CN108694957A - The echo cancelltion design method formed based on circular microphone array beams - Google Patents
The echo cancelltion design method formed based on circular microphone array beams Download PDFInfo
- Publication number
- CN108694957A CN108694957A CN201810304397.1A CN201810304397A CN108694957A CN 108694957 A CN108694957 A CN 108694957A CN 201810304397 A CN201810304397 A CN 201810304397A CN 108694957 A CN108694957 A CN 108694957A
- Authority
- CN
- China
- Prior art keywords
- signal
- microphone array
- array
- sound
- echo
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 54
- 230000004044 response Effects 0.000 claims abstract description 23
- 239000013598 vector Substances 0.000 claims abstract description 20
- 230000003044 adaptive effect Effects 0.000 claims abstract description 19
- 230000008569 process Effects 0.000 claims abstract description 12
- 238000010586 diagram Methods 0.000 claims description 8
- 238000001914 filtration Methods 0.000 claims description 8
- 238000005314 correlation function Methods 0.000 claims description 6
- 230000035939 shock Effects 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 4
- 238000001228 spectrum Methods 0.000 claims description 4
- 239000002131 composite material Substances 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 2
- 230000000644 propagated effect Effects 0.000 claims description 2
- 230000017105 transposition Effects 0.000 claims description 2
- 238000009795 derivation Methods 0.000 claims 1
- 230000005764 inhibitory process Effects 0.000 abstract description 4
- 230000000694 effects Effects 0.000 abstract description 3
- 238000006386 neutralization reaction Methods 0.000 abstract description 2
- 239000011159 matrix material Substances 0.000 description 6
- 230000008901 benefit Effects 0.000 description 2
- 230000021615 conjugation Effects 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000000926 separation method Methods 0.000 description 2
- 241000209140 Triticum Species 0.000 description 1
- 235000021307 Triticum Nutrition 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000005236 sound signal Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
- 230000002087 whitening effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
- G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
- G10L21/0208—Noise filtering
- G10L21/0216—Noise filtering characterised by the method used for estimating noise
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
- G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
- G10L21/0208—Noise filtering
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
- G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
- G10L21/0208—Noise filtering
- G10L2021/02082—Noise filtering the noise being echo, reverberation of the speech
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
- G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
- G10L21/0208—Noise filtering
- G10L21/0216—Noise filtering characterised by the method used for estimating noise
- G10L2021/02161—Number of inputs available containing the signal or the noise to be suppressed
- G10L2021/02166—Microphone arrays; Beamforming
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Quality & Reliability (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Circuit For Audible Band Transducer (AREA)
Abstract
The present invention proposes a kind of echo cancelltion design method formed based on circular microphone array beams, mainly solves the problems, such as that the prior art can not precise gain signal source and inhibition noise source.Its realization process is:(1)The sound field impulse response in room is obtained according to sef-adapting filter;(2)Loud speaker plays calibration sound, designs corresponding steering vector, circular microphone array is made to recognize interference source direction;(3)The weight coefficient of spatial filter is designed according to gained steering vector, weighting obtains the optimal direction figure of circular microphone array;(4)It is filtered using the Subband adaptive filters of closed loop configuration in time domain, offsets the echo signal of loud speaker.This method combines spatial domain Wave beam forming and time-domain adaptive sub-band filter method, and it is notable that design obtains echo neutralization effect, improves quality of speech signal and echo cancellor speed.
Description
Technical field
The invention belongs to voice process technology fields, and in particular to one kind is formed based on circular microphone array beams
Echo cancelltion design method, for precise gain signal source and inhibit the signal in noise source direction.
Background technology
Microphone array signals processing is a kind of emerging technology, has become a research of field of voice signal
Hot spot.Single microphone received signal is superimposed by multi-acoustical and ambient noise, and the separation of each sound source is difficult to realize,
It thus cannot achieve auditory localization and separation.In order to solve these limitations of single microphone, carried out using microphone array
The method of speech processes is seasonable and gives birth to.The wheat that microphone array is put by one group by certain geometry (common linear, annular)
Gram wind composition carries out space time processing to the voice signal in the different spaces direction of acquisition, realizes noise suppressed, reverberation removal, people
The functions such as acoustic jamming inhibition, sound source direction finding, audio source tracking, array gain, and then Speech processing quality is improved, it is true to improve
Phonetic recognization rate under real environment.Therefore, the echo cancelltion based on microphone array is studied, is had great importance.
Currently, the research for microphone array echo cancelltion, the method for focusing primarily upon self-adaptive filters in time area is improved.
For example, least mean-square error LMS algorithm or recursive least squares, the latter is intended between expected response and filter output
Quadratic sum is minimum.When receiving the new sampled value of input signal in each iteration, minimum two is solved using recursive form
Multiply problem, the coefficient of sef-adapting filter is made to be updated, expected response and filter export in least square meaning most
Match.LMS filtering is a kind of transient analysis method, i.e., will reappraise quadratic sum to all with input signal at each moment,
And reach minimum by adjusting weight vector.But the performance of this method is influenced by input signal, LMS filters can be met
The problem of amplifying to gradient noise.
Invention content
It is an object of the invention to:For above-mentioned existing methods disadvantage, propose a kind of based on circular microphone array wave
The echo cancelltion design method that beam is formed, the signal for precise gain signal source and inhibition noise source direction.
