(19) United States
`(12) Patent Application Publication (10) Pub. No.: US 2009/0271187 A1
`Yen et al.
`(43) Pub. Date:
`Oct. 29, 2009
`
`US 20090271 187A1
`
`(54) TWO MICROPHONE NOISE REDUCTION
`SYSTEM
`
`(76) Inventors:
`
`Kuan-Chieh Yen, Northville, MI
`(US); Rogerio Guedes Alves,
`Macomb Twp, MI (US
`p
`(US)
`
`Correspondence Address:
`BLACKLOWE & GRAHAM, PLLC
`701 FIFTHAVENUE, SUITE 4800
`SEATTLE, WA 98104 (US)
`
`(21) Appl. No.:
`
`12/110, 182
`
`(22) Filed:
`
`Apr. 25, 2008
`
`Publication Classification
`
`(51) Int. Cl.
`(2006.01)
`GIOL 2L/02
`(52) U.S. Cl. ................................. 704/226; 704/E21.001
`(57)
`ABSTRACT
`A two microphone noise reduction system is described. In an
`embodiment, input signals from each of the microphones are
`divided into subbands and each subband is then filtered inde
`pendently to separate noise and desired signals and to Sup
`press non-stationary and stationary noise. Filtering methods
`used include adaptive decorrelation filtering. A post-process
`ing module using adaptive noise cancellation like filtering
`algorithms may be used to further Suppress stationary and
`non-stationary noise in the output signals from the adaptive
`decorrelation filtering and a single microphone noise reduc
`tion algorithm may be used to further provide optimal sta
`tionary noise reduction performance of the system.
`
`x (n)
`
`
`
`H
`
`vo (n)
`
`So(n)
`
`Page 1 of 29
`
`GOOGLE EXHIBIT 1013
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 1 of 17
`
`US 2009/0271187 A1
`
`
`
`a(n)
`
`in
`
`1 - A B
`
`v(n)
`
`s(n)
`
`F.G. 1
`
`Page 2 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 2 of 17
`
`US 2009/0271187 A1
`
`x (m)
`
`X (n)
`0- AFB
`\-
`201
`
`
`
`vo (m)
`-
`
`26
`
`l
`
`LR
`
`S(n)
`--> SFB -—b
`V
`204
`
`S. (m)
`
`2O1
`
`AFB
`
`x (m)
`
`x, (n)
`
`$ (m)
`
`H SFB HD
`v. (m) t
`V-
`s, (n)
`205
`204
`
`FIG 2
`
`Page 3 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 3 of 17
`
`US 2009/0271187 A1
`
`DeCompose the input signals into Subband
`Signals
`
`301
`
`Apply an ADF algorithm independently in each
`Subband
`
`302
`
`
`
`
`
`Combine the restored signals
`
`303
`
`FIG 3
`
`Page 4 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 4 of 17
`
`US 2009/0271187 A1
`
`Compute latest samples of the separated
`Signals based on Current filter estimates
`
`4O1
`
`Next m
`
`Update the coefficients of the filters
`
`
`
`Apply a filter to compensate distortion
`
`402
`
`403
`
`FG. 4
`
`Page 5 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 5 of 17
`
`US 2009/0271187 A1
`
`Computing a subband step-size function
`
`
`
`Updating the Coefficients using the step-size
`
`FIG. 5
`
`Page 6 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 6 of 17
`
`US 2009/0271187 A1
`
`Computing power levels of first and second channel
`Subband input signals
`
`6O1
`
`Computing a phase of a cross-Correlation between
`the second channel subband input signal and a
`second channel Subband separated signal
`
`6O2
`
`Computing a power level of a first channel Subband
`restored signal
`
`603
`
`Computing the Subband step-size function based on
`one or more of the previously computed values
`
`604
`
`FIG. 6
`
`Page 7 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 7 of 17
`
`US 2009/0271187 A1
`
`d(n) = tin) + n(n)
`
`
`
`FG. 7
`
`Page 8 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 8 of 17
`
`US 2009/0271187 A1
`
`Synthesis Filter Bank
`
`Analysis Filter Bank
`
`
`
`|
`4
`
`cd) AD Fale)
`
`
`
`
`> H,(z)
`>| yD
`> ‘pd
`>
`F,(z)
`
`>| Holz)
`
`Mic_0
`
`d(n)= t(n) + n(n)
`
`
`
`> Lo a
`|
`
`am
`
`
`
`v
`
`
`
`em)
`
`
`
`oe- H(z). —> fo t
`
`
`
`
`
`
`
`
`
`
`
`
`
`l.
