CN110211601A - A kind of acquisition methods, the apparatus and system of spatial filter parameter matrix - Google Patents

A kind of acquisition methods, the apparatus and system of spatial filter parameter matrix Download PDF

Info

Publication number
CN110211601A
CN110211601A CN201910425432.XA CN201910425432A CN110211601A CN 110211601 A CN110211601 A CN 110211601A CN 201910425432 A CN201910425432 A CN 201910425432A CN 110211601 A CN110211601 A CN 110211601A
Authority
CN
China
Prior art keywords
matrix
microphone
bands
sub
frequencies
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
Application number
CN201910425432.XA
Other languages
Chinese (zh)
Other versions
CN110211601B (en
Inventor
周舒然
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chumen Wenwen Information Technology Co Ltd
Original Assignee
Chumen Wenwen Information Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Chumen Wenwen Information Technology Co Ltd filed Critical Chumen Wenwen Information Technology Co Ltd
Priority to CN201910425432.XA priority Critical patent/CN110211601B/en
Publication of CN110211601A publication Critical patent/CN110211601A/en
Application granted granted Critical
Publication of CN110211601B publication Critical patent/CN110211601B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L21/00Processing of the speech or voice signal to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
    • G10L21/02Speech enhancement, e.g. noise reduction or echo cancellation
    • G10L21/0208Noise filtering
    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L21/00Processing of the speech or voice signal to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
    • G10L21/02Speech enhancement, e.g. noise reduction or echo cancellation
    • G10L21/0208Noise filtering
    • G10L21/0216Noise filtering characterised by the method used for estimating noise
    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L21/00Processing of the speech or voice signal to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
    • G10L21/02Speech enhancement, e.g. noise reduction or echo cancellation
    • G10L21/0208Noise filtering
    • G10L21/0216Noise filtering characterised by the method used for estimating noise
    • G10L2021/02161Number of inputs available containing the signal or the noise to be suppressed
    • G10L2021/02166Microphone arrays; Beamforming

Abstract

The present embodiments relate to technical field of data processing, specifically disclose acquisition methods, the apparatus and system of a kind of spatial filter parameter matrix, this method comprises: determining that first under the first sub-bands of frequencies is directed toward matrix according to the first sub-bands of frequencies matrix, number of microphone, each microphone position and each corresponding angle in microphone position under the first sub-bands of frequencies of pre-acquiring;According to each microphone position, corresponding angle determines the element numerical value in constraint matrix, wherein the element number in constraint matrix is identical as number of microphone;It is directed toward matrix according to constraint matrix and first, determines the spatial filter parameter matrix under the first sub-bands of frequencies.It by the above-mentioned means, may be implemented while the short spacing setting for meeting microphone array, and can guarantee that the microphone array receives the sound of preset direction, reach speech enhan-cement.