In order to achieve the above object, the technical solution adopted in the present invention is:
The echo cancelltion design method formed based on circular microphone array beams, which is characterized in that the method includes
Following steps:
(1) according to the principle of sef-adapting filter, the sound field in room is estimated using the IMAGE methods based on ray model
Impulse response, that is, loud speaker provide to be canceled to the space echo path between microphone array for subsequent echo cancelltion
Loud speaker interference signal;
(2) loud speaker plays calibration sound, and circular microphone array makes according to direction of arrival estimation method of the TDOA based on time delay
Circular microphone array recognizes loud speaker interference source direction;
(3) orientation that teller is recognized according to direction of arrival estimation method of the TDOA based on time delay, is interfered using loud speaker
With the azimuth information of teller, array steering vector is obtained, the weight coefficient of spatial filter is then designed according to steering vector, is added
Power obtains the spatial domain optimal direction figure of circular microphone array, and the main lobe direction of array is made to be directed toward teller, and loud speaker interferes
Direction is located at array low sidelobe, and the purpose of airspace filter is to make the voice signal of microphone array acquisition teller, gain signal
Inhibit the signal in interference source direction while source;
(4) airspace filter has cut off the acoustics circuit of loud speaker and microphone array, remaining residual echo signal by when
Domain is handled, since voice signal is broadband signal, so being filtered using the Subband adaptive filters of closed loop configuration in time domain
Wave offsets the residual echo signal of loud speaker.
Further, the specific implementation process of the step (1) is:If describing room sound with linear time invariant system
Field channel model, sound-source signal is s (t), therefore microphone array reception signal x (t) can be expressed as s (t) and sound field channel
The convolution form of impulse response h (t), i.e.,:
X (t)=s (t) * h (t)+n (t) (1)
Wherein * indicates that convolution algorithm, n (t) are noise, and the purpose of channel estimation is exactly the condition in known s (t) and x (t)
Lower solution shock response h (t), in ray sound-field model, the shock response of channel may be considered by a series of intrinsic sound rays
Determining, at this moment impulse response is:
Wherein, M is the number of intrinsic sound ray, AiAnd τiIt is i-th of intrinsic corresponding decay factor and propagation delay respectively.
Further, the realization process of the step (2) is:According to the geometric format of array, (time delay is based in conjunction with TDOA
The arrival bearing angular estimation of estimation) method, the sound of stationary sound source is by room propagated, due to the geometry of microphone array
Form, there are one time delay, this time delays to be solved according to cross-correlation function for the voice signal that microphone receives two-by-two,
Again to arbitrarily microphone receives signal simultaneous solution two-by-two, where deriving sound bearing, loudspeaker calibration sound side can be oriented
Position.
Further, the specific implementation process of the step (3) is:It finds sound bearing and obtains one group of Space Angle letter later
Horizontal azimuth and vertical elevation (α, β) are ceased, the steric direction vector A of array, () then can be obtainedHIndicate conjugate transposition;
Wherein, τijFor i and j, array element receives the time delay of signal, { ω two-by-twoij, i, j=1,2 ..., N } it is that spatial domain is weighed;
Spatial domain weighting is carried out per voice signal all the way to microphone array, obtains optimal direction figure;
The directional diagram of array is defined as:
P (θ)=s |ωHA(θ)| (4)
Different weight vectors can make the signal on different directions have different responses, to form the space wave of different directions
Beam so that beam main lobe is directed toward effective Sounnd source direction, and null beam position interferes Sounnd source direction;
Realize that Spatially adaptive filtering, array received signal are expressed as using maximum signal noise ratio principle
X (t)=xs(t)+xn(t) (5)
xs(t) it is corresponding useful signal part, xn(t) it is interference and noise section;
Then the output of array is after Wave beam forming:
Y (t)=ωHX (t)=ωHxs(t)+ωHxn(t)(6)
Further, the specific implementation process of the step (4) is:Processing in time domain uses the subband of closed loop configuration certainly
Adaptive filter filters, and essence is that the voice signal that will be received is divided into several subsignals on frequency spectrum, in each subband,
Using normalization minimum mean-square adaptive algorithm so that subband mean square error is minimum;
The sub-filter algorithm renewal equation of the closed loop configuration of core is:
Input signal extremely closes weight relative to Δ is postponed existing for desired signal, to the compensation of delay in closed loop sub-band structure
It wants;
Wherein,L is the subfilter length , [ of analysis and composite filter group;·]Indicate round numbers
Part.
Compared with prior art, the beneficial effects of the invention are as follows:It is proposed by the present invention to be based on circular microphone array beams
The echo cancelltion design method of formation, the advantage is that:
(1) spatial domain Wave beam forming and time-domain adaptive sub-band filter method are combined, it is aobvious that design obtains echo neutralization effect
It writes, improves quality of speech signal and echo cancellor speed.
(2) it uses sub-band adaptive filtering method to realize time-domain filtering, the broadband signal of input can be divided into several
The subband signal for carrying a small amount of source signal, to reduce the complexity of system processing.
(3) the microphone array array number of the acoustic echo cancellation system of mainstream is small now, 3~5 average, strictly can not
Circular array is constituted, positioning accuracy is low, and airspace filter effect is poor, and the present invention is using frequency dividing sub-band filter, more efficient essence in time domain
Really.