`|
`dxe(™)
`&xo(m)
`
`
`
`
`
`
`
`
`
`> to
`>) Falz) —]-———>
`> }o
`ro} HalZ)
`
`I, e{n)
`ae, i ‘
`|
`Yerr(M)E=
`
`aa
`801
`;
` Gxrlz}
`802
`803
`
`Mic_1
`Analysis Filter Bank
`x“
`“| ~~
`
`
`
`
`
`
`
`
`
`.
`
`yo(m
`
`y.{m
`
`G,(z)
`
`G,(z)
`
`~
`
`x,(m)
`
`x,(m)
`
`
`
`|
`
`x(n}
`
`
`
`
`> H,(z)
`
`> 4D
`
`
`
`»|Xxa(m)
`7
`801
`
`
`| Higo(z), HK yD - nce in
`
`FIG. 8
`
`Page 9 of 29
`
`Page 9 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 9 of 17
`
`US 2009/0271187 A1
`
`
`
`901
`
`
`
`902
`
`Compute latest Samples of the Subband Output
`Signal
`
`Next in
`
`Update the coefficients of the filter
`
`FIG. 9
`
`Page 10 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 10 of 17
`
`US 2009/0271187 A1
`
`Compute latest samples of the subband output
`Signal
`
`1001
`
`
`
`Next m
`
`
`
`Update the coefficients of the filter
`
`Finalise output signal and filter coefficients
`
`1 OO2 for r
`
`1 to R
`
`1003
`
`FIG 10
`
`Page 11 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 11 of 17
`
`US 2009/0271187 A1
`
`
`
`
`
`
`
`
`
`Output
`power >
`reference
`OOwer?
`
`
`
`Compute step-size
`function
`
`
`
`Increment A(m)
`
`
`
`
`
`Step-size = 0
`
`FIG 11
`
`Page 12 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 12 of 17
`
`US 2009/0271187 A1
`
`
`
`d(n)
`
`Synthesis Filter Bank
`
`tD
`
`TD
`
`Fo(z)
`
`F(z)
`
`td H Fk12(z)
`S
`12O1
`
`Z(n)
`
`F.G. 12
`
`Page 13 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 13 of 17
`
`US 2009/0271187 A1
`
`Track average power of quasi-stationary noise
`in a Subband signal
`
`1301
`
`Determine again factor using the average
`power
`
`
`
`Apply the gain factor to the subband signal
`
`
`
`1302
`
`1303
`
`F.G. 13
`
`Page 14 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 14 of 17
`
`US 2009/0271187 A1
`
`d(n) = tin) + n(n)
`
`
`
`Z(m)
`
`Z(n)
`
`SFB
`
`Output
`
`1402
`
`x(n)
`
`1401
`
`F.G. 14
`
`Page 15 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 15 of 17
`
`US 2009/0271187 A1
`
`
`
`Mic 1
`ro
`
`x(m)
`
`AFB
`
`x(n)
`
`1401
`
`FG. 15
`
`Page 16 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 16 of 17
`
`US 2009/0271187 A1
`
`
`
`1603 Xo(n)
`-
`b- AFB
`Mic 0
`?o
`
`Xok (m)
`
`Mic 1
`
`16O1
`N/
`p- AFB
`X (n)
`1604
`
`16O7
`
`FG 16
`
`Page 17 of 29
`
`

`

`Patent Application Publication
`
`Oct. 29, 2009 Sheet 17 of 17
`
`US 2009/0271187 A1
`
`Decompose the input signals into subband
`Signals
`
`1701
`
`Apply an ADF algorithm independently in each
`Subband
`
`1702
`
`Correct for distortion in each Subband
`
`Post-process the signals using ANC
`
`Suppress stationary noise
`
`Combine the output signals
`
`1703
`
`1704
`
`1705
`
`1706
`
`FIG. 17
`
`Page 18 of 29
`
`

`

`US 2009/02711. 87 A1
`
`Oct. 29, 2009
`
`TWO MCROPHONE NOISE REDUCTION
`SYSTEM
`
`FIELD OF THE INVENTION
`
`0001. This invention relates generally to voice communi
`cation systems and, more specifically, to microphone noise
`reduction systems to suppress noise and provide optimal
`audio quality.