Description

A kind of acquisition methods, the apparatus and system of spatial filter parameter matrix
Technical field
The present embodiments relate to technical field of data processing, and in particular to a kind of acquisition of spatial filter parameter matrix Method, apparatus, system and storage medium.
Background technique
Beam forming, need to control the direction of wave beam by the time delay between microphone, and then by being differently directed Wave beam constitutes ideal shape.So that it is determined that beam forming is wanted to receive the direction of sound.By this kind of mode, certain spies are realized The sound for determining direction is all received or part is received, and the sound in certain directions is then all suppressed.Typically, all Repressed sound belongs to " noise " for the sound being received, it is therefore desirable to be suppressed.
And if it is desired to if the time delay between microphone is controlled in strict accordance with presetting time delay, it is necessary to microphone Between there are certain spacing.But microphone array once be applied to earphone in equal volume in lesser equipment when, due to equipment The limitation of volume itself can not conventionally arrange the spacing between microphone array, to realize beam forming.It is short The direct time delay of the microphone array of spacing will inaccuracy or almost nil.That is, if according to traditional wave beam Forming Theory arranges microphone array, will seriously limit the design of many hardware devices.
So, it in order to guarantee to realize the setting of microphone array short spacing in lesser volume range, and can also protect Demonstrate,proving the microphone array can be realized the beam forming of preset state, receive the sound of preset direction, inhibit " making an uproar for other directions Sound ", reaches speech enhan-cement, and most important step is exactly to determine an optimal spatial filter parameter matrix, airspace filter The setting of device parameter matrix directly affects the setting of microphone array, and then influences therefore how beam forming obtains optimal Spatial filter parameter matrix, then become technical problems to be solved in this application.
Summary of the invention
For this purpose, the embodiment of the present invention provides acquisition methods, device, system and the storage of a kind of spatial filter parameter matrix Medium, with solve can not obtain in the prior art optimal beam forming spatial filter parameter matrix, and then cannot be guaranteed compared with While realizing the setting of microphone array short spacing in small volume range, additionally it is possible to it is pre- to guarantee that the microphone array can be realized If the beam forming of state receives the sound of preset direction, the problem of inhibiting " noise " in other directions, reach speech enhan-cement.
To achieve the goals above, the embodiment of the present invention provides the following technical solutions:
In a first aspect, the embodiment of the invention provides a kind of acquisition methods of spatial filter parameter matrix, this method packet It includes:
According to the first sub-bands of frequencies matrix, number of microphone, each microphone under the first sub-bands of frequencies of pre-acquiring Position and each corresponding angle in microphone position determine that first under the first sub-bands of frequencies is directed toward matrix;
According to each microphone position, corresponding angle determines the element numerical value in constraint matrix, wherein constraint Element number in matrix is identical as number of microphone;
It is directed toward matrix according to constraint matrix and first, determines the spatial filter parameter matrix under the first sub-bands of frequencies, In, number of sub-bands is at least one, and the first subband is any one subband at least one subband.
The embodiment of the present invention is further characterized in that, according to the first sub-bands of frequencies square under the first sub-bands of frequencies of pre-acquiring Battle array, number of microphone, each microphone position and each corresponding angle in microphone position determine First under one sub-bands of frequencies is directed toward matrix, specifically includes:
According to each microphone position, determine under the first sub-bands of frequencies, sound field reaches each microphone institute Delay at position;
Delay and i-th of wheat when reaching each microphone position according to the first sub-bands of frequencies matrix, sound field Gram corresponding angle in wind position, determines i-th of direction vector, wherein the initial value of i is 1, i successively progressive value, until I value is the corresponding numerical value of number of microphone;
According to all direction vectors, determine that first under the first sub-bands of frequencies is directed toward matrix.
The embodiment of the present invention is further characterized in that each corresponding angle in microphone position includes pointing direction class Type and inhibition direction type;According to each microphone position, corresponding angle determines the element numerical value in constraint matrix, It specifically includes:
When determining the corresponding angle in the first microphone position is pointing direction type, it is arranged in constraint matrix and the Element numerical value corresponding to one microphone is 1;Alternatively, when determining that the corresponding angle in the first microphone position is to inhibit direction When type, being arranged in constraint matrix with element numerical value corresponding to the first microphone is 0.
The embodiment of the present invention is further characterized in that, is directed toward matrix according to constraint matrix and first, is determined the first sub-bands of frequencies Under spatial filter parameter matrix, specifically include:
Set the initialization matrix of spatial filter parameter;
The initialization matrix for being directed toward matrix, constraint matrix and spatial filter parameter according to first generates cost function;
To cost function solve optimal solution, using the solution of cost function be it is optimal when spatial filter parameter matrix as most Whole spatial filter parameter matrix.
The embodiment of the present invention is further characterized in that, solves optimal solution to cost function, and airspace filter when by optimal solution Device parameter matrix is specifically included as final spatial filter parameter matrix:
According to Lagrangian, convex Optimization Solution is carried out to cost function, when obtaining optimal solution, and will acquire optimal solution Spatial filter parameter matrix as final spatial filter parameter matrix.
Second aspect, the embodiment of the invention also provides a kind of acquisition device of spatial filter parameter matrix, the devices Include:
It is directed toward matrix determination unit, for the first sub-bands of frequencies matrix under the first sub-bands of frequencies according to pre-acquiring, wheat Gram wind quantity, each microphone position and each corresponding angle in microphone position, determine the first subband First under frequency is directed toward matrix;