(4) have the advantages that can null interference radiating way and the effective information source direction of gain, can be used for higher to performance requirement
Tele-conferencing system.
Description of the drawings
Fig. 1 is the realization general flow chart of the present invention;
Fig. 2 is airspace filter distribution map of the present invention;
Fig. 3 is that optimal direction diagram of the present invention is intended to;
Fig. 4 is the sub-process figure of Wave beam forming of the present invention;
Fig. 5 is uniform circular array aspect figure of the present invention;
Fig. 6 is time domain sub-band filter method schematic diagram of the present invention.
Specific implementation mode
For the ease of those of ordinary skill in the art understand and implement the present invention, with reference to embodiment to the present invention make into
The detailed description of one step, it should be understood that implementation example described herein is merely to illustrate and explain the present invention, and is not used to limit
The fixed present invention.
Based on the echo cancelltion design method that circular microphone array beams are formed, mainly solving the prior art can not be accurate
The problem of gain signal source and inhibition noise source.Its realization process is:
(1) the sound field impulse response in room is obtained according to sef-adapting filter;
(2) loud speaker plays calibration sound, designs corresponding steering vector, circular microphone array is made to recognize interference source side
To;
(3) weight coefficient of spatial filter is designed according to gained steering vector, weighting obtains circular microphone array most
Excellent directional diagram;
(4) it is filtered using the Subband adaptive filters of closed loop configuration in time domain, offsets the echo signal of loud speaker.
The calibration sound that this method is released according to loud speaker first obtains microphone using convex optimization method, matrix disassembling method
Then the covariance matrix of array obtains the angle information of interference signal, design corresponding steering vector, is oriented to and is sweared according to gained
The weight coefficient of amount design spatial filter so that the weight coefficient on interference information source direction is small as possible, to inhibit to acquire to greatest extent
The interference signal arrived.The impulse response time in acoustic echo path is very long, and can reach 200ms for general room arrives
If 300ms needs thousands of ranks are even more to can be only achieved simulation actual ghosts path impulse using auto-adaptive fir filter
The requirement of receptance function can divide the broadband signal of input so realizing time-domain filtering using sub-band adaptive filtering method
The subband signal that a small amount of source signal is carried for several implements step to reduce the complexity of system processing
Including as follows:
1) according to meeting room space, determine that the Homogeneous Circular microphone array of N array numbers, N number of array element composition periphery are low
Sample rate circular array, for calculating sound bearing angle, independent one high sampling array element is located at circle battle array center, high-quality for acquiring
Measure voice signal.The calibration sound signal relationship played using array received signal and loud speaker, adaptively obtains room sound field mould
The impulse response of type.If describing room sound field channel model with linear time invariant system, transmitting signal s (t) receives signal
X (t) can be expressed as the convolution form of transmitting signal and sound field channel impulse response h (t), i.e.,:
X (t)=s (t) * h (t)+n (t) (1)
Wherein * indicates that convolution algorithm, n (t) are noise.The purpose of channel estimation is exactly the condition in known s (t) and x (t)
It is lower to solve specific shock response h (t), for receiving the voice signal that signal estimation loud speaker is sent out by microphone array.
In ray sound-field model, the shock response of channel may be considered to be determined by a series of intrinsic sound rays, and at this moment impulse response is:
Wherein, M is the number of intrinsic sound ray, AiAnd τiIt is i-th of intrinsic corresponding decay factor and propagation delay respectively.
2) array element minimum spacingλ is the corresponding wavelength of information source maximum frequency.Meet spatial sampling and determines rate.
In order to calculate azimuth where sound source, to justify battle array dot as in the three-dimensional coordinate of origin, any two array element
Coordinate can be expressed as Bi(xi,yi,zi), Bj(xj,yj,zj), shown is the geometric representation of information source angle-of- arrival estimation, three
In dimension coordinate Oxyz, s is information source arrival direction, and α indicates that horizontal azimuth, β indicate vertical elevation, all array element positions of uniform circular array
In x/y plane, the vertical line that a wherein array element Bi is s is crossed, s is handed over to reach microphone array since sound source is far field model at this time in A points
It is plane wave when row, so it is OA that sound wave, which reaches Bi array elements with the range difference for reaching origin O points, it can thus be concluded that Bi and O point phases
DifferenceIt can show that the phase difference of origin and arbitrary two array element, formula 8 are what Fig. 3 was derived by the geometrical relationship of array
Information source direction and array element geometrical relationship:
Formula (9), (10) are center of circle array element and i-th and j-th of array element received signal in Homogeneous Circular microphone array
Phase difference, be transformed into time domain be signal reach array element time delay.
(i, j=1,2,3...N)
The signal phase difference expression formula that any two array element receives may finally be determined by formula (9), (10) simultaneous:
The phase difference of same information source is received by calculating array element two-by-two, you can obtain one group of horizontal azimuth and vertical elevation
(α, β), the i.e. spatial information (si) of information source, in the present invention, this step is for calculating the orientation for fixing loud speaker in space.Phase
Difference can be solved by broad sense cross-correlation function.