`
`BACKGROUND OF THE INVENTION
`
`0002 Voice communications systems have traditionally
`used single-microphone noise reduction (NR) algorithms to
`Suppress noise and provide optimal audio quality. Such algo
`rithms, which depend on statistical differences between
`speech and noise, provide effective Suppression of stationary
`noise, particularly where the signal to noise ratio (SNR) is
`moderate to high. However, the algorithms are less effective
`where the SNR is very low.
`0003 Mobile devices, such as cellular telephones, are
`used in many diverse environments, such as train stations,
`airports, busy streets and bars. Traditional single-microphone
`NR algorithms do not work effectively in these environments
`where the noise is dynamic (or non-stationary), e.g., back
`ground speech, music, passing vehicles etc. In order to Sup
`press dynamic noise and further optimize NR performance on
`stationary noise, multiple-microphone NR algorithms have
`been proposed to address the problem using spatial informa
`tion. However, these are typically computationally intensive
`and therefore are not suited to use in embedded devices,
`where processing power and battery life are constrained.
`0004 Further challenges to noise reduction are introduced
`by the reducing size of devices, such as cellular telephones
`and Bluetooth R) headsets. This reduction in size of a device
`generally increases the distance between the microphone and
`the mouth of the user and results in lower user speech power
`at the microphone (and therefore lower SNR).
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`0005 Preferred and alternative examples of the present
`invention are described in detail below with reference to the
`following drawings:
`0006 FIG. 1 shows a block diagram of an adaptive deco
`rrelation filtering (ADF) signal separation system;
`0007 FIG. 2 shows a block diagram of the preferred ADF
`algorithm;
`0008 FIG. 3 shows a flow diagram of an exemplary
`method of operation of the algorithm shown in FIG. 2;
`0009 FIG. 4 shows a flow diagram of an exemplary sub
`band implementation of ADF:
`0010 FIG. 5 shows a flow diagram of a method of updat
`ing the filter coefficients in more detail;
`0011
`FIG. 6 shows a flow diagram of an exemplary
`method of computing a Subband step-size function;
`0012 FIG. 7 is a schematic diagram of a fullband imple
`mentation of an adaptive noise cancellation (ANC) applica
`tion using two inputs;
`0013 FIG. 8 is a schematic diagram of a subband imple
`mentation of an ANC application using two inputs;
`0014 FIG. 9 shows a flow diagram of an exemplary
`method of ANC;
`
`0015 FIG. 10 shows a flow diagram of data re-using:
`0016 FIG. 11 shows a flow diagram of an exemplary
`control mechanism for ANC;
`0017 FIG. 12 shows a block diagram of a single-channel
`NR algorithm;
`0018 FIG. 13 is a flow diagramofan exemplary method of
`operation of the algorithm shown in FIG. 12;
`(0019 FIGS. 14 and 15 show block diagrams of two exem
`plary arrangements which integrate ANC and NRalgorithms;
`0020 FIG. 16 shows a block diagram of a two-micro
`phone based NR system; and
`0021
`FIG. 17 shows a flow diagram of an exemplary
`method of operation of the system of FIG. 16.
`0022 Common reference numerals are used throughout
`the Figures to indicate similar features.
