`10
`
`15
`
`25
`
`30
`
`From the previous definition of the coherence function, it
`can be shown that the output noise spectrum is given by
`Equation (16) as follows:
`
`35
`
`and the coherent output power is given by Equation (17) as
`follows:
`
`G.(a) = yin (co)G2p2(co)
`
`(17)
`
`Thus the signal-to-noise ratio is given by Equation (18) as
`follows:
`
`40
`
`45
`
`50
`
`SNR
`
`... -
`
`in 2(co)
`G.(a)
`() - Go-1,2,...,
`
`(18)
`
`55
`
`Using the expression for the Wiener filter given by Equation
`(13) Suggests a simple Wiener-type spectral Suppression
`algorithm according to Equation (19) as follows:
`
`60
`
`Hopi (co) = yin (co)
`
`(19)
`
`65
`
`US 7,171,008 B2
`
`8
`FIG. 6 shows the amount of noise suppression that is applied
`as a function of coherence between the two microphone
`signals.
`One major issue with implementing a Wiener noise reduc
`tion scheme as outlined above is that typical acoustic signals
`are not stationary random processes. As a result, the esti
`mation of the coherence function should be done over short
`time windows so as to allow tracking of dynamic changes.
`This problem turns out to be substantial when dealing with
`turbulent wind-noise that is inherently highly non-stationary.
`Fortunately, there are other ways to detect incoherent signals
`between multi-channel microphone systems with highly
`non-stationary noise signals. One way that is effective for
`wind-noise turbulence, slowly propagating signals, and
`microphone self-noise, is described in the next section.
`It is straightforward to extend the two-channel results
`presented above to any number of channels by the use of
`partial coherence functions that provide a measure of the
`linear dependence between a collection of inputs and out
`puts. A multi-channel least-squares estimator can also be
`employed for the signals that are linearly related between the
`channels.
`Wind-Noise Suppression
`The goal of turbulent wind-noise Suppression is to deter
`mine what frequency components are due to turbulence
`(noise) and what components are desired acoustic signal.
`Combining the results of the previous sections indicates how
`to proceed. The noise power estimation algorithm is based
`on the difference in the powers of the sum and difference
`signals. If these differences are much smaller than the
`maximum predicted for acoustic signals (i.e., signals propa
`gating along the axis of the microphones), then the signal
`may be declared turbulent and used to update the noise
`estimation. The gain that is applied can be the Wiener gain
`as given by Equations (14) and (19), or a weighting (pref
`erably less than 1) that can be uniform across frequency. In
`general, the gain can be any desired function of frequency.
`One possible general weighting function would be to
`enforce the difference-to-sum power ratio that would exist
`for acoustic signals that are propagating along the axis of the
`microphones. The fluctuating acoustic pressure signals trav
`eling along the microphone axis can be written for both
`microphones as follows:
`
`p2(t)=S(t-t')+v (t-t')+n2(i)
`
`(20)
`
`where t is the delay for the propagating acoustic signal s(t),
`T, is the delay for the convective or slow propagating waves,
`and n (t) and n(t) represent microphone self-noise and/or
`incoherent turbulent noise at the microphones. If the signals
`are represented in the frequency domain, the power spec
`trum of the pressure sum (p(t)+p(t)) and difference signals
`(p(t)-p(t)) can be written as follows:
`
`Gd (co) = 4Picosin
`
`God
`2c
`
`+4Y? (oricosin God --
`C
`2U
`
`and
`
`G, (a) = 4P(a) + 4Y'(a)(a)) +
`
`(21)
`
`(22)
`
`Page 18 of 27
`
`
`
`The ratio of these factors (denoted as PR) gives the expected
`power ratio of the difference and sum signals between the
`microphones as follows:
`
`Gd(co)
`PR(co) =
`G(co)
`
`(23)
`
`where Y is the turbulence coherence as measured or pre
`dicted by the Corcos or other turbulence model, Y(co) is the
`RMS power of the turbulent noise, and N and N represent
`the RMS power of the independent noise at the microphones
`due to sensor self-noise. For turbulent flow where the
`convective wave speed is much less than the speed of Sound,
`the power ratio will be much greater (by approximately the
`ratio of propagation speeds) and thereby moves the power
`ratio to unity. Also, as discussed earlier, the convective
`turbulence spatial correlation function decays rapidly, and
`this term becomes dominant when turbulence (or indepen
`dent sensor self-noise is present) and thereby moves the
`power ratio towards unity. For a purely propagating acoustic
`signal traveling along the microphone axis, the power ratio
`is as follows:
`
`PR, (o)=sin()
`
`- sin2
`
`God
`
`(24)
`
`For general orientation of a single plane-wave where the
`angle between the planewave and the microphone axis is 0.
`
`10
`
`15
`
`25
`
`30
`
`dcost
`PR (co, 0) =sin() COS
`
`(25)
`
`35
`
`40
`
`The results shown in Equations (24)–(25) lead to an algo
`rithm for suppression of airflow turbulence and sensor
`self-noise. The rapid decay of spatial coherence or large
`difference in propagation speeds, results in the relative
`powers between the sums and differences of the closely
`spaced pressure (Zero-order) microphones to be much
`Smaller than for an acoustic planewave propagating along
`45
`the microphone array axis. As a result, it is possible to detect
`whether the acoustic signals transduced by the microphones
`are turbulent-like noise or propagating acoustic signals by
`comparing the Sum and difference powers.
`FIG. 3 shows the difference-to-sum power ratio for a pair
`of omnidirectional microphones spaced at 2 cm in a con
`vective fluid flow propagating at 5 m/s. It is clearly seen in
`this figure that there is a relatively wide difference between
`the acoustic and turbulent sum-difference power ratios. The
`ratio differences become more pronounced at low frequen
`cies since the differential microphone rolls off at-6 dB/oc
`tave, where the predicted turbulent component rolls off at a
`much slower rate.
`If sound arrives from off-axis from the microphone array,
`the ratio of the difference-to-sum power levels becomes
`even smaller as shown in Equation (25). Note that it has been
`assumed that the coherence decay is similar in directions
`that are normal to the flow. The closest the sum and
`difference powers come to each other is for acoustic signals
`propagating along the microphone axis. Therefore, if acous
`tic waves are assumed to be propagating along the micro
`phone axis, the power ratio for acoustic signals will be less
`
`50
`
`55
`
`60
`
`65
`
`US 7,171,008 B2
`
`10
`than or equal to acoustic signals arriving along the micro
`phone axis. This limiting approximation is the key to pre
`ferred embodiments of the present invention relating to
`noise detection and the resulting Suppression of signals that
`are identified as turbulent and/or noise. The proposed sup
`pression gain SG(co) can thus be stated as follows: If the
`measured ratio exceeds that given by Equation (25),