Broad sense cross-correlation function method is by finding out the crosspower spectrum between two signals first, one then being given in frequency domain
Fixed weighting carries out whitening processing with this to signal and noise, to enhance the higher frequency content of signal-to-noise ratio in signal, inhibits
The influence of noise, finally inverse transformation obtains the cross-correlation function between two signals, i.e., to time domain again:
BiWith BjIndicate the voice signal forms of time and space that i-th and j-th of microphone receive.In conjunction with geometry shown in Fig. 5
Structure receives signal simultaneous solution to arbitrary three groups of microphones, where deriving sound bearing, can orient loud speaker reference note
Orientation.WhereinIt is a symmetric function related with τ, there is unique peak value, pass through the corresponding horizontal seat of search peak
It marks, the delay, τ of available two signals, in formulaFor cross-spectral density function, ψij(ω) is the weighting of broad sense cross-correlation
Function.
3) optimal beam forming directional diagram, the spatial domain weight coefficient method for solving of uniform circular array are as follows in order to obtain:
The mathematical model of microphone array output signal data:
Rx=E[x(t)xH(t)]=ARxAH+Rn (13)
RnIt is noise covariance matrix, RxIt is signal covariance matrix, A is array steering vector.(·)HIndicate that conjugation turns
It sets.Spatial domain steering vector:
Wherein, τijFor i and j, array element receives the time delay of signal, { ω two-by-twoij, i, j=1,2 ..., N } it is that spatial domain is weighed.
And then, the output of array can be obtained in the weighting output of all array elements after being added:
Wherein, { } * indicates conjugation.
The directional diagram of array is defined as:
P (θ)=s |ωHA(θ)(16)
Different weight vectors can make the signal on different directions have different responses, to form the space wave of different directions
Beam.
Realize that Spatially adaptive filtering, array received signal are expressed as x (t)=x using maximum signal noise ratio principles(t)+
xn(t), xs(t) it is corresponding useful signal part, xn(t) it is interference and noise section.
Then the output of array is after Wave beam forming:
Y (t)=ωHX (t)=ωHxs(t)+ωHxn(t) (17)
Adaptive weighted vector ω is acquired by maximum signal noise ratio principle:
WhereinFor signal covariance matrix,For noise jamming covariance square
Battle array.Make the maximum optimal weight vector ω of above formula output signal-to-noise ratiooptIt is matrix to (RS,Rn) the corresponding spy of maximum generalized characteristic value
Sign vector.
For example, far field there are one be located at θ0The signal source of angle, then have
Obtained maximum signal noise ratio principle isα is and θ in formula0Unrelated constant.
4) processing in time domain is filtered using the Subband adaptive filters of closed loop configuration, and essence is the voice that will be received
Signal is divided into several subsignals on frequency spectrum, and the method makes all subband convergence rates that can all improve, and computation complexity
Also it is greatly reduced, this just improves computational efficiency to a certain extent.
The sub-filter algorithm renewal equation of closed loop configuration is:
Input signal extremely closes weight relative to Δ is postponed existing for desired signal, to the compensation of delay in closed loop sub-band structure
It wants.
L is the subfilter length , [ of analysis and composite filter group;·]Indicate round numbers part.
Inventive microphone array and loud speaker orientation diagram are as shown in Figure 2.The performance parameter of inventive microphone array is such as
Shown in following table;
Referring to Fig.1, steps are as follows for realization of the invention:
Step 1, it is assumed that two loud speakers (having secondary reflection) and two tellers determine that uniform circular array row element number of array is
7=2*2+2+1, the number that can obtain forming wave beam is 6.
Step 2, after fixing microphone array column position, the specific calibration audio of loud speaker output, with above-mentioned offer
Azimuth method of estimation positions the orientation of two loud speakers interference and corresponding four secondary reflection interference sources.
Step 3, the corresponding weighting coefficient in interference source direction is calculated using above-mentioned Adaptive beamformer method so that side
Four null angles are generated on collection of illustrative plates, to reduce the gain on interference source direction.
After system enters operating mode, microphone array starts the voice signal of collection site people, is denoted as x (t), with identification
It is the same to calibrate sound direction, azimuth information where array real-time estimation teller, Adaptive beamformer updates weighting coefficient
ωij, so that the main lobe of direction collection of illustrative plates is directed toward voice source direction, to improve the voice signal gain of teller.
It should be understood that the part that this specification does not elaborate belongs to the prior art.
It should be understood that the above-mentioned description for preferred embodiment is more detailed, can not therefore be considered to this
The limitation of invention patent protection range, those skilled in the art under the inspiration of the present invention, are not departing from power of the present invention
Profit requires under protected ambit, can also make replacement or deformation, each fall within protection scope of the present invention, this hair
It is bright range is claimed to be determined by the appended claims.
Claims (5)
1. the echo cancelltion design method formed based on circular microphone array beams, which is characterized in that the method includes such as
Lower step:
(1) according to the principle of sef-adapting filter, the sound field impulse in room is estimated using the IMAGE methods based on ray model
Response, that is, loud speaker provide to be canceled raise to the space echo path between microphone array for subsequent echo cancelltion
Sound device interference signal;
(2) loud speaker plays calibration sound, and circular microphone array makes circle according to direction of arrival estimation method of the TDOA based on time delay
Microphone array recognizes loud speaker interference source direction;
(3) orientation that teller is recognized according to direction of arrival estimation method of the TDOA based on time delay is interfered and is said using loud speaker
The azimuth information for talking about people, obtains array steering vector, and the weight coefficient of spatial filter is then designed according to steering vector, is weighted
To the spatial domain optimal direction figure of circular microphone array, the main lobe direction of array is made to be directed toward teller, and loud speaker interference radiating way
Positioned at array low sidelobe, the purpose of airspace filter is to make the voice signal of microphone array acquisition teller, gain signal source
Inhibit the signal in interference source direction simultaneously;
(4) airspace filter has cut off the acoustics circuit of loud speaker and microphone array, remaining residual echo signal by time domain into
Row processing, so being filtered using the Subband adaptive filters of closed loop configuration in time domain, is supported since voice signal is broadband signal
Disappear the residual echo signal of loud speaker.