Constraint matrix determination unit, for being determined in constraint matrix according to each corresponding angle in microphone position Element numerical value, wherein the element number in constraint matrix is identical as number of microphone;
Processing unit determines the airspace filter under the first sub-bands of frequencies for being directed toward matrix according to constraint matrix and first Device parameter matrix, wherein number of sub-bands is at least one, and the first subband is any one subband at least one subband.
The embodiment of the present invention is further characterized in that, is directed toward matrix determination unit and is specifically used for, according to each microphone institute In position, determine under the first sub-bands of frequencies, sound field reaches delay when each microphone position;
Delay and i-th of wheat when reaching each microphone position according to the first sub-bands of frequencies matrix, sound field Gram corresponding angle in wind position, determines i-th of direction vector, wherein the initial value of i is 1, i successively progressive value, until I value is the corresponding numerical value of number of microphone;
According to all direction vectors, determine that first under the first sub-bands of frequencies is directed toward matrix.
The embodiment of the present invention is further characterized in that each corresponding angle in microphone position includes pointing direction class Type and inhibit direction type, constraint matrix determination unit is specifically used for, when determining the corresponding angle in the first microphone position When for pointing direction type, being arranged in constraint matrix with element numerical value corresponding to the first microphone is 1;Alternatively, when determining the The corresponding angle in one microphone position is when inhibiting direction type, to be arranged in constraint matrix and corresponding to the first microphone Element numerical value is 0.
The third aspect, the embodiment of the invention also provides a kind of acquisition system of spatial filter parameter matrix, the systems It include: processor and memory;
Memory is for storing one or more program instructions;
Processor, for running one or more program instructions, to execute a kind of spatial filter parameter matrix as above Acquisition methods in either method step.
Fourth aspect includes in computer storage medium the embodiment of the invention also provides a kind of computer storage medium One or more program instructions, one or more program instructions are used for by a kind of acquisition system of spatial filter parameter matrix A kind of server spatial filter parameter matrix for executing first aspect as above acquisition methods in either method step.
Embodiment according to the present invention has the advantages that the subband frequency under the first sub-bands of frequencies according to pre-acquiring Rate matrix, number of microphone, each microphone position and each corresponding angle in microphone position determine First under first sub-bands of frequencies is directed toward matrix.Then, constraint is determined according to the corresponding angle in each microphone position Element numerical value in matrix;Finally, matrix is directed toward according to constraint matrix and first, determines the airspace filter under the first sub-bands of frequencies Device parameter matrix.Spatial filter parameter matrix, the precisely important parameter of beam forming, and this important parameter, can be with While the short spacing setting for meeting microphone array, and it can guarantee that the microphone array can be realized the wave beam of preset state Forming, receives the sound of preset direction, inhibits " noise " in other directions, reach speech enhan-cement.
Detailed description of the invention
It, below will be to embodiment party in order to illustrate more clearly of embodiments of the present invention or technical solution in the prior art Formula or attached drawing needed to be used in the description of the prior art are briefly described.It should be evident that the accompanying drawings in the following description is only It is merely exemplary, it for those of ordinary skill in the art, without creative efforts, can also basis The attached drawing of offer, which is extended, obtains other implementation attached drawings.
Fig. 1 is a kind of acquisition methods flow diagram for spatial filter parameter matrix that one embodiment of the invention provides;
Fig. 2 be another embodiment of the present invention provides a kind of spatial filter parameter matrix acquisition device structural representation Figure;
Fig. 3 be another embodiment of the present invention provides a kind of spatial filter parameter matrix acquisition system structure signal Figure.
Specific embodiment
Embodiments of the present invention are illustrated by particular specific embodiment below, those skilled in the art can be by this explanation Content disclosed by book is understood other advantages and efficacy of the present invention easily, it is clear that described embodiment is the present invention one Section Example, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not doing Every other embodiment obtained under the premise of creative work out, shall fall within the protection scope of the present invention.
The embodiment of the present invention 1 provides a kind of acquisition methods of spatial filter parameter matrix, specific as shown in Figure 1, should Method can be adapted for the multi-microphone array of earphone, VR glasses or the similar wearable device for having strict demand to volume It is required that electronic equipment.This method comprises:
Step 110, according to the first sub-bands of frequencies matrix, number of microphone, each under the first sub-bands of frequencies of pre-acquiring A microphone position and each corresponding angle in microphone position determine that first under the first sub-bands of frequencies refers to To matrix.
Specifically, the first sub-bands of frequencies matrix, number of microphone, each microphone under the first sub-bands of frequencies obtained Position and each corresponding angle in microphone position, actually according to user demand setting.At one In specific example, frequency of source 0-8000Hz, subband number can be set to 256.First subband is in 256 subbands Any one subband.When actually executing, it is contemplated that under different frequency bands, sound propagation velocity is different, then time delay is also different. Therefore need to calculate the spatial filter parameter matrix under each subband.The acquisition modes of first sub-bands of frequencies matrix are existing Technology does not do excessive explanation here.