`
`DETAILED DESCRIPTION OF THE PREFERRED
`EMBODIMENT
`0023. A two microphone noise reduction system is
`described. In an embodiment, input signals from each of the
`microphones are divided into subbands and each subband is
`then filtered independently to separate noise and desired sig
`nals and to Suppress non-stationary and stationary noise. Fil
`tering methods used include adaptive decorrelation filtering.
`A post-processing module using adaptive noise cancellation
`like filtering algorithms may be used to further Suppress sta
`tionary and non-stationary noise in the output signals from
`the adaptive decorrelation filtering and a single microphone
`noise reduction algorithm may be used to further optimize the
`stationary noise reduction performance of the system.
`0024. A first aspect provides a method of noise reduction
`comprising: decomposing each of a first and a second input
`signal into a plurality of Subbands, the first and second input
`signals being received by two closely spaced microphones;
`applying at least one filter independently in each Subband to
`generate a plurality of filtered subband signals from the first
`input signal, wherein said at least one filter comprises an
`adaptive decorrelation filter, and combining said plurality of
`filtered Subband signals from the first input signal to generate
`a restored fullband signal.
`0025. The step of applying at least one filter independently
`in each subband to generate a plurality of filtered subband
`signals from the first input signal may comprise: applying an
`adaptive decorrelation filter in each subband for each of the
`first and second signals to generate a plurality of filtered
`Subband signals from each of the first and second input sig
`nals; and adapting the filter in each subband for each of the
`input signals based on a step-size function associated with the
`Subband and the input signal.
`0026. The step-size function associated with a subband
`and an input signal may be normalized against a total power
`in the subband for both the first and second input signals.
`0027. The direction of the step-size function associated
`with a Subband and one of the first and second input signals
`may be adjusted according to a phase of a cross-correlation
`between an input subband signal from the other of the first and
`second input signals and a filtered Subband signal from said
`other of the first and second input signals.
`0028. The step-size function associated with a subband
`and an input signal may be adjusted based on a ratio of a
`power level of the filtered subband signal from said subband
`input signal to a power level of said Subband input signal.
`0029. The step of applying at least one filter independently
`in each subband to generate a plurality of filtered subband
`
`Page 19 of 29
`
`

`

`US 2009/02711. 87 A1
`
`Oct. 29, 2009
`
`signals from the first input signal may comprise: applying an
`adaptive decorrelation filter independently in each subband to
`generate a plurality of separated Subband signals from each of
`the first and second input signals; and applying an adaptive
`noise cancellation filter to the separated Subband signals inde
`pendently in each subband to generate a plurality of filtered
`Subband signals from the first input signal.
`0030 The step of applying an adaptive noise cancellation
`filter to the separated Subband signals independently in each
`Subband may comprise: applying an adaptive noise cancella
`tion filter independently to a first and a second separated
`Subband signal in each Subband; and adapting each said adap
`tive noise cancellation filter in each Subband based on a step
`size function associated with the separated Subband signal.
`0031. The method may further comprise, for each sepa
`rated Subband signal: if a Subband is in a defined frequency
`range, setting the associated step-size function to Zero if
`power in the separated Subband signal exceeds power in a
`corresponding filtered Subband signal; and if a Subband is not
`in the defined frequency range, setting the associated step
`size function to Zero based on a determination of a number of
`Subbands in the defined frequency range having an associated
`step-size set to Zero.
`0032. The step of applying at least one filter independently
`in each subband to generate a plurality of filtered subband
`signals from the first input signal may comprise: applying an
`adaptive decorrelation filter independently in each subband to
`generate a plurality of separated Subband signals from each of
`the first and second input signals; applying an adaptive noise
`cancellation filter to the separated Subband signals indepen
`dently in each subband to generate a plurality of error sub
`band signals from the first input signal; and applying a single
`microphone noise reduction algorithm to the error Subband
`signals to generate a plurality offiltered Subband signals from
`the first input signal.
`0033. A second aspect provides a noise reduction system
`comprising: a first input from a first microphone; a second
`input from a second microphone closely spaced from the first
`microphone; an analysis filter bank coupled to the first input
`and arranged to decompose a first input signal into Subbands;
`an analysis filter bank coupled to the second input and
`arranged to decompose a second input signal into Subbands;
`at least one adaptive filter element arranged to be applied
`independently in each subband, the at least one adaptive filter
`element comprising an adaptive decorrelation filter element;
`and a synthesis filter bank arranged to combine a plurality of
`restored Subband signals output from the at least one adaptive
`filter element.