2. the echo cancelltion design method formed as described in claim 1 based on circular microphone array beams, feature are existed
In the specific implementation process of the step (1) is:If describing room sound field channel model, sound with linear time invariant system
Source signal is s (t), therefore microphone array reception signal x (t) can be expressed as s (t) and sound field channel impulse response h (t)
Convolution form, i.e.,:
X (t)=s (t) * h (t)+n (t) (1)
Wherein * indicates that convolution algorithm, n (t) are noise, and the purpose of channel estimation is exactly to be asked under conditions of known s (t) and x (t)
Shock response h (t) is solved, in ray sound-field model, the shock response of channel may be considered to be determined by a series of intrinsic sound rays
, at this moment impulse response is:
Wherein, M is the number of intrinsic sound ray, AiAnd τiIt is i-th of intrinsic corresponding decay factor and propagation delay respectively.
3. the echo cancelltion design method formed as described in claim 1 based on circular microphone array beams, feature are existed
In the realization process of the step (2) is:According to the geometric format of array, in conjunction with TDOA methods, the sound of stationary sound source passes through
Room propagated, due to the geometric format of microphone array, time delay that there are one the voice signals that microphone receives two-by-two, this
A time delay can be solved according to cross-correlation function, then to arbitrarily microphone receives signal simultaneous solution two-by-two, derivation is spoken
Where the orientation of source, loudspeaker calibration sound orientation can be oriented.
4. the echo cancelltion design method formed as described in claim 1 based on circular microphone array beams, feature are existed
In the specific implementation process of the step (3) is:Find obtain after sound bearing one group of Space Angle information level azimuth and
The steric direction vector A of array, () then can be obtained in vertical elevation (α, β)HIndicate conjugate transposition;
Wherein, τijFor i and j, array element receives the time delay of signal, { ω two-by-twoij, i, j=1,2 ..., N } it is that spatial domain is weighed;
Spatial domain weighting is carried out per voice signal all the way to microphone array, obtains optimal direction figure;
The directional diagram of array is defined as:
P (θ)=s |ωHA(θ)| (4)
Different weight vectors can make the signal on different directions have different responses, to form the spatial beams of different directions,
So that beam main lobe is directed toward effective Sounnd source direction, null beam position interferes Sounnd source direction;
Realize that Spatially adaptive filtering, array received signal are expressed as using maximum signal noise ratio principle
X (t)=xs(t)+xn(t) (5)
xs(t) it is corresponding useful signal part, xn(t) it is interference and noise section;
Then the output of array is after Wave beam forming:
Y (t)=ωHX (t)=ωHxs(t)+ωHxn(t) (6)。
5. the echo cancelltion design method formed as described in claim 1 based on circular microphone array beams, feature are existed
In the specific implementation process of the step (4) is:Processing in time domain is filtered using the Subband adaptive filters of closed loop configuration
Wave, essence is that the voice signal that will be received is divided into several subsignals on frequency spectrum, in each subband, most using normalization
Small square adaptive algorithm so that subband mean square error is minimum;
The sub-filter algorithm renewal equation of the closed loop configuration of core is:
Input signal is most important in closed loop sub-band structure to the compensation of delay relative to postponing Δ existing for desired signal;
Wherein,L is the subfilter length , [ of analysis and composite filter group;·]Indicate round numbers part.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810304397.1A CN108694957B (en) | 2018-04-08 | 2018-04-08 | Echo cancellation design method based on circular microphone array beam forming |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810304397.1A CN108694957B (en) | 2018-04-08 | 2018-04-08 | Echo cancellation design method based on circular microphone array beam forming |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108694957A true CN108694957A (en) | 2018-10-23 |
| CN108694957B CN108694957B (en) | 2021-08-31 |
Family
ID=63845425
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810304397.1A Active CN108694957B (en) | 2018-04-08 | 2018-04-08 | Echo cancellation design method based on circular microphone array beam forming |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108694957B (en) |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109782228A (en) * | 2019-02-21 | 2019-05-21 | 哈尔滨工程大学 | More positioning nodes of transient signal combine waveform estimation method |