And number of microphone, microphone position and its corresponding angle are set according to actual conditions completely, such as wheat Gram wind quantity can be set as 6, and microphone array constitutes a circle, and each microphone position is respectively 0 degree, 60 Degree, 120 degree, 180 degree, 240 degree, the positions such as 300 degree, then, the corresponding angle in microphone position naturally namely on These angles described in text.
Optionally, during specifically executing step 110, it is necessary first to according to each microphone position, determine Under the first sub-bands of frequencies, sound field reaches delay when each microphone position, this process is routine techniques, here not Do excessive description.
Then, delay when each microphone position, Yi Ji are reached according to the first sub-bands of frequencies matrix, sound field The corresponding angle in i microphone position, determines i-th of direction vector.
Specifically, may refer to following formula:
Wherein, d (w, θi) it is i-th of direction vector, w is θiFor the corresponding angle in i-th of microphone position,For TiDelay when i-th of microphone is reached for sound source, if using the first microphone as reference microphone, T1 It is 0.M is the quantity of microphone, in a specific example, M 6.The dimension for being directed toward vector is 1 × 6 dimension.Wherein, i is initial Value is 1, i successively progressive value, until i value is the corresponding numerical value of number of microphone.
First is directed toward matrix, really has multiple direction vectors to constitute, and under the first sub-bands of frequencies, it is directed toward vector Number is identical as microphone number, and when number of microphone is 6, being directed toward vector number is 6, therefore first is directed toward matrix Dimension is 6 × 6 dimensions.
Step 120, according to each microphone position, corresponding angle determines the element numerical value in constraint matrix.
Specifically, the corresponding angle in microphone position includes two types, the first seed type is pointing direction type; Second is inhibition direction type.So, according to each microphone position, corresponding angle is determined in constraint matrix Element numerical value, specifically includes:
When determining the corresponding angle in the first microphone position is pointing direction type, it is arranged in constraint matrix and the Element numerical value corresponding to one microphone is 1;
Alternatively, constraint matrix is arranged when determining the corresponding angle in the first microphone position to inhibit direction type In with element numerical value corresponding to the first microphone be 0.
In fact, each of constraint matrix element represents a constraint, it is corresponding even more to represent microphone position Angle on constraint.For example, when the corresponding angle in first microphone position is the angle that user wishes acquisition sound, It that is to say that angular type is pointing direction type, then, the element of corresponding position will be set as 1 in constraint matrix, the second wheat Gram wind be to the 6th corresponding angle in microphone position inhibit direction type, then, corresponding position in constraint matrix Element will be set as 0.Naturally, the element number in constraint matrix is identical as number of microphone.
Constraint matrix expression formula is as follows at this time:
B=[1 0000 0] ' (formula 2)
Step 130, matrix is directed toward according to constraint matrix and first, determines the spatial filter parameter under the first sub-bands of frequencies Matrix.
Specifically, first is directed toward between matrix and constraint matrix there are certain functional relation, and this functional relation with Spatial filter parameter matrix is related.Referring specifically to formula 3:
Loss function=A × H (w)-B (formula 3)
Wherein, A is the first direction matrix, and B is constraint matrix, and H (w) is spatial filter parameter matrix, and H (w) The matrix exactly taken required by us.Loss function is cost function, is also called loss function.
Ideally, when only loss function is 0, available A × H (the w)=B of.But in actual application, Loss function can not be zero, then just needing in continuous iterative process, allow it to minimize, finally seek H (w).Specifically hold During row, in order to realize iterative process, the initialization matrix h (w) of a spatial filter parameter can be first set.In h (w) Element be it is random generate, h (w) dimension is 6 × 1 dimensions.Then, matrix, constraint matrix and airspace filter are directed toward according to first The initialization matrix of device parameter generates cost function.
To cost function solve optimal solution, using the solution of cost function be it is optimal when spatial filter parameter matrix as most Whole spatial filter parameter matrix.
Optionally, root can carry out convex Optimization Solution to cost function according to Lagrangian, obtain optimal solution, and will Spatial filter parameter matrix when acquisition optimal solution is as final spatial filter parameter matrix.The realization process belongs to existing There is technology, does not do excessive explanation here.
In addition, can also additionally add a restrictive condition, i.e. white noise during carrying out convex Optimization Solution to cost function Sound gain, so that each of filter spatial filter parameter matrix element will not amplify white noise, and then can be to avoid The noise of low-frequency range is amplified.
Moreover, can use the filter spatial filter parameter matrix finally obtained frame by frame after executing the above process The audio signal of microphone acquisition is filtered, and will treated that signal is input to next noise reduction module, to Execute subsequent operation.
It should be noted that above-mentioned calculating obtain optimal spatial filter parameter matrix process can be by counting offline It calculates, that is to say can obtain much covering filter spatial filter parameter matrix in advance, to adapt to the feelings of user's different demands It is called under condition.This design is not rely on the electronic equipment for being equipped with microphone array, any can have place completely The electronic equipment of reason ability is realized.And the electronic equipment for being equipped with microphone array only needs to store corresponding filter airspace filter Wave device parameter matrix both can, without storing other data, that is, for the electronic equipment for being equipped with microphone array, Storage capacity and operational capability require all can suitably reduce.Moreover, have apparent noise reduction effect to noisy noise and interference environment, The final enhancing realized to target voice.