`0034. The adaptive decorrelation filter element may be
`arranged to control adaptation of the filter element for each
`subband based on power levels of a first input subband signal
`and a second input Subband signal.
`0035. The adaptive decorrelation filter element may be
`further arranged to control a direction of adaptation of the
`filter element for each subband for a first input based on a
`phase of a cross correlation of a second input Subband signal
`and a second Subband signal output from the adaptive deco
`rrelation filter element.
`0036. The adaptive decorrelation filter element may be
`further arranged to control adaptation of the filter element for
`each subband for the first input based on a ratio of a power
`
`level of a first Subband signal output from the adaptive deco
`rrelation filter element to a power level of a first Subband input
`signal.
`0037. The at least one adaptive filter element may further
`comprise an adaptive noise cancellation filter element.
`0038. The adaptive noise cancellation filter element may
`be arranged to: Stop adaptation of the adaptive noise cancel
`lation filter element for Subbands in a defined frequency range
`where the subband power input to the adaptive noise cancel
`lation filter element exceeds the subband power output from
`the adaptive noise cancellation filter element; and to stop
`adaptation of the adaptive noise cancellation filter element for
`Subbands not in the defined frequency range based on an
`assessment of adaptation rates in Subbands in the defined
`frequency range.
`0039. The at least one adaptive filter element may further
`comprise a single-microphone noise reduction element.
`0040. A third aspect provides a method of noise reduction
`comprising: receiving a first signal from a first microphone;
`receiving a second signal from a second microphone; decom
`posing the first and second signals into a plurality of Sub
`bands; and for each Subband, applying an adaptive decorre
`lation filter independently.
`0041. The step of applying an adaptive decorrelation filter
`independently may comprise, for each adaptation step m:
`computing samples of separated signals Vo (m) and V1 (m)
`corresponding to the first and second signals in a Subband k
`based on estimates of filters of length M with coefficients a
`and be using:
`
`where:
`
`a fa(0)a(1)... a (M-1)
`
`b=fb (O)b (1)... b. (M-1)
`0042 and; updating the filter coefficients, using:
`a"-a"'+1(m)v (m)vo (m)
`
`10043 where * denotes a complex conjugate, u(m) and
`u (m) are Subband step-size functions and where:
`vo (m) vo(m)vo (m-1)... vo (m-M+1)
`
`(m) v1.(m)vi (m-1)... v(m-M+1)
`0044) The subband step-size functions may be given by:
`
`1 Co. O
`2yexp(-jLO,11)
`Atak M (Ofot + O.E.) XaX - Ofot
`and:
`
`(3)
`
`Page 20 of 29
`
`

`

`US 2009/02711. 87 A1
`
`Oct. 29, 2009
`
`-continued
`2yexp(-jLO,00)
`(i.
`pub = -
`- Xmax 1 - Er, O
`M(Oio + O.E.)
`Ok
`
`where:
`
`(4)
`
`where so(m)ands (m) comprise restored Subband signals.
`0045. The method may further comprise, for each sub
`band, applying an adaptive noise cancellation filter indepen
`dently to signals output from the adaptive decorrelation filter.
`0046. The methods described herein may be performed by
`firmware or software in machine readable form on a storage
`medium. The software can be suitable for execution on a
`parallel processor or a serial processor Such that the method
`steps may be carried out in any Suitable order, or simulta
`neously.
`0047 A fourth aspect provides one or more tangible com
`puter readable media comprising executable instructions for
`performing steps of any of the methods described herein.
`0048. This acknowledges that firmware and software can
`be valuable, separately tradable commodities. It is intended to
`encompass software, which runs on or controls “dumb' or
`standard hardware, to carry out the desired functions. It is also
`intended to encompass software which “describes' or defines
`the configuration of hardware, such as HDL (hardware
`description language) software, as is used for designing sili
`con chips, or for configuring universal programmable chips,
`to carry out desired functions.
`0049. The preferred features may be combined as appro
`priate, as would be apparent to a skilled person, and may be
`combined with any of the aspects of the invention.
`0050 Embodiments of the present invention are described
`below by way of example only. These exemplary embodi
`ments represent the best ways of putting the invention into
`practice that are currently known to the Applicant although
`they are not the only ways in which this could be achieved.
`The description sets forth the functions of the exemplary
`embodiments and the sequence of steps for constructing and
`operating the exemplary embodiment. However, the same or
`equivalent functions and sequences may be accomplished by
`different embodiments.
`0051. There are a number of different multiple-micro
`phone signal separation algorithms which have been devel
`oped. One exemplary embodiment is adaptive decorrelation
`filtering (ADF) which is an adaptive filtering type of signal
`separation algorithm based on second-order statistics. The
`algorithm is designed to deal with convolutive mixtures,
`which is often more realistic than instantaneous mixtures due
`to the transmission delay from Source to microphone and the
`reverberation in the acoustic environment. The algorithm also
`assumes that the number of microphones is equal to the num
`ber of sources. However, with careful system design and
`adaptation control, the algorithm can group several noise
`sources into one and performs reasonably well with fewer
`microphones than sources. ADF is described in detail in
`
`“Multi-channel signal separation by decorrelation” by Wein
`stein, Feder and Oppenheim, (IEEE Transactions on Speech
`and Audio Processing, vol. 1, no. 4, pp. 405-413, October
`1993) and a simplification and further discussion on adaptive
`step control is described in Adaptive Co-channel speech
`separation and recognition’ by Yen and Zhao, (IEEE Trans
`actions on Speech and Audio Processing, Vol. 7, no. 2, pp.
`138-151, March 1999).
`0.052 The ADF was developed based on a model for co
`channel environment. Under this environment, the signals
`captured by the microphones, X(n) and X (n), are convolutive
`mixtures of signals from two independent Sound sources,
`so(n) and s(n). Here n is the time index in the fullband
`domain. Without losing generality, So(n) can be defined as the
`target Source for Xo (n) and S (n) as the target source for X (n).
`For a given microphone, the Source that is not the target is the
`interfering source. The relation between the source and
`microphone signals can be modelled mathematically as:
`
`0053 where linear filters Ho (Z) and Ho (Z) model the
`relative cross acoustic paths. These filters can be approxi
`mated by N-tap finite impulse response (FIR) filters. The
`sources are relatively better captured by the microphones that
`target them if:
`
`0054 for all frequencies. This is the preferable condition
`for the ADF algorithm to prevent permutation problem due to
`the ambiguity on target sources. This co-channel model and
`the ADF algorithm can both be extended for more micro
`phones and signal sources.
`0055 FIG. 1 shows a block diagram of the ADF signal
`separation system for two microphones, which uses two
`adaptive filters 101, 102 to estimate and track the underlying
`relative cross acoustic paths from signals Xo(n) and X (n)
`received from the two microphones. Using these filters, the
`system can separate the sources from these convolutive mix
`tures, and thus restore the source signals. Depending on the
`sampling frequency, the reverberation in the environment,
`and the separation of Sources and microphones, acoustic
`paths typically require FIR filters with hundreds or even thou
`sands of taps to be modeled digitally. Therefore, the tail
`lengths of the adaptive filters A(Z) and B(Z) can be quite
`Substantial. This is further complicated because audio signals
`are usually highly colored and dynamic and acoustic environ
`ments are often time-varying. As a result, satisfactory track
`ing performance may require a large amount of computa
`tional power.
`0056 FIG. 2 shows a block diagram of an optimized ADF
`algorithm where the signal separation is implemented in the
`frequency (Subband) domain. The block diagram shows two
`input signals, X(n), X (n), which are received by different
`microphones. Where one of the microphones is located closer
`to the user's mouth, the signal received by that microphone
`(e.g., Xo(n)) can comprise relatively more speech (e.g., So(n))
`whilst the signal received by the other microphone (e.g.,
`X (n)) can comprise relatively more noise (e.g., s(n)). There
`fore, the speech is the target source in X(n) and the interfering
`Source in X (n), while the noise is the target source in X (n)
`and the interfering source in Xo(n). The operation of the algo
`rithm can be described with reference to the flow diagram
`shown in FIG. 3. Although the exemplary embodiments
`
`Page 21 of 29
`
`

`

`US 2009/02711. 87 A1
`
`Oct. 29, 2009
`
`shown and described herein relate to two microphones, the
`systems and methods described may be extended to more than
`two microphones.
`0057 The term speech is used herein in relation to a
`Source signal to refer to the desired speech signal from a user
`that is to be preserved and restored in the output. The term
`noise is used herein in relation to a source signal to refer to
`an undesired competing signal (which originates from mul
`tiple actual sources), including background speech, which is
`to be suppressed or removed in the output.
`0058. The input signals x(n), X (n) are decomposed into
`Subband signals Xo,(m), X (m) (block 301) using an analy
`sis filter bank (AFB) 201, where k is the subband index and m
`is the time index in the Subband domain. Because the band
`width of each subband signal is only a fraction of the full
`bandwidth, the subband signals can be down-sampled for
`processing efficiency without losing information (i.e., with
`out violating the Nyquist sampling theorem). An exemplary
`embodiment of the AFB is the Discrete Fourier Transform
`(DFT) analysis filter bank, which decomposes a fullband
`signal into Subband signals of equally spaced bandwidths:
`
`k
`f2ink
`W-
`Xik (m) = =0 x(mD +n)wn exp(-i). k = 0, 1, ... , 2.
`
`(7)
`
`0059 where D is the down-sample factor, K is the DFT
`size, and w(n) is the prototype window of length W designed
`to achieve the intended cross-band rejection. This shows just
`one example of an AFB which may be used and depending on
`the type of the AFB, the subband signals can be either real or
`complex, and the bandwidth of the subbands can be either
`uniform or non-uniform. For AFB with non-uniform sub
`bands, different down-sampling factor may be used in each
`subband.
`0060 Having decomposed the input signals (in block
`301), an ADF algorithm is applied independently to each
`subband (block 302) using subband ADF filters A, (z) and
`B(Z). 202, 203. These filters are adapted by estimating and
`tracking the relative cross acoustic paths from the micro
`phone signals (Ho (Z) and Ho (Z) respectively), with filter
`A (Z) providing the coupling from the second channel (chan
`nel 1) into the first channel (channel 0) and filter B(Z) pro
`viding the coupling from the first channel (channel 0) into the
`second channel (channel 1). The subband ADF algorithm is
`described in more detail below. The output of the ADF algo
`rithms comprises restored Subband signals so(m), S(m)
`and these separated signals are then combined (block 303) to
`generate the fullband restored signals So(n) and S (n) using a
`synthesis filter bank (SFB) 204 that matches the AFB 201.
`0061. By using subbands as shown in FIGS. 2 and 3, each
`Subband comprising a whiter input signal and a shorter filter
`tail can be used in each Subband due to down-sampling. This
`reduces the computational complexity and optimizes the con
`Vergence performance.
`0062. The subband filters A(Z) and B(z) are FIR filters of
`length M with coefficients:
`
`mately N/D, due to the down-sampling, in order to provide
`similar temporal coverage as a fullband ADF filter of length
`N. It will be appreciated that the filter length, M. may be
`different to (e.g., longer than) N/D.
`0064 FIG. 4 shows a flow diagram of an example subband
`implementation of ADF. The flow diagram shows the imple
`mentation for a single Subband and the method is performed
`independently for each Subband k. In each adaptation step m,
`the latest samples of the separated signals Vo(m) and V. (m)
`are computed (block 401) based on the current estimates of
`filters A (Z) and B(Z), where:
`
`(9)
`v(m)=x(m)-vo (m)5."'
`0065 where the subband input signal vectors are defined
`aS
`
`I0066. These computed values of the latest samples vo(m)
`and V. (m) are then used to update the coefficients of filters
`A (Z) and B(Z) (block 402) using the following adaptation
`equations:
`
`I0067 where * denotes a complex conjugate, u(m) and
`u (m) are Subband step-size functions (as described in more
`detail below) and where the subband separated signal vectors
`are defined as:
`vo(m)-vo (m)vo (m-1)
`
`vo (m-M+1)
`
`0068. The separated signals may then be filtered (block
`403) to compensate for distortion using the filter (1-A, (z)B.
`(z))' 205. The output of the ADF algorithm comprises
`restored Subband signals so (m) and S. (m).
`0069. In this example, the control mechanism is imple
`mented independently in each Subband. In other examples,
`the control mechanism may be implemented across the full
`band or across a number of Subbands (e.g., cross-band con
`trol).
`0070 FIG. 5 shows a flow diagram of the methodofupdat
`ing the filter coefficients (e.g., block 402 from FIG. 4) in more
`detail. The method comprises computing a Subband step-size
`function (block 501) and then using the computed subband
`step-size function to update the coefficients (block 502), e.g.,
`using the adaptation equations given above.
`(0071. The step-size functions u (m) and u (m) control
`the rate of filter adaptation and may also be referred to the
`adaption gain function or adaptation gain. An upper bound of
`step-size for the Subband implementation is:
`
`0 < uk, it
`
`2
`3 M (Ofot + O.E.)
`
`(11)
`
`(8)
`b=fb (O)b (1)b (M-1)
`0063 where the superscript T denotes vector transpose.
`The subband filter length, M, preferably needs to be approxi
`
`E{Ix, (m)}, i=0,1, represents the
`(0072 where of
`power of Subband microphone signal x, (m).
`
`Page 22 of 29
`
`

`

`US 2009/02711. 87 A1
`
`Oct. 29, 2009
`
`0073. According to this upper bound, the step-size may be
`defined as:
`
`Atak Filbk
`
`2y
`, 0 < y < 1
`M (Ofot + O.E.)
`
`(12)
`
`0074 This provides a power-normalized ADF algorithm
`whose adaptation is insensitive to the input level of the micro
`phone signals. This step-size function is generally sufficient
`for applications with stationary signals, time-invariant mix
`ing channels, and moderate cross-channel leakage.
`0075. However, for applications with dynamic signals,
`time-varying channels, and high cross-channel leakage. Such
`as when separating target speech from dynamic interfering
`noise with closely-spaced microphones, the adjustment of
`step-size may be further refined to optimize performance.
`Three further optimizations are described below:
`0076 Power normalization
`0077 Adaption direction control
`0078 Target ratio control
`0079 Any one or more of these optimizations may be used
`in combination with the methods described above, or alter
`natively none of these optimizations may be used.
`0080. The input signals are time-varying and as a result the
`power levels of the input signals, Oo, and O,
`are also
`time-varying. Dynamic tracking of the power levels in each
`Subband can beachieved by averaging the input power in each
`subband with a 1-tap recursive filter with adjustable time
`coefficient or weighted windows with adjustable time span.
`The resulting input power estimates, Go, and 6, are used
`in place of Oof and O.in the step-size function. The ability
`to follow the increase in input power levels promptly reduces
`instability and the ability to follow the decrease in input
`power levels within a reasonable time frame avoids unneces
`sarily stalling of the adaptation (because the step-size is
`reduced as power increases) and enhances the dynamic track
`ing ability of the system. However, when source signals are
`absent, it is desirable that the input power estimates do not
`drop to the level of noise floor. This prevents the negative
`impact on filter adaption during these idle periods. Therefore,
`the time coefficient or weighted windows should be adjusted
`Such that the averaging period of the input power estimates
`are short within normal power level variation but long when
`incoming power level is significantly lower.
`0081 Adaptation direction control comprises controlling
`the direction of the step-size, Landu, through the addition
`of an extra term in the equation. This term stops the filter from
`diverging under certain circumstances. The following
`description provides a derivation of the extra term.
`0082 Previous work (as descri