| CN110133578A (en) * | 2019-05-08 | 2019-08-16 | 西北工业大学 | A kind of sub-bottom reflection sound ray incident angle estimation method based on semicolumn volume array |
| CN110211600A (en) * | 2019-05-17 | 2019-09-06 | 北京华控创为南京信息技术有限公司 | For orienting the intelligent microphone array module for monitoring communication |
| CN110364176A (en) * | 2019-08-21 | 2019-10-22 | 百度在线网络技术（北京）有限公司 | Audio signal processing method and device |
| CN110379439A (en) * | 2019-07-23 | 2019-10-25 | 腾讯科技（深圳）有限公司 | A kind of method and relevant apparatus of audio processing |
| CN110415720A (en) * | 2019-07-11 | 2019-11-05 | 湖北工业大学 | The constant Beamforming Method of the super directional frequency of quaternary difference microphone array |
| CN110456309A (en) * | 2019-02-28 | 2019-11-15 | 深圳壹秘科技有限公司 | Sound localization method, device and computer readable storage medium |
| CN110493690A (en) * | 2019-08-29 | 2019-11-22 | 北京搜狗科技发展有限公司 | A kind of sound collection method and device |
| CN110517703A (en) * | 2019-08-15 | 2019-11-29 | 北京小米移动软件有限公司 | A kind of sound collection method, device and medium |
| CN111273230A (en) * | 2020-03-02 | 2020-06-12 | 开放智能机器（上海）有限公司 | Sound source positioning method |
| CN111327984A (en) * | 2020-02-27 | 2020-06-23 | 北京声加科技有限公司 | Earphone auxiliary listening method based on null filtering and ear-worn equipment |
| CN111651719A (en) * | 2020-06-04 | 2020-09-11 | 湖北工业大学 | Robust adaptive beam forming method based on contraction estimation covariance matrix reconstruction |
| CN111785290A (en) * | 2020-05-18 | 2020-10-16 | 深圳市东微智能科技股份有限公司 | Microphone array voice signal processing method, device, equipment and storage medium |
| CN111863012A (en) * | 2020-07-31 | 2020-10-30 | 北京小米松果电子有限公司 | Audio signal processing method and device, terminal and storage medium |
| CN111866439A (en) * | 2020-07-21 | 2020-10-30 | 厦门亿联网络技术股份有限公司 | Conference device and system for optimizing audio and video experience and operation method thereof |
| CN113329288A (en) * | 2021-04-29 | 2021-08-31 | 开放智能技术(南京)有限公司 | Bluetooth headset noise reduction method based on notch technology |
| CN113406561A (en) * | 2021-05-31 | 2021-09-17 | 中国电子科技集团公司第三十六研究所 | Direction finding method and device |
| CN113645546A (en) * | 2020-05-11 | 2021-11-12 | 阿里巴巴集团控股有限公司 | Voice signal processing method and system and audio and video communication equipment |
| CN113782046A (en) * | 2021-09-09 | 2021-12-10 | 清华大学 | Microphone array pickup method and system for remote speech recognition |
| CN114268883A (en) * | 2021-11-29 | 2022-04-01 | 苏州君林智能科技有限公司 | Method and system for selecting microphone placement position |
| CN114913868A (en) * | 2022-05-17 | 2022-08-16 | 电子科技大学 | FPGA-based acoustic array directional pickup method |
| WO2023125537A1 (en) * | 2021-12-31 | 2023-07-06 | 华为技术有限公司 | Sound signal processing method and apparatus, and device and storage medium |
| CN116567515A (en) * | 2023-07-11 | 2023-08-08 | 无锡聚诚智能科技有限公司 | Microphone array calibration method |
| CN116582803A (en) * | 2023-06-01 | 2023-08-11 | 广州市声讯电子科技股份有限公司 | Self-adaptive control method, system, storage medium and terminal for loudspeaker array |
| CN117956370A (en) * | 2024-03-26 | 2024-04-30 | 苏州声学产业技术研究院有限公司 | Dynamic sound pointing method and system based on linear loudspeaker array |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101682808A (en) * | 2007-06-04 | 2010-03-24 | 雅马哈株式会社 | Acoustic device |
| CN103856871A (en) * | 2012-12-06 | 2014-06-11 | 华为技术有限公司 | Device and method for collecting multi-channel sound through microphone array |
| CN105355210A (en) * | 2015-10-30 | 2016-02-24 | 百度在线网络技术（北京）有限公司 | Preprocessing method and device for far-field speech recognition |
| US9508357B1 (en) * | 2014-11-21 | 2016-11-29 | Apple Inc. | System and method of optimizing a beamformer for echo control |
| CN106611600A (en) * | 2016-12-02 | 2017-05-03 | 广州音书科技有限公司 | Audio processing device and system for far-field pickup and mobile charging |
| US9659576B1 (en) * | 2016-06-13 | 2017-05-23 | Biamp Systems Corporation | Beam forming and acoustic echo cancellation with mutual adaptation control |
| CN106935246A (en) * | 2015-12-31 | 2017-07-07 | 芋头科技（杭州）有限公司 | A kind of voice acquisition methods and electronic equipment based on microphone array |
| CN206322472U (en) * | 2016-12-29 | 2017-07-11 | 厦门大洋通信有限公司 | A kind of ZigBee voice controllers |
| CN107248413A (en) * | 2017-03-19 | 2017-10-13 | 临境声学科技江苏有限公司 | Hidden method for acoustic based on Difference Beam formation |
| CN108447479A (en) * | 2018-02-02 | 2018-08-24 | 上海大学 | The robot voice control system of noisy work condition environment |
-
2018
- 2018-04-08 CN CN201810304397.1A patent/CN108694957B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101682808A (en) * | 2007-06-04 | 2010-03-24 | 雅马哈株式会社 | Acoustic device |