The acquisition methods of a kind of spatial filter parameter matrix provided in an embodiment of the present invention, according to the first of pre-acquiring the son In place with sub-bands of frequencies matrix, number of microphone, each microphone position and each microphone institute under frequency Corresponding angle is set, determines that first under the first sub-bands of frequencies is directed toward matrix.Then, according to each microphone position pair The angle answered determines the element numerical value in constraint matrix;Finally, matrix is directed toward according to constraint matrix and first, determines the first subband Spatial filter parameter matrix under frequency.Spatial filter parameter matrix, the precisely important parameter of beam forming, and this A important parameter can guarantee that the microphone array can be real while the short spacing setting for meeting microphone array The beam forming of existing preset state, receives the sound of preset direction, inhibits " noise " in other directions, reach speech enhan-cement.
Corresponding with above-described embodiment 1, the embodiment of the invention also provides a kind of acquisitions of spatial filter parameter matrix Device, it is specific as shown in Fig. 2, the device includes: to be directed toward matrix determination unit 201, constraint matrix determination unit 202 and processing Unit 203.
Be directed toward matrix determination unit 201, under the first sub-bands of frequencies according to pre-acquiring the first sub-bands of frequencies matrix, Number of microphone, each microphone position and each corresponding angle in microphone position determine the first son Matrix is directed toward with first under frequency;
Constraint matrix determination unit 202, for determining constraint square according to the corresponding angle in each microphone position Element numerical value in battle array, wherein the element number in constraint matrix is identical as number of microphone;
Processing unit 203 determines the airspace filter under the first sub-bands of frequencies for being directed toward matrix according to constraint matrix and first Wave device parameter matrix, wherein number of sub-bands is at least one, and the first subband is any one subband at least one subband.
Optionally, it is directed toward matrix determination unit 201 to be specifically used for, according to each microphone position, determine the Under one sub-bands of frequencies, sound field reaches delay when each microphone position;
Delay and i-th of wheat when reaching each microphone position according to the first sub-bands of frequencies matrix, sound field Gram corresponding angle in wind position, determines i-th of direction vector, wherein the initial value of i is 1, i successively progressive value, until I value is the corresponding numerical value of number of microphone;
According to all direction vectors, determine that first under the first sub-bands of frequencies is directed toward matrix.
Optionally, each corresponding angle in microphone position include pointing direction type and inhibit direction type, Constraint matrix determination unit 202 is specifically used for, when determine the corresponding angle in the first microphone position be pointing direction type When, being arranged in constraint matrix with element numerical value corresponding to the first microphone is 1;Alternatively, when determining that the first microphone institute is in place When setting corresponding angle to inhibit direction type, being arranged in constraint matrix with element numerical value corresponding to the first microphone is 0.
Function performed by each component in a kind of acquisition device of spatial filter parameter matrix provided in an embodiment of the present invention It can be discussed in detail in above-described embodiment 1, therefore not do excessively repeat here.
The acquisition device of a kind of spatial filter parameter matrix provided in an embodiment of the present invention, according to the first of pre-acquiring the son In place with sub-bands of frequencies matrix, number of microphone, each microphone position and each microphone institute under frequency Corresponding angle is set, determines that first under the first sub-bands of frequencies is directed toward matrix.Then, according to each microphone position pair The angle answered determines the element numerical value in constraint matrix;Finally, matrix is directed toward according to constraint matrix and first, determines the first subband Spatial filter parameter matrix under frequency.Spatial filter parameter matrix, the precisely important parameter of beam forming, and this A important parameter can guarantee that the microphone array can be real while the short spacing setting for meeting microphone array The beam forming of existing preset state, receives the sound of preset direction, inhibits " noise " in other directions, reach speech enhan-cement.
Corresponding with above-described embodiment 1, the embodiment of the invention also provides a kind of obtaining for spatial filter parameter matrix System is taken, specifically as shown in figure 3, the system includes processor 301 and memory 302.
Memory 302 is for storing one or more program instructions;
Processor 301, for running one or more program instructions, to execute a kind of spatial filter parameter square as above Method step either in the acquisition methods of battle array.
Function performed by each component in a kind of acquisition system of spatial filter parameter matrix provided in an embodiment of the present invention It can be discussed in detail in above-described embodiment 1, therefore not do excessively repeat here.
The acquisition system of a kind of spatial filter parameter matrix provided in an embodiment of the present invention, according to the first of pre-acquiring the son In place with sub-bands of frequencies matrix, number of microphone, each microphone position and each microphone institute under frequency Corresponding angle is set, determines that first under the first sub-bands of frequencies is directed toward matrix.Then, according to each microphone position pair The angle answered determines the element numerical value in constraint matrix;Finally, matrix is directed toward according to constraint matrix and first, determines the first subband Spatial filter parameter matrix under frequency.Spatial filter parameter matrix, the precisely important parameter of beam forming, and this A important parameter can guarantee that the microphone array can be real while the short spacing setting for meeting microphone array The beam forming of existing preset state, receives the sound of preset direction, inhibits " noise " in other directions, reach speech enhan-cement.
Corresponding with above-described embodiment, the embodiment of the invention also provides a kind of computer storage medium, the computers Include one or more program instructions in storage medium.Wherein, one or more program instructions are used for by a kind of spatial filter Server in the acquisition system of parameter matrix executes a kind of acquisition for the spatial filter parameter matrix introduced such as embodiment 1 Method.
Although above having used general explanation and specific embodiment, the present invention is described in detail, at this On the basis of invention, it can be made some modifications or improvements, this will be apparent to those skilled in the art.Therefore, These modifications or improvements without departing from theon the basis of the spirit of the present invention are fallen within the scope of the claimed invention.

Claims (10)

1. a kind of acquisition methods of spatial filter parameter matrix, which is characterized in that the described method includes:
According to the first sub-bands of frequencies matrix under the first sub-bands of frequencies of pre-acquiring, number of microphone, each microphone place Position and each corresponding angle in microphone position determine that first under first sub-bands of frequencies is directed toward matrix;
The element numerical value in constraint matrix is determined according to each described corresponding angle in microphone position, wherein described Element number in constraint matrix is identical as the number of microphone;
It is directed toward matrix according to the constraint matrix and described first, determines the spatial filter parameter under first sub-bands of frequencies Matrix, wherein the number of sub-bands is at least one, and first subband is any height at least one described subband Band.
2. the method according to claim 1, wherein first under first sub-bands of frequencies according to pre-acquiring Sub-bands of frequencies matrix, the number of microphone, each described microphone position and each microphone position Corresponding angle determines that first under first sub-bands of frequencies is directed toward matrix, specifically includes:
According to each described microphone position, determine under the first sub-bands of frequencies, sound field reaches each microphone institute Delay at position;
According to the first sub-bands of frequencies matrix, the sound field reaches each microphone position when delay and i-th The corresponding angle in a microphone position, determines i-th of direction vector, wherein the initial value of i is 1, i successively progressive value, Until i value is the corresponding numerical value of the number of microphone;
According to all direction vectors, determine that first under first sub-bands of frequencies is directed toward matrix.
3. the method according to claim 1, wherein the corresponding angle packet in each described microphone position It includes pointing direction type and inhibits direction type;The corresponding angle in described each microphone position according to determines about Element numerical value in beam matrix, specifically includes:
When determining the corresponding angle in the first microphone position is pointing direction type, it is arranged in the constraint matrix and institute Stating element numerical value corresponding to the first microphone is 1;
Alternatively, the constraint is arranged when determining the corresponding angle in first microphone position to inhibit direction type In matrix with element numerical value corresponding to first microphone be 0.
4. method according to claim 1-3, which is characterized in that described according to the constraint matrix and described One is directed toward matrix, determines the spatial filter parameter matrix under first sub-bands of frequencies, specifically includes:
Set the initialization matrix of the spatial filter parameter;
The initialization matrix for being directed toward matrix, the constraint matrix and the spatial filter parameter according to described first generates Cost function;
To the cost function solve optimal solution, by the solution of the cost function be it is optimal when spatial filter parameter matrix make For final spatial filter parameter matrix.
5., and will most according to the method described in claim 4, it is characterized in that, described solve optimal solution to the cost function Spatial filter parameter matrix when excellent solution is specifically included as final spatial filter parameter matrix:
According to Lagrangian, convex Optimization Solution is carried out to the cost function, when obtaining optimal solution, and will acquire optimal solution Spatial filter parameter matrix as final spatial filter parameter matrix.
6. a kind of acquisition device of spatial filter parameter matrix, which is characterized in that described device includes:
It is directed toward matrix determination unit, for the first sub-bands of frequencies matrix, the microphone under the first sub-bands of frequencies according to pre-acquiring Quantity, each microphone position and each corresponding angle in microphone position, determine first subband First under frequency is directed toward matrix;
Constraint matrix determination unit, for being determined in constraint matrix according to each described corresponding angle in microphone position Element numerical value, wherein the element number in the constraint matrix is identical as the number of microphone;
Processing unit determines under first sub-bands of frequencies for being directed toward matrix according to the constraint matrix and described first Spatial filter parameter matrix, wherein the number of sub-bands is at least one, and first subband is at least one described subband In any one subband.
7. device according to claim 6, which is characterized in that the direction matrix determination unit is specifically used for, according to institute It states each microphone position, determines under the first sub-bands of frequencies, when sound field reaches each microphone position Delay;
According to the first sub-bands of frequencies matrix, the sound field reaches each microphone position when delay and i-th The corresponding angle in a microphone position, determines i-th of direction vector, wherein the initial value of i is 1, i successively progressive value, Until i value is the corresponding numerical value of the number of microphone;
According to all direction vectors, determine that first under first sub-bands of frequencies is directed toward matrix.
8. device according to claim 6, which is characterized in that the corresponding angle packet in described each microphone position It includes pointing direction type and inhibits direction type, the constraint matrix determination unit is specifically used for, when determining the first microphone institute When in position, corresponding angle is pointing direction type, it is arranged in the constraint matrix and member corresponding to first microphone Prime number value is 1;Alternatively, institute is arranged when determining the corresponding angle in first microphone position to inhibit direction type State in constraint matrix with element numerical value corresponding to first microphone be 0.
9. a kind of acquisition system of spatial filter parameter matrix, which is characterized in that the system comprises: processor and storage Device;
The memory is for storing one or more program instructions;
The processor, for running one or more of program instructions, to execute as described in claim any one of 1-5 Method.
10. a kind of computer storage medium, which is characterized in that refer in the computer storage medium comprising one or more programs It enables, one or more of program instructions are used to be executed by the processor in a kind of acquisition system of spatial filter parameter matrix The method according to claim 1 to 5 step.
CN201910425432.XA 2019-05-21 2019-05-21 Method, device and system for acquiring parameter matrix of spatial filter Active CN110211601B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910425432.XA CN110211601B (en) 2019-05-21 2019-05-21 Method, device and system for acquiring parameter matrix of spatial filter

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910425432.XA CN110211601B (en) 2019-05-21 2019-05-21 Method, device and system for acquiring parameter matrix of spatial filter

Publications (2)

Publication Number Publication Date
CN110211601A true CN110211601A (en) 2019-09-06
CN110211601B CN110211601B (en) 2020-05-08

Family

ID=67787951

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910425432.XA Active CN110211601B (en) 2019-05-21 2019-05-21 Method, device and system for acquiring parameter matrix of spatial filter

Country Status (1)

Country Link
CN (1) CN110211601B (en)

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102214464A (en) * 2010-04-02 2011-10-12 飞思卡尔半导体公司 Transient state detecting method of audio signals and duration adjusting method based on same
CN102440002A (en) * 2009-04-09 2012-05-02 挪威科技大学技术转让公司 Optimal modal beamformer for sensor arrays
CN103181190A (en) * 2010-10-22 2013-06-26 高通股份有限公司 Systems, methods, apparatus, and computer-readable media for far-field multi-source tracking and separation
CN103282961A (en) * 2010-12-21 2013-09-04 日本电信电话株式会社 Speech enhancement method, device, program, and recording medium
CN103632675A (en) * 2012-08-24 2014-03-12 奥迪康有限公司 Noise estimation for use with noise reduction and echo cancellation in personal communication
CN106100608A (en) * 2015-05-01 2016-11-09 中国人民解放军海军大连舰艇学院 Weighted least-squares spatial domain matrix filter design method
CN108447499A (en) * 2018-04-18 2018-08-24 佛山市顺德区中山大学研究院 A kind of double-layer circular ring microphone array voice enhancement method
CN108717495A (en) * 2018-05-22 2018-10-30 出门问问信息科技有限公司 The method, apparatus and electronic equipment of multi-beam beam forming
CN109102822A (en) * 2018-07-25 2018-12-28 出门问问信息科技有限公司 A kind of filtering method and device formed based on fixed beam
CN109616136A (en) * 2018-12-21 2019-04-12 出门问问信息科技有限公司 A kind of Adaptive beamformer method, apparatus and system

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102440002A (en) * 2009-04-09 2012-05-02 挪威科技大学技术转让公司 Optimal modal beamformer for sensor arrays
CN102214464A (en) * 2010-04-02 2011-10-12 飞思卡尔半导体公司 Transient state detecting method of audio signals and duration adjusting method based on same
CN103181190A (en) * 2010-10-22 2013-06-26 高通股份有限公司 Systems, methods, apparatus, and computer-readable media for far-field multi-source tracking and separation
CN103282961A (en) * 2010-12-21 2013-09-04 日本电信电话株式会社 Speech enhancement method, device, program, and recording medium
CN103632675A (en) * 2012-08-24 2014-03-12 奥迪康有限公司 Noise estimation for use with noise reduction and echo cancellation in personal communication
CN106100608A (en) * 2015-05-01 2016-11-09 中国人民解放军海军大连舰艇学院 Weighted least-squares spatial domain matrix filter design method
CN108447499A (en) * 2018-04-18 2018-08-24 佛山市顺德区中山大学研究院 A kind of double-layer circular ring microphone array voice enhancement method
CN108717495A (en) * 2018-05-22 2018-10-30 出门问问信息科技有限公司 The method, apparatus and electronic equipment of multi-beam beam forming
CN109102822A (en) * 2018-07-25 2018-12-28 出门问问信息科技有限公司 A kind of filtering method and device formed based on fixed beam
CN109616136A (en) * 2018-12-21 2019-04-12 出门问问信息科技有限公司 A kind of Adaptive beamformer method, apparatus and system

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
HAOMING LI等: ""Low Complexity Zero-Forcing Beamforming for Distributed Massive MIMO Systems in Large Public Venues"", 《JOURNAL OF COMMUNICATIONS AND NETWORKS》 *
韩东 等: ""零点约束矩阵滤波设计"", 《声学学报》 *

Also Published As

Publication number Publication date
CN110211601B (en) 2020-05-08

Similar Documents

Publication Publication Date Title
US11081123B2 (en) Microphone array-based target voice acquisition method and device
US9591404B1 (en) Beamformer design using constrained convex optimization in three-dimensional space
US9641929B2 (en) Audio signal processing method and apparatus and differential beamforming method and apparatus
CN105355210B (en) Preprocessing method and device for far-field speech recognition
US7274794B1 (en) Sound processing system including forward filter that exhibits arbitrary directivity and gradient response in single wave sound environment
JP4782202B2 (en) Method and apparatus for improving noise discrimination using enhanced phase difference values
KR101060301B1 (en) Method and apparatus for adjusting mismatch of device or signal in sensor array
JP4782201B2 (en) Method and apparatus for improving noise discrimination using attenuation factors
EP1571875A2 (en) A system and method for beamforming using a microphone array
US9106196B2 (en) Sound field spatial stabilizer with echo spectral coherence compensation
JP2013543987A (en) System, method, apparatus and computer readable medium for far-field multi-source tracking and separation
US11381909B2 (en) Method and apparatus for forming differential beam, method and apparatus for processing signal, and chip
US9949034B2 (en) Sound field spatial stabilizer
CN108877828B (en) Speech enhancement method/system, computer-readable storage medium, and electronic device
US9743179B2 (en) Sound field spatial stabilizer with structured noise compensation
CN110211601A (en) A kind of acquisition methods, the apparatus and system of spatial filter parameter matrix
McDonough et al. Microphone arrays
CN110661510B (en) Beam former forming method, beam forming device and electronic equipment
CN113491137B (en) Flexible differential microphone array with fractional order
EP2816818B1 (en) Sound field spatial stabilizer with echo spectral coherence compensation
CN106448693A (en) Speech signal processing method and apparatus
KR102649227B1 (en) Double-microphone array echo eliminating method, device and electronic equipment
EP4138412A1 (en) A method for designing a line array loudspeaker arrangement
Elko et al. Adaptive beamformer for spherical eigenbeamforming microphone arrays
EP2816817A1 (en) Sound field spatial stabilizer with spectral coherence compensation

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