| CN103856871A (en) * | 2012-12-06 | 2014-06-11 | 华为技术有限公司 | Device and method for collecting multi-channel sound through microphone array |
| US9508357B1 (en) * | 2014-11-21 | 2016-11-29 | Apple Inc. | System and method of optimizing a beamformer for echo control |
| CN105355210A (en) * | 2015-10-30 | 2016-02-24 | 百度在线网络技术（北京）有限公司 | Preprocessing method and device for far-field speech recognition |
| CN106935246A (en) * | 2015-12-31 | 2017-07-07 | 芋头科技（杭州）有限公司 | A kind of voice acquisition methods and electronic equipment based on microphone array |
| US9659576B1 (en) * | 2016-06-13 | 2017-05-23 | Biamp Systems Corporation | Beam forming and acoustic echo cancellation with mutual adaptation control |
| CN106611600A (en) * | 2016-12-02 | 2017-05-03 | 广州音书科技有限公司 | Audio processing device and system for far-field pickup and mobile charging |
| CN206322472U (en) * | 2016-12-29 | 2017-07-11 | 厦门大洋通信有限公司 | A kind of ZigBee voice controllers |
| CN107248413A (en) * | 2017-03-19 | 2017-10-13 | 临境声学科技江苏有限公司 | Hidden method for acoustic based on Difference Beam formation |
| CN108447479A (en) * | 2018-02-02 | 2018-08-24 | 上海大学 | The robot voice control system of noisy work condition environment |
Non-Patent Citations (2)
| Title |
|---|
| JENS MEYER ETC: "Spherical harmonic modal beamforming for an augmented circular microphone array", 《ACOUSTICS,SPEECH AND SIGNALPROCESSING,2008,ICASSP 2008.IEEE INTERNATIONAL CONFERENCE ON》 * |
| 张晓永: "数字助听器的自适应回波抵消算法的研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (40)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109782228A (en) * | 2019-02-21 | 2019-05-21 | 哈尔滨工程大学 | More positioning nodes of transient signal combine waveform estimation method |
| CN109782228B (en) * | 2019-02-21 | 2022-08-23 | 哈尔滨工程大学 | Multi-positioning-node joint waveform estimation method for transient signal |
| CN110456309B (en) * | 2019-02-28 | 2021-06-08 | 深圳壹秘科技有限公司 | Sound source positioning method, device and computer readable storage medium |
| CN110456309A (en) * | 2019-02-28 | 2019-11-15 | 深圳壹秘科技有限公司 | Sound localization method, device and computer readable storage medium |
| CN110133578A (en) * | 2019-05-08 | 2019-08-16 | 西北工业大学 | A kind of sub-bottom reflection sound ray incident angle estimation method based on semicolumn volume array |
| CN110133578B (en) * | 2019-05-08 | 2023-02-28 | 西北工业大学 | Seabed reflection sound ray incident angle estimation method based on semi-cylindrical volume array |
| CN110211600A (en) * | 2019-05-17 | 2019-09-06 | 北京华控创为南京信息技术有限公司 | For orienting the intelligent microphone array module for monitoring communication |
| CN110211600B (en) * | 2019-05-17 | 2021-08-03 | 北京华控创为南京信息技术有限公司 | Intelligent microphone array module for directional monitoring communication |
| CN110415720A (en) * | 2019-07-11 | 2019-11-05 | 湖北工业大学 | The constant Beamforming Method of the super directional frequency of quaternary difference microphone array |
| CN110415720B (en) * | 2019-07-11 | 2020-05-12 | 湖北工业大学 | Quaternary differential microphone array super-directivity frequency-invariant beam forming method |
| CN110379439B (en) * | 2019-07-23 | 2024-05-17 | 腾讯科技（深圳）有限公司 | Audio processing method and related device |
| CN110379439A (en) * | 2019-07-23 | 2019-10-25 | 腾讯科技（深圳）有限公司 | A kind of method and relevant apparatus of audio processing |
| CN110517703A (en) * | 2019-08-15 | 2019-11-29 | 北京小米移动软件有限公司 | A kind of sound collection method, device and medium |
| CN110517703B (en) * | 2019-08-15 | 2021-12-07 | 北京小米移动软件有限公司 | Sound collection method, device and medium |
| CN110364176A (en) * | 2019-08-21 | 2019-10-22 | 百度在线网络技术（北京）有限公司 | Audio signal processing method and device |
| CN110493690A (en) * | 2019-08-29 | 2019-11-22 | 北京搜狗科技发展有限公司 | A kind of sound collection method and device |
| CN111327984A (en) * | 2020-02-27 | 2020-06-23 | 北京声加科技有限公司 | Earphone auxiliary listening method based on null filtering and ear-worn equipment |
| CN111327984B (en) * | 2020-02-27 | 2022-04-22 | 北京声加科技有限公司 | Earphone auxiliary listening method based on null filtering and ear-worn equipment |
| CN111273230A (en) * | 2020-03-02 | 2020-06-12 | 开放智能机器（上海）有限公司 | Sound source positioning method |
| CN111273230B (en) * | 2020-03-02 | 2022-06-07 | 开放智能机器（上海）有限公司 | Sound source positioning method |
| CN113645546A (en) * | 2020-05-11 | 2021-11-12 | 阿里巴巴集团控股有限公司 | Voice signal processing method and system and audio and video communication equipment |
| CN113645546B (en) * | 2020-05-11 | 2023-02-28 | 阿里巴巴集团控股有限公司 | Voice signal processing method and system and audio and video communication equipment |
| CN111785290A (en) * | 2020-05-18 | 2020-10-16 | 深圳市东微智能科技股份有限公司 | Microphone array voice signal processing method, device, equipment and storage medium |
| CN111785290B (en) * | 2020-05-18 | 2023-12-26 | 深圳市东微智能科技股份有限公司 | Microphone array voice signal processing method, device, equipment and storage medium |
| CN111651719A (en) * | 2020-06-04 | 2020-09-11 | 湖北工业大学 | Robust adaptive beam forming method based on contraction estimation covariance matrix reconstruction |
| CN111651719B (en) * | 2020-06-04 | 2023-04-25 | 湖北工业大学 | Robust self-adaptive beam forming method based on shrinkage estimation covariance matrix reconstruction |
| CN111866439A (en) * | 2020-07-21 | 2020-10-30 | 厦门亿联网络技术股份有限公司 | Conference device and system for optimizing audio and video experience and operation method thereof |
| CN111866439B (en) * | 2020-07-21 | 2022-07-05 | 厦门亿联网络技术股份有限公司 | Conference device and system for optimizing audio and video experience and operation method thereof |
| CN111863012A (en) * | 2020-07-31 | 2020-10-30 | 北京小米松果电子有限公司 | Audio signal processing method and device, terminal and storage medium |
| CN113329288A (en) * | 2021-04-29 | 2021-08-31 | 开放智能技术(南京)有限公司 | Bluetooth headset noise reduction method based on notch technology |
| CN113406561A (en) * | 2021-05-31 | 2021-09-17 | 中国电子科技集团公司第三十六研究所 | Direction finding method and device |
| CN113782046A (en) * | 2021-09-09 | 2021-12-10 | 清华大学 | Microphone array pickup method and system for remote speech recognition |
| CN114268883A (en) * | 2021-11-29 | 2022-04-01 | 苏州君林智能科技有限公司 | Method and system for selecting microphone placement position |
| WO2023125537A1 (en) * | 2021-12-31 | 2023-07-06 | 华为技术有限公司 | Sound signal processing method and apparatus, and device and storage medium |
| CN114913868A (en) * | 2022-05-17 | 2022-08-16 | 电子科技大学 | FPGA-based acoustic array directional pickup method |
| CN116582803B (en) * | 2023-06-01 | 2023-10-20 | 广州市声讯电子科技股份有限公司 | Self-adaptive control method, system, storage medium and terminal for loudspeaker array |
| CN116582803A (en) * | 2023-06-01 | 2023-08-11 | 广州市声讯电子科技股份有限公司 | Self-adaptive control method, system, storage medium and terminal for loudspeaker array |
| CN116567515B (en) * | 2023-07-11 | 2023-09-22 | 无锡聚诚智能科技有限公司 | Microphone array calibration method |
| CN116567515A (en) * | 2023-07-11 | 2023-08-08 | 无锡聚诚智能科技有限公司 | Microphone array calibration method |
| CN117956370A (en) * | 2024-03-26 | 2024-04-30 | 苏州声学产业技术研究院有限公司 | Dynamic sound pointing method and system based on linear loudspeaker array |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108694957B (en) | 2021-08-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108694957A (en) | The echo cancelltion design method formed based on circular microphone array beams | |
| CN108370470B (en) | Conference system and voice acquisition method in conference system | |
| Nishiura et al. | Localization of multiple sound sources based on a CSP analysis with a microphone array | |
| Lockwood et al. | Performance of time-and frequency-domain binaural beamformers based on recorded signals from real rooms | |
| Flanagan et al. | Autodirective microphone systems | |
| US7577266B2 (en) | Systems and methods for interference suppression with directional sensing patterns | |
| Ryan et al. | Array optimization applied in the near field of a microphone array | |
| EP1658751B1 (en) | Audio input system | |
| CN106448722A (en) | Sound recording method, device and system | |
| US20130258813A1 (en) | Apparatus and method for spatially selective sound acquisition by acoustictriangulation | |
| CN106710601A (en) | Voice signal de-noising and pickup processing method and apparatus, and refrigerator | |
| Lockwood et al. | Beamformer performance with acoustic vector sensors in air | |
| JP2013543987A (en) | System, method, apparatus and computer readable medium for far-field multi-source tracking and separation | |
| CN102831898A (en) | Microphone array voice enhancement device with sound source direction tracking function and method thereof | |
| CA2685434A1 (en) | Interference suppression techniques | |
| CN105869651A (en) | Two-channel beam forming speech enhancement method based on noise mixed coherence | |
| CN101828407A (en) | Microphone array processor based on spatial analysis | |
| CN105044675B (en) | A kind of Fast implementation of SRP auditory localizations | |
| Koldovský et al. | Semi-blind noise extraction using partially known position of the target source | |
| Benesty et al. | Array beamforming with linear difference equations | |
| Niwa et al. | Optimal microphone array observation for clear recording of distant sound sources | |
| Shabtai | Optimization of the directivity in binaural sound reproduction beamforming | |
| Lleida et al. | Robust continuous speech recognition system based on a microphone array | |
| Mizumachi | Neural network-based broadband beamformer with less distortion | |
| Shen et al. | A modified cross power-spectrum phase method based on microphone array for acoustic source localization |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |