Dual-microphone array echo cancellation method, device and electronic equipment

Abstract

The embodiment of the present application relates to a dual-microphone array echo cancellation method, device and electronic device, the method comprising: obtaining a far-end signal, a first near-end signal and a second near-end signal; performing linear filtering on the far-end signal, the first near-end signal and the second near-end signal; performing centroid fitting on the first initial error spectrum and the second initial error spectrum respectively to obtain the initial suppression gain factor and variable step size factor of the first near-end signal and the second near-end signal; performing residual echo suppression on the first error spectrum and the second error spectrum using an adaptive zero-pole echo canceller to obtain a target spectrum signal; performing sub-band range selection on the initial suppression gain factor to obtain a smoothing factor, and performing full-band smoothing and exponential operations to obtain a secondary suppression gain factor; using the secondary suppression gain factor to filter the target spectrum signal to obtain a target near-end voice signal. The present invention effectively improves the echo cancellation effect and improves the voice call quality.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of echo cancellation, and in particular to a dual-microphone array echo cancellation method, device and electronic device. Background Art

[0002] Acoustic echo cancellation (AEC) technology is widely used in modern communication electronic equipment. During a call between two parties, the near-end microphone will collect the sound of the speaker, so that the far-end can hear its own echo. Therefore, the performance of echo cancellation seriously affects the call quality of both parties.

[0003] During the process of implementing the embodiments of the present application, the inventors of the present application found that in actual environments, due to the influence of factors such as communication equipment and environmental noise, the echo cancellation performance is usually unable to reach the optimal level, resulting in a large residual echo. Summary of the invention

[0004] The purpose of the embodiments of the present application is to provide a dual-microphone array echo cancellation method, device and electronic device, which can effectively improve the echo cancellation effect and improve the voice call quality.

[0005] In a first aspect, an embodiment of the present invention provides a dual-microphone array echo cancellation method, the method comprising:

[0006] Acquire a far-end signal generated by the electronic device during a call, and a first near-end signal and a second near-end signal of the dual-microphone array;

[0007] Performing linear filtering processing on the far-end signal, the first near-end signal, and the second near-end signal to obtain a first initial error spectrum and a second initial error spectrum;

[0008] Performing centroid fitting on the first initial error spectrum and the second initial error spectrum respectively to obtain initial suppression gain factors of the first near-end signal and the second near-end signal, and using the initial suppression gain factors as variable step size factors for linear filtering processing of the next frame;

[0009] Performing residual echo suppression on the first error spectrum and the second error spectrum using an adaptive zero-pole echo canceller to obtain a target spectrum signal;

[0010] After selecting the sub-band range of the initial suppression gain factor, a smoothing factor is obtained, and full-band smoothing and exponential operation are performed to obtain a secondary suppression gain factor;

[0011] The target spectrum signal is filtered using the secondary suppression gain factor to obtain a target near-end speech signal.

[0012] In some embodiments, performing linear filtering on the far-end signal, the first near-end signal, and the second near-end signal to obtain a first initial error spectrum and a second initial error spectrum includes:

[0013] Performing Fourier transform on the far-end signal to obtain far-end frequency domain information, performing Fourier transform on the first near-end signal to obtain first near-end frequency domain information, and performing Fourier transform on the second near-end signal to obtain second near-end frequency domain information;

[0014] Using the filter weight coefficient updated in the previous frame to filter the remote frequency domain information to obtain an echo spectrum;

[0015] Subtracting the echo spectrum from the first near-end frequency domain information to obtain a first initial error spectrum;

[0016] The echo spectrum is subtracted from the second near-end frequency domain information to obtain a second initial error spectrum.

[0017] In some embodiments, performing centroid fitting on the first initial error spectrum and the second initial error spectrum respectively to obtain initial suppression gain factors of the first near-end signal and the second near-end signal includes:

[0018] calculating a first cross-correlation coefficient between the far-end frequency domain information and the first initial error spectrum, and a second cross-correlation coefficient between the far-end frequency domain information and the second initial error spectrum;

[0019] Performing centroid fitting on the first cross-correlation coefficient to obtain a first initial suppression gain factor, and performing centroid fitting on the second cross-correlation coefficient to obtain a second initial suppression gain factor;

[0020] The first initial error spectrum is multiplied by the first initial suppression gain factor to obtain the first error spectrum, and the second initial error spectrum is multiplied by the second initial suppression gain factor to obtain the second error spectrum.

[0021] In some embodiments, the using the initial suppression gain factor as a variable step size factor for linear filtering of the next frame comprises:

[0022] updating the filter weight coefficient of the previous frame based on the first initial suppression gain factor and the second initial suppression gain factor to obtain the filter weight coefficient of the current frame;

[0023] If the first initial suppression gain factor and the second initial suppression gain factor are smaller than a preset threshold, the first initial suppression gain factor and the second initial suppression gain factor are used as variable step size factors for linear filtering processing of the next frame.

[0024] In some embodiments, the using an adaptive pole-zero echo canceller to perform residual echo suppression on the first error spectrum and the second error spectrum to obtain a target spectrum signal includes:

[0025] Calculate a first branch spectrum signal based on the first error spectrum and delay compensation, and calculate a second branch spectrum signal based on the second error spectrum and delay compensation;

[0026] Based on the first branch spectrum signal and the second branch spectrum signal, calculating a differential signal cross-correlation power spectrum coefficient between the first branch spectrum signal and the second branch spectrum signal and a weight coefficient of an adaptive zero-pole echo canceller;

[0027] The weight coefficients of the adaptive pole-zero echo canceller are used to perform residual echo suppression on the first error spectrum and the second error spectrum to obtain a target spectrum signal.

[0028] In some embodiments, after selecting the sub-band range of the initial suppression gain factor, a smoothing factor is obtained, and full-band smoothing and exponential operation are performed to obtain a secondary suppression gain factor, including:

[0029] Selecting a sub-band range for the initial suppression gain factor to obtain initial suppression gain factors within a plurality of sub-band ranges;

[0030] Sorting a plurality of initial suppression gain factors within the sub-band range, and calculating a suppression gain coefficient smoothing factor;

[0031] The initial suppression gain factor is subjected to full-band smoothing processing and exponential operation using the suppression gain coefficient smoothing factor to obtain the secondary suppression gain factor.

[0032] In some embodiments, the filtering process of the target spectrum signal by using the secondary suppression gain factor to obtain the target near-end speech signal includes:

[0033] Performing noise reduction processing on the target spectrum signal;

[0034] After the target spectrum signal after noise reduction processing is multiplied by the secondary suppression gain factor, an inverse Fourier transform is performed, and overlap-addition is performed to obtain a target near-end speech signal.

[0035] In a second aspect, an embodiment of the present invention provides a dual-microphone array echo cancellation device, the device comprising:

[0036] A signal acquisition module, used to acquire a far-end signal generated by the electronic device during a call, and a first near-end signal and a second near-end signal of the dual-microphone array;

[0037] A linear filtering module, configured to perform linear filtering processing on the far-end signal, the first near-end signal, and the second near-end signal to obtain a first initial error spectrum and a second initial error spectrum;

[0038] a centroid fitting module, configured to perform centroid fitting on the first initial error spectrum and the second initial error spectrum respectively, obtain initial suppression gain factors of the first near-end signal and the second near-end signal, and use the initial suppression gain factors as variable step size factors for linear filtering processing of a next frame;

[0039] A residual echo suppression module, used for performing residual echo suppression on the first error spectrum and the second error spectrum by using an adaptive zero-pole echo canceller to obtain a target spectrum signal;

[0040] A smoothing module is used to select a sub-band range for the initial suppression gain factor, obtain a smoothing factor, and perform full-band smoothing and exponential operation to obtain a secondary suppression gain factor;

[0041] The target acquisition module is used to filter the target spectrum signal using the secondary suppression gain factor to obtain a target near-end speech signal.

[0042] In a third aspect, an embodiment of the present invention provides an electronic device, including:

[0043] a dual microphone array and at least one processor, and

[0044] A memory, wherein the memory is communicatively connected to the at least one processor, and the memory stores instructions executable by the at least one processor, and when the instructions are executed by the at least one processor, the electronic device implements any of the methods described above.

[0045] In a fourth aspect, an embodiment of the present invention provides a non-volatile computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions. When the computer-executable instructions are executed by an electronic device, the electronic device executes the method as described above.

[0046] The dual-microphone array echo cancellation method, device and electronic device of the embodiment of the present invention obtain the far-end signal generated by the electronic device during a call, as well as the first near-end signal and the second near-end signal of the dual-microphone array, and then perform initial filtering processing on the far-end signal and the near-end signal of the dual-microphone array. The far-end signal, the first near-end signal and the second near-end signal can be linearly filtered to obtain a first initial error spectrum and a second initial error spectrum, thereby preliminarily eliminating the echo; then, the first initial error spectrum and the second initial error spectrum are respectively centroidally fitted to obtain the initial suppression of the first near-end signal and the second near-end signal. gain factor, and use the initial suppression gain factor as a variable step size factor for the next frame of linear filtering processing; use an adaptive zero-pole echo canceller to perform residual echo suppression on the first error spectrum and the second error spectrum to obtain a target spectrum signal, thereby achieving residual echo suppression; after selecting a sub-band range for the initial suppression gain factor, obtain a smoothing factor, perform full-band smoothing and exponential operations to obtain a secondary suppression gain factor, use the secondary suppression gain factor to filter the target spectrum signal, achieve secondary suppression gain filtering, obtain a target near-end voice signal, effectively improve the echo cancellation effect, and effectively improve the voice call quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] One or more embodiments are exemplarily described by pictures in the corresponding drawings, and these exemplified descriptions do not constitute limitations on the embodiments. Elements with the same reference numerals in the drawings represent similar elements, and unless otherwise stated, the figures in the drawings do not constitute proportional limitations.

[0048] Figure 1 is a schematic block diagram of an electronic device according to an embodiment of the present invention;

[0049] Figure 2 is a circuit structure diagram of an electronic device according to an embodiment of the present invention;

[0050] Figure 3 is a schematic structural diagram of a dual-microphone array of an electronic device according to an embodiment of the present invention;

[0051] Figure 4 It is a flowchart of an embodiment of a dual-microphone array echo cancellation method according to an embodiment of the present invention;

[0052] Figure 5 It is a structural schematic diagram of an embodiment of a dual-microphone array echo cancellation device of the present invention;

[0053] Figure 6 It is a schematic diagram of the hardware structure of a controller in one embodiment of the electronic device of the present invention. DETAILED DESCRIPTION

[0054] The present application is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present application, but do not limit the present application in any form. It should be noted that, for those of ordinary skill in the art, several variations and improvements can also be made without departing from the concept of the present application. These all belong to the protection scope of the present application.

[0055] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0056] It should be noted that, if there is no conflict, the various features in the embodiments of the present application can be combined with each other, all within the scope of protection of the present application. In addition, although the functional module division is performed in the device schematic diagram and the logical order is shown in the flow chart, in some cases, the steps shown or described can be performed in a sequence different from the module division in the device or the flow chart. In addition, the words "first", "second", "third", etc. used herein do not limit the data and execution order, but only distinguish the same items or similar items with basically the same functions and effects.

[0057] Unless otherwise defined, all technical and scientific terms used in this specification have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used in this specification and in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application. The term "and / or" used in this specification includes any and all combinations of one or more of the related listed items.

[0058] In addition, the technical features involved in each embodiment of the present application described below can be combined with each other as long as they do not conflict with each other.

[0059] The dual-microphone array echo cancellation method and device provided in the embodiments of the present application can be applied to electronic devices, which may be smart phones, Bluetooth headsets, wired headsets, and other devices used for making calls. Taking Bluetooth headsets as an example, the Bluetooth headsets are connected to the smart phone that needs to make a call through Bluetooth, and the user wears the Bluetooth headset to make a call.

[0060] like Figure 1 and Figure 2As shown, the electronic device 100 includes a controller 11, a speaker 12 and a dual microphone array 13, and the controller 11 is connected to the speaker 12 and the dual microphone array 13. The speaker 12 is used to play sound, which is a far-end signal; the dual microphone array 13 is used to collect sound, and the sound collected by the dual microphone array 13 includes the user's speaking sound, echo and background noise, which is a near-end signal. In addition, the dual microphone array 13 includes two microphones, namely a first microphone mic1 and a second microphone mic2, and the dual microphone array 13 is an end-fire setting, as shown in FIG. Figure 3 As shown, the dual microphone array 13 adopts an end-fire mode to suppress the sound in the non-gain direction, that is, the sound in the direction of the speaker 12, so as to retain the sound in the direction where the gain is required, that is, the sound in the direction of the near-end user speaking. The controller 11 is used to obtain the far-end signal generated by the user during the call, as well as the first near-end signal of the first microphone mic1 and the second near-end signal of the second microphone mic2.

[0061] The electronic device 100 also includes a linear adaptive filter module 14, an adaptive zero-pole echo canceller 15, a single-channel noise reduction module NR and a suppression gain filter output module 16, and the controller 11 is connected to the linear adaptive filter module 14, the adaptive zero-pole echo canceller 15, the single-channel noise reduction module NR and the suppression gain filter output module 16.

[0062] The linear adaptive filter module 14 is used to perform linear filtering on the far-end signal, the first near-end signal and the second near-end signal to obtain a first initial error spectrum and a second initial error spectrum, and the controller 11 obtains the first initial error spectrum and the second initial error spectrum from the linear adaptive filter module 14 .

[0063] The linear adaptive filter module 14 is also used to perform centroid fitting on the first initial error spectrum and the second initial error spectrum respectively, obtain the initial suppression gain factors of the first near-end signal and the second near-end signal, and use the initial suppression gain factors as variable step factors for the next frame of linear filtering processing. The controller 11 obtains the first error spectrum and the first initial suppression gain factor, the second error spectrum and the second initial suppression gain factor from the linear adaptive filter module 14, and the first initial suppression gain factor and the second initial suppression gain factor constitute the initial suppression gain factor.

[0064] The adaptive zero-pole echo canceller 15 is used to perform residual echo suppression on the first error spectrum and the second error spectrum to obtain a target spectrum signal.

[0065] The suppression gain filter output module 16 is used to select the sub-band range of the initial suppression gain factor, obtain the smoothing factor, and perform full-band smoothing and exponential operation to obtain the secondary suppression gain factor; use the secondary suppression gain factor to filter the target spectrum signal to obtain the target near-end speech signal, and output the target near-end speech signal.

[0066] The electronic device 100 further includes a delay compensation calculation module d / c, which is used to calculate the delay compensation between the far-end signal and the first near-end signal and / or the second near-end signal.

[0067] The controller 11, the linear adaptive filter module 14, the adaptive zero-pole echo canceller 15 and the suppressed gain filter output module 16 perform primary filtering processing, centroid fitting, residual echo suppression and secondary filtering processing on the far-end signal and the near-end signal, so as to realize echo cancellation of the electronic device 100 during a call, solve the problems of poor echo cancellation effect and excessive residual echo, and thus effectively improve the voice call quality.

[0068] See also Figure 4 , Figure 4 1 is a flow chart of a dual-microphone array echo cancellation method provided in an embodiment of the present invention. The method is applied to an electronic device 100 and can be executed by a processor 111 in the electronic device 100. Figure 4 As shown, the method includes:

[0069] S401: Acquire a far-end signal generated by the electronic device during a call, as well as a first near-end signal and a second near-end signal of a dual-microphone array.

[0070] The controller of the electronic device obtains the far-end signal and near-end signal generated by the electronic device during the call, the far-end signal is the sound signal played by the speaker, and the near-end signal includes the user's voice, echo and / or background noise. In addition, the dual microphone array includes a first microphone mic1 and a second microphone mic2, the first microphone generates a first near-end signal, and the second microphone generates a second near-end signal.

[0071] The far-end signal is denoted by x(n), and the far-end signal x(n) is expressed by Formula 1:

[0072] Formula 1;

[0073] The first near-end signal is denoted by d1(n), and the first near-end signal d1(n) is expressed by Formula 2:

[0074] Formula 2;

[0075] The second near-end signal is represented by d2(n), and the second near-end signal d2(n) is expressed by Formula 3:

[0076] Formula 3;

[0077] Wherein, M represents the number of sampling points of a frame of the far-end signal x(n) or the first near-end signal d1(n) or the second near-end signal d2(n), and i represents the i-th sampling point at which the controller of the electronic device collects the far-end signal x(n) or the near-end signal d1(n) or the second near-end signal d2(n).

[0078] S402: Perform linear filtering on the far-end signal, the first near-end signal, and the second near-end signal to obtain a first initial error spectrum and a second initial error spectrum.

[0079] After the controller of the electronic device obtains the far-end signal x(n), the first near-end signal d1(n), and the second near-end signal d2(n), the linear adaptive filter module performs linear filtering processing on the far-end signal x(n), the first near-end signal d1(n), and the second near-end signal d2(n) to obtain the first initial error spectrum and the second initial error spectrum.

[0080] In some implementation modes, performing linear filtering on the far-end signal, the first near-end signal, and the second near-end signal to obtain a first initial error spectrum and a second initial error spectrum includes:

[0081] Performing Fourier transform on the far-end signal to obtain far-end frequency domain information, performing Fourier transform on the first near-end signal to obtain first near-end frequency domain information, and performing Fourier transform on the second near-end signal to obtain second near-end frequency domain information;

[0082] Using the filter weight coefficient updated in the previous frame to filter the remote frequency domain information to obtain an echo spectrum;

[0083] Subtracting the echo spectrum from the first near-end frequency domain information to obtain a first initial error spectrum;

[0084] The echo spectrum is subtracted from the second near-end frequency domain information to obtain a second initial error spectrum.

[0085] Specifically, first, the far-end signal x(n) is Fourier transformed to obtain the far-end frequency domain information X(n), the first near-end signal d1(n) is Fourier transformed to obtain the first near-end frequency domain information D1(n), and the second near-end signal d2(n) is Fourier transformed to obtain the second near-end frequency domain information D2(n). The far-end frequency domain information X(n) is expressed by Formula 4, the first near-end frequency domain information D1(n) is expressed by Formula 5, and the second near-end frequency domain information D2(n) is expressed by Formula 6:

[0086] Formula 4;

[0087] Formula 5;

[0088] Formula 6;

[0089] Among them, X(n) represents the far-end frequency domain information of the nth frame, D1(n) represents the first near-end frequency domain information of the nth frame, D2(n) represents the second near-end frequency domain information of the nth frame, win represents the Hanning window, the length of which is 2*M, and fft represents Fourier transform. The near-end frequency domain information D(n) includes the first near-end frequency domain information D1(n) and the second near-end frequency domain information D2(n).

[0090] Then, the far-end frequency domain information X(n) is filtered using the filter weight coefficient updated in the previous frame to obtain the echo spectrum. Since the echo path and the call process are fixed, the far-end frequency domain information can be delayed and aligned with the first near-end frequency domain information D1(n) and the second near-end frequency domain information D2(n) before filtering, so as to estimate the first echo spectrum Y1(n) and the second echo spectrum Y2(n). The filter weight coefficient updated in the previous frame includes the filter weight coefficient of the first microphone updated in the previous frame. and the filter weight coefficients of the second microphone updated in the previous frame That is, the remote frequency domain information X(n) is filtered in the frequency domain to obtain an echo spectrum represented by Y(n). The echo spectrum Y(n) includes a first echo spectrum Y1(n) and a second echo spectrum Y2(n), which are calculated using Formula 7 and Formula 8:

[0091] Formula 7;

[0092] Formula 8.

[0093] in, represents the filter weight coefficient updated by the first microphone in the previous frame, Represents the filter weight coefficients of the second microphone updated in the previous frame.

[0094] Finally, the echo spectrum Y(n) is subtracted from the near-end frequency domain information D(n) to obtain the first initial error spectrum E1(n) and the second initial error spectrum E2(n), that is, the first echo spectrum Y1(n) is subtracted from the first near-end frequency domain information D1(n) to obtain the first initial error spectrum E1(n); the second echo spectrum Y2(n) is subtracted from the second near-end frequency domain information D2(n) to obtain the second initial error spectrum E2(n). It is expressed by Formula 9 and Formula 10:

[0095] Formula 9;

[0096] Formula 10.

[0097] The obtained first initial error spectrum E1 (n) and the second initial error spectrum E2 (n) can preliminarily eliminate the linear echo of the first microphone and the second microphone.

[0098] S403: Perform centroid fitting on the first initial error spectrum and the second initial error spectrum respectively to obtain initial suppression gain factors of the first near-end signal and the second near-end signal, and use the initial suppression gain factors as variable step size factors for linear filtering processing of the next frame.

[0099] In order to reduce most of the linear echoes of the near-end signal, it is first necessary to calculate the initial suppression gain factor, which includes a first initial suppression gain factor and a second initial suppression gain factor. Performing centroid fitting on the first initial error spectrum and the second initial error spectrum respectively to obtain the initial suppression gain factors of the first near-end signal and the second near-end signal may include:

[0100] calculating a first cross-correlation coefficient between the far-end frequency domain information and the first initial error spectrum, and a second cross-correlation coefficient between the far-end frequency domain information and the second initial error spectrum;

[0101] Performing centroid fitting on the first cross-correlation coefficient to obtain a first initial suppression gain factor, and performing centroid fitting on the second cross-correlation coefficient to obtain a second initial suppression gain factor;

[0102] The first initial error spectrum is multiplied by the first initial suppression gain factor to obtain the first error spectrum, and the second initial error spectrum is multiplied by the second initial suppression gain factor to obtain the second error spectrum.

[0103] Specifically, first, a first cross-correlation coefficient between the far-end frequency domain information X(n) and the first initial error spectrum E1(n) is calculated. Similarly, a second cross-correlation coefficient between the far-end frequency domain information X(n) and the second initial error spectrum E2(n) is calculated.

[0104] Using formulas 11-17, the first correlation coefficient and the second correlation coefficient are calculated.

[0105] Formula 11;

[0106] Formula 12;

[0107] Formula 13;

[0108] Formula 14;

[0109] Formula 15;

[0110] Formula 16;

[0111] Formula 17.

[0112] Among them, gamma represents the smoothing factor, which is generally a number less than 1 but close to 1, such as 0.9; σ represents the division protection factor, which can be set to be greater than 0; conj represents the conjugate operation; a smoothed power spectrum representing the first initial error spectrum; a smoothed power spectrum representing a second initial error spectrum; The smoothed power spectrum representing the far-end frequency domain information, A smoothed power spectrum representing the far-end frequency domain information, the first near-end frequency domain information, and the first initial error spectrum, A smoothed power spectrum representing the far-end frequency domain information, the second near-end frequency domain information, and the second initial error spectrum; represents the first correlation coefficient, represents the second correlation coefficient.

[0113] Next, the first correlation coefficient is subjected to centroid fitting to obtain a first initial suppression gain factor, and the second correlation coefficient is subjected to centroid fitting to obtain a second initial suppression gain factor, which can be calculated using Formula 18 and Formula 19:

[0114] Formula 18;

[0115] Formula 19;

[0116] Where k represents the centroid fitting distance, i represents the i-th frequency point of the n-th frame, represents the first initial suppression gain factor; Represents the second initial suppression gain factor.

[0117] In order to obtain the first initial suppression gain factor and the second initial suppression gain factor Then, the first initial suppression gain factor is used Multiply the first initial error spectrum E1(n) to obtain the first error spectrum, and use the second initial suppression gain factor Multiply by the second initial error spectrum E2(n) to obtain the second error spectrum, which is expressed by Formula 20 and Formula 21:

[0118] Formula 20;

[0119] Formula 21.

[0120] in, represents the first error spectrum, represents the second error spectrum.

[0121] The first error spectrum obtained and the second error spectrum The near-end frequency domain signal of most linear echoes can be reduced.

[0122] In some implementations, in order to improve the accuracy and robustness of the filter, the filter weight coefficient may be updated in real time. Therefore, the initial suppression gain factor may be used as a variable step size factor for the next frame of linear filtering processing, which may include:

[0123] updating the filter weight coefficient of the previous frame based on the first initial suppression gain factor and the second initial suppression gain factor to obtain the filter weight coefficient of the current frame;

[0124] If the first initial suppression gain factor and the second initial suppression gain factor are smaller than a preset threshold, the first initial suppression gain factor and the second initial suppression gain factor are used as variable step size factors for linear filtering processing of the next frame.

[0125] Specifically, based on the first initial suppression gain factor , the second initial suppression gain factor Update the filter weight coefficient of the previous frame , the filter weight coefficient of the previous frame includes the filter weight coefficient of the first microphone updated in the previous frame and the filter weight coefficients of the second microphone updated in the previous frame , obtain the current frame filter weight coefficient, expressed by Formula 22 and Formula 23:

[0126] Formula 22;

[0127] Formula 23.

[0128] in, Indicates the current frame filter weight coefficient of the first microphone mic1, represents the weight coefficient of the current frame filter of the second microphone mic2; μ represents a fixed step size factor, usually in the range of 0<μ<2; represents the filter coefficient adjustment amount of the first microphone mic1, represents the filter coefficient adjustment amount of the second microphone mic2, the calculation method of the filter coefficient adjustment amount includes but is not limited to the normalized least mean square error (NLMS), recursive least squares (RLS) and Kalman algorithms; represents the first initial suppression gain factor, represents the second initial suppression gain factor, and It acts as a variable step size factor for the linear filtering process of the next frame, and performs the step of filtering the remote frequency domain information, so that the filter is more robust.

[0129] Therefore, when the filter has not converged, the residual echo is too large, and the step size becomes larger to help the filter converge quickly; when the filter gradually converges, such as the first initial suppression gain factor , the second initial suppression gain factor When the value of the first initial error spectrum E1(n) and the second initial error spectrum E2(n) gradually decrease, and is less than the preset threshold and gradually approaches 0, for example, when the preset threshold is 0.01, the step size gradually decreases, indicating that the residual echo in the first initial error spectrum E1(n) and the second initial error spectrum E2(n) is small or in an echo-free state. Then, according to step S402, the current frame filter weight coefficient ( and ) performs filtering on the far-end frequency domain information X(n) to obtain the echo spectrum, and then calculates the initial error spectrum.

[0130] S404: Perform residual echo suppression on the first error spectrum and the second error spectrum using an adaptive pole-zero echo canceller to obtain a target spectrum signal.

[0131] In some implementations, performing residual echo suppression on the first error spectrum and the second error spectrum using an adaptive pole-zero echo canceller to obtain a target spectrum signal may include:

[0132] Calculate a first branch spectrum signal based on the first error spectrum and delay compensation, and calculate a second branch spectrum signal based on the second error spectrum and delay compensation;

[0133] Based on the first branch spectrum signal and the second branch spectrum signal, calculating a differential signal cross-correlation power spectrum coefficient between the first branch spectrum signal and the second branch spectrum signal and a weight coefficient of an adaptive zero-pole echo canceller;

[0134] The weight coefficients of the adaptive pole-zero echo canceller are used to perform residual echo suppression on the first error spectrum and the second error spectrum to obtain a target spectrum signal.

[0135] Specifically, first, the spectrum signals of the two branches are calculated, that is, the spectrum signal of the first branch is calculated based on the first error spectrum and delay compensation, and the spectrum signal of the second branch is calculated based on the second error spectrum and delay compensation; then, based on the first branch spectrum signal and the second branch spectrum signal, the differential signal cross-correlation power spectrum coefficient between the first branch spectrum signal and the second branch spectrum signal and the weight coefficient of the adaptive zero-pole echo canceller are calculated, and finally, the weight coefficient of the adaptive zero-pole echo canceller is used to suppress the residual echo of the first error spectrum and the second error spectrum to obtain the target spectrum signal. It can be obtained using formula 24-formula 29:

[0136] Formula 24;

[0137] Formula 25;

[0138] Formula 26;

[0139] Formula 27;

[0140] Formula 28;

[0141] Formula 29.

[0142] in, represents the second branch spectrum signal; represents the spectrum signal of the first branch; delay represents the delay compensation operation, represents the delay compensation of the first branch, Indicates the delay compensation of the second branch; the value range of α is 0≤α≤1, the value range of β is 0≤β≤1, and α+β=1; represents the weight coefficient of the adaptive zero-pole echo canceller, represents the cross-correlation power spectrum coefficient of the differential signal, represents the second branch autocorrelation power spectrum; Represents the target spectrum signal.

[0143] S405: After selecting a sub-band range for the initial suppression gain factor, a smoothing factor is obtained, and full-band smoothing and exponential operations are performed to obtain a secondary suppression gain factor.

[0144] In order to improve the echo cancellation effect, the initial suppression gain factor may be smoothed.

[0145] In some implementation modes, the initial suppression gain factor includes the first initial suppression gain factor and the second initial suppression gain factor. After selecting the sub-band range of the initial suppression gain factor, obtaining the smoothing factor, and performing full-band smoothing and exponential operation to obtain the secondary suppression gain factor may include:

[0146] Selecting a sub-band range for the initial suppression gain factor to obtain a plurality of sub-band ranges;

[0147] Sorting a plurality of initial suppression gain factors within the sub-band range, and calculating a suppression gain coefficient smoothing factor;

[0148] The initial suppression gain factor is subjected to full-band smoothing processing and exponential operation using the suppression gain coefficient smoothing factor to obtain the secondary suppression gain factor.

[0149] Specifically, the subband range may be selected by selecting the subband minimum value minFre, the subband maximum value maxFre and the reference subband point refFre, and then the subband range is sorted, for example, in ascending order, to obtain a sequence hNI, and a reference value Refer is selected according to formula 30, which is as follows:

[0150] Formula 30.

[0151] in, represents the reference value; refFre represents the reference subband point; hNI represents the sequence of subband range from small to large.

[0152] The suppression gain coefficient smoothing factor can be calculated using Formula 31, which is as follows:

[0153] Formula 31.

[0154] in, represents the smoothing factor of the suppression gain coefficient; minFre represents the minimum value of the subband; maxFre represents the maximum value of the subband; refFre represents the reference subband point; i represents the i-th frequency point of the n-th frame.

[0155] In order to obtain the smoothing factor of the suppression gain coefficient Then, the suppression gain coefficient smoothing factor is used The initial suppression gain factor is subjected to full-band smoothing and exponential operation to obtain the secondary suppression gain factor, which is expressed by formula 32:

[0156] Formula 32.

[0157] The initial suppression gain factor includes a first initial suppression gain factor g1(i) and a second initial suppression gain factor g2(i), and G(i) represents a secondary suppression gain factor.

[0158] S406: Filter the target spectrum signal using the secondary suppression gain factor to obtain a target near-end speech signal.

[0159] In order to eliminate the residual echo as much as possible and to avoid the possibility of secondary filtering damaging the near-end signal and obtain the optimal output, the target spectrum signal is filtered using the secondary suppression gain factor to obtain the target near-end speech signal, which may include:

[0160] Performing noise reduction processing on the target spectrum signal;

[0161] After the target spectrum signal after noise reduction processing is multiplied by the secondary suppression gain factor, an inverse Fourier transform is performed, and overlap-addition is performed to obtain a target near-end speech signal.

[0162] Specifically, the target spectrum signal is first subjected to noise reduction processing, and then the target spectrum signal after noise reduction processing is multiplied by the secondary suppression gain factor, and then an inverse Fourier transform is performed, and overlap-addition is performed to obtain the target near-end speech signal. The following is calculated using Formula 33-Formula 36:

[0163] Formula 33;

[0164] Formula 34;

[0165] Formula 35;

[0166] Formula 36.

[0167] in, represents overlap-add retention block, ifft represents inverse Fourier transform; NR represents noise reduction algorithm processing, represents the target spectrum signal after noise reduction processing, represents the error spectrum, including the first error spectrum and the second error spectrum ; Indicates the target near-end speech signal.

[0168] In an embodiment of the present invention, a far-end signal generated by an electronic device during a call, as well as a first near-end signal and a second near-end signal of a dual-microphone array are obtained, and then, the far-end signal and the near-end signal of the dual-microphone array are subjected to initial filtering processing, and the far-end signal, the first near-end signal and the second near-end signal can be subjected to linear filtering processing to obtain a first initial error spectrum and a second initial error spectrum, thereby preliminarily eliminating echoes; then, the first initial error spectrum and the second initial error spectrum are subjected to centroid fitting respectively to obtain initial suppression gain factors of the first near-end signal and the second near-end signal, and the initial suppression gain factors are used as variable step size factors for linear filtering processing of the next frame; residual echo suppression is performed on the first error spectrum and the second error spectrum using an adaptive zero-pole echo canceller to obtain a target spectrum signal, thereby achieving residual echo suppression; after selecting a sub-band range for the initial suppression gain factor, a smoothing factor is obtained, and full-band smoothing and exponential operations are performed to obtain a secondary suppression gain factor, and the target spectrum signal is subjected to filtering processing using the secondary suppression gain factor to achieve secondary suppression gain filtering, thereby obtaining a target near-end voice signal, thereby effectively improving the echo cancellation effect and the voice call quality.

[0169] Correspondingly, such as Figure 5 As shown, the embodiment of the present invention also provides a dual-microphone array echo cancellation method, which can be used in electronic equipment. The dual-microphone array echo cancellation device 500 includes:

[0170] The signal acquisition module 501 is used to acquire a far-end signal generated by the electronic device during a call, and a first near-end signal and a second near-end signal of the dual-microphone array;

[0171] A linear filtering module 502 is used to perform linear filtering processing on the far-end signal, the first near-end signal and the second near-end signal to obtain a first initial error spectrum and a second initial error spectrum;

[0172] A centroid fitting module 503 is used to perform centroid fitting on the first initial error spectrum and the second initial error spectrum respectively to obtain initial suppression gain factors of the first near-end signal and the second near-end signal, and use the initial suppression gain factors as variable step size factors for linear filtering processing of the next frame;

[0173] A residual echo suppression module 504 is used to perform residual echo suppression on the first error spectrum and the second error spectrum using an adaptive pole-zero echo canceller to obtain a target spectrum signal;

[0174] A smoothing module 505 is used to select a sub-band range for the initial suppression gain factor, obtain a smoothing factor, and perform full-band smoothing and exponential operation to obtain a secondary suppression gain factor;

[0175] The target acquisition module 506 is used to filter the target spectrum signal using the secondary suppression gain factor to obtain a target near-end speech signal.

[0176] In an embodiment of the present invention, a far-end signal generated by an electronic device during a call, as well as a first near-end signal and a second near-end signal of a dual-microphone array are obtained, and then, the far-end signal and the near-end signal of the dual-microphone array are subjected to initial filtering processing, and the far-end signal, the first near-end signal and the second near-end signal can be subjected to linear filtering processing to obtain a first initial error spectrum and a second initial error spectrum, thereby preliminarily eliminating echoes; then, the first initial error spectrum and the second initial error spectrum are subjected to centroid fitting respectively to obtain initial suppression gain factors of the first near-end signal and the second near-end signal, and the initial suppression gain factors are used as variable step size factors for linear filtering processing of the next frame; residual echo suppression is performed on the first error spectrum and the second error spectrum using an adaptive zero-pole echo canceller to obtain a target spectrum signal, thereby achieving residual echo suppression; after selecting a sub-band range for the initial suppression gain factor, a smoothing factor is obtained, and full-band smoothing and exponential operations are performed to obtain a secondary suppression gain factor, and the target spectrum signal is subjected to filtering processing using the secondary suppression gain factor to achieve secondary suppression gain filtering, thereby obtaining a target near-end voice signal, thereby effectively improving the echo cancellation effect and the voice call quality.

[0177] In other embodiments, the linear filtering module 502 is further configured to:

[0178] Performing Fourier transform on the far-end signal to obtain far-end frequency domain information, performing Fourier transform on the first near-end signal to obtain first near-end frequency domain information, and performing Fourier transform on the second near-end signal to obtain second near-end frequency domain information;

[0179] Using the filter weight coefficient updated in the previous frame to filter the remote frequency domain information to obtain an echo spectrum;

[0180] Subtracting the echo spectrum from the first near-end frequency domain information to obtain a first initial error spectrum;

[0181] The echo spectrum is subtracted from the second near-end frequency domain information to obtain a second initial error spectrum.

[0182] In other embodiments, the centroid fitting module 503 is further used to:

[0183] calculating a first cross-correlation coefficient between the far-end frequency domain information and the first initial error spectrum, and a second cross-correlation coefficient between the far-end frequency domain information and the second initial error spectrum;

[0184] Performing centroid fitting on the first cross-correlation coefficient to obtain a first initial suppression gain factor, and performing centroid fitting on the second cross-correlation coefficient to obtain a second initial suppression gain factor;

[0185] The first initial error spectrum is multiplied by the first initial suppression gain factor to obtain the first error spectrum, and the second initial error spectrum is multiplied by the second initial suppression gain factor to obtain the second error spectrum.

[0186] In other embodiments, the dual-microphone array echo cancellation device 500 further includes an updating module 507, which is used to:

[0187] updating the filter weight coefficient of the previous frame based on the first initial suppression gain factor and the second initial suppression gain factor to obtain the filter weight coefficient of the current frame;

[0188] If the first initial suppression gain factor and the second initial suppression gain factor are smaller than a preset threshold, the first initial suppression gain factor and the second initial suppression gain factor are used as variable step size factors for linear filtering processing of the next frame.

[0189] In other embodiments, the residual echo suppression module 504 is further configured to:

[0190] Calculate a first branch spectrum signal based on the first error spectrum and delay compensation, and calculate a second branch spectrum signal based on the second error spectrum and delay compensation;

[0191] Based on the first branch spectrum signal and the second branch spectrum signal, calculating a differential signal cross-correlation power spectrum coefficient between the first branch spectrum signal and the second branch spectrum signal and a weight coefficient of an adaptive zero-pole echo canceller;

[0192] The weight coefficients of the adaptive pole-zero echo canceller are used to perform residual echo suppression on the first error spectrum and the second error spectrum to obtain a target spectrum signal.

[0193] In other embodiments, the smoothing module 505 is further configured to:

[0194] Selecting a sub-band range for the initial suppression gain factor to obtain initial suppression gain factors within a plurality of sub-band ranges;

[0195] Sorting a plurality of initial suppression gain factors within the sub-band range, and calculating a suppression gain coefficient smoothing factor;

[0196] The initial suppression gain factor is subjected to full-band smoothing processing and exponential operation using the suppression gain coefficient smoothing factor to obtain the secondary suppression gain factor.

[0197] In other embodiments, the target acquisition module 506 is further configured to:

[0198] Performing noise reduction processing on the target spectrum signal;

[0199] After the target spectrum signal after noise reduction processing is multiplied by the secondary suppression gain factor, an inverse Fourier transform is performed, and overlap-addition is performed to obtain a target near-end speech signal.

[0200] It should be noted that the above device can execute the method provided in the embodiment of the present application, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not fully described in the device embodiment, please refer to the method provided in the embodiment of the present application.

[0201] Figure 6 FIG. 1 is a schematic diagram of the hardware structure of the controller of the electronic device 100 in one embodiment of the electronic device 100. Figure 6 As shown, the controller 110 includes:

[0202] One or more processors 111 and memory 112 . Figure 6 In the figure, a processor 111 and a memory 112 are taken as an example.

[0203] The processor 111 and the memory 112 may be connected via a bus or other means. Figure 6 The example of connecting through bus is taken in the following.

[0204] The memory 112 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions / modules corresponding to the dual-microphone array echo cancellation method in the embodiment of the present application (for example, the attached Figure 5 The processor 111 executes various functional applications and data processing of the controller by running the non-volatile software programs, instructions and modules stored in the memory 112, that is, the dual-microphone array echo cancellation method of the above method embodiment is implemented.

[0205] The memory 112 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the electronic device 100, etc. In addition, the memory 112 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 112 may optionally include a memory remotely arranged relative to the processor 111, and these remote memories may be connected to a signal long-term recording device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0206] The one or more modules are stored in the memory 112, and when executed by the one or more processors 111, the dual-microphone array echo cancellation method in any of the above method embodiments is executed, for example, the above described Figure 4 Method steps S401 to S406 in the method; implement Figure 5 The functions of modules 501-507 in.

[0207] The above-mentioned product can execute the method provided in the embodiment of the present application, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not fully described in this embodiment, please refer to the method provided in the embodiment of the present application.

[0208] The embodiment of the present application provides a non-volatile computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are executed by one or more processors, such as Figure 6 A processor 111 in the embodiment may enable the one or more processors to execute the dual-microphone array echo cancellation method in any of the above method embodiments, for example, to execute the above described Figure 4 Method steps S401 to S406 in the method; implement Figure 5 The functions of modules 501-407 in.

[0209] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0210] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a general hardware platform, and of course, can also be implemented by hardware. Those skilled in the art can understand that all or part of the processes in the above embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above methods. Among them, the storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0211] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them. Under the concept of the present invention, the technical features in the above embodiments or different embodiments may also be combined, the steps may be implemented in any order, and there are many other changes in different aspects of the present invention as described above, which are not provided in detail for the sake of simplicity. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they may still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A dual-microphone array echo cancellation method, It is characterized in that The method comprises: Acquire a far-end signal, a first near-end signal, and a second near-end signal, wherein the far-end signal is a sound signal played by a speaker of an electronic device, and the first near-end signal and the second near-end signal are sound signals collected by a dual-microphone array of the electronic device; Performing Fourier transform on the far-end signal to obtain far-end frequency domain information, performing Fourier transform on the first near-end signal to obtain first near-end frequency domain information, and performing Fourier transform on the second near-end signal to obtain second near-end frequency domain information; Using the filter weight coefficient updated in the previous frame to filter the remote frequency domain information to obtain an echo spectrum; Subtracting the echo spectrum from the first near-end frequency domain information to obtain a first initial error spectrum; Subtracting the echo spectrum from the second near-end frequency domain information to obtain a second initial error spectrum; calculating a first cross-correlation coefficient between the far-end frequency domain information and the first initial error spectrum, and a second cross-correlation coefficient between the far-end frequency domain information and the second initial error spectrum; Performing centroid fitting on the first cross-correlation coefficient to obtain a first initial suppression gain factor of the first near-end signal, and performing centroid fitting on the second cross-correlation coefficient to obtain a second initial suppression gain factor of the second near-end signal; Multiply the first initial error spectrum by the first initial suppression gain factor to obtain a first error spectrum, and multiply the second initial error spectrum by the second initial suppression gain factor to obtain a second error spectrum; The first error spectrum and the second error spectrum are subjected to residual echo suppression using an adaptive zero-pole echo canceller based on the following formula to obtain a target spectrum signal: in, represents the first error spectrum, represents the second error spectrum, represents the second branch spectrum signal, represents the spectrum signal of the first branch, delay represents the delay compensation operation, represents the delay compensation of the first branch, represents the delay compensation of the second branch. The value range of α is 0≤α≤1, the value range of β is 0≤β≤1, and α+β=1. represents the weight coefficient of the adaptive zero-pole echo canceller, represents the cross-correlation power spectrum coefficient of the differential signal, represents the second branch autocorrelation power spectrum, Represents the target spectrum signal.

2. The method according to claim 1, It is characterized in that The first cross-correlation coefficient is subjected to centroid fitting by the following formula to obtain a first initial suppression gain factor, and the second cross-correlation coefficient is subjected to centroid fitting to obtain a second initial suppression gain factor: Where k represents the centroid fitting distance, i represents the i-th frequency point of the n-th frame, represents the first initial suppression gain factor; represents the second initial suppression gain factor, represents the first mutual correlation coefficient, represents the second correlation coefficient.

3. The method according to claim 1, It is characterized in that Also includes: updating the filter weight coefficient of the previous frame based on the first initial suppression gain factor and the second initial suppression gain factor to obtain the filter weight coefficient of the current frame; If the first initial suppression gain factor and the second initial suppression gain factor are smaller than a preset threshold, the first initial suppression gain factor and the second initial suppression gain factor are used as variable step-size factors for linear filtering processing of the next frame.

4. The method according to claim 1, It is characterized in that Also includes: The initial suppression gain factor is selected in a sub-band range to obtain initial suppression gain factors in several sub-band ranges, wherein the minimum value of the sub-band range is minFre and the maximum value of the sub-band range is maxFre; The initial suppression gain factors within the sub-band range are sorted to obtain a sequence ; Select reference subband point , get the reference value , ; The suppression gain coefficient smoothing factor is calculated by the following formula: ； The first initial suppression gain factor and the second initial suppression gain factor are subjected to full-band smoothing processing and exponential operation using the suppression gain coefficient smoothing factor through the following formula to obtain a secondary suppression gain factor: ； The target spectrum signal is filtered using the secondary suppression gain factor to obtain a target near-end speech signal. represents the first initial suppression gain factor, Represents the second initial suppression gain factor.

5. The method according to claim 4, It is characterized in that The filtering process of the target spectrum signal by using the secondary suppression gain factor to obtain the target near-end speech signal includes: Performing noise reduction processing on the target spectrum signal; After the target spectrum signal after noise reduction processing is multiplied by the secondary suppression gain factor, an inverse Fourier transform is performed, and overlap-addition is performed to obtain a target near-end speech signal.

6. A dual-microphone array echo cancellation device, It is characterized in that The device comprises: a signal acquisition module, configured to acquire a far-end signal, a first near-end signal, and a second near-end signal, wherein the far-end signal is a sound signal played by a speaker of the electronic device, and the first near-end signal and the second near-end signal are sound signals collected by a dual-microphone array of the electronic device; a linear filtering module, configured to perform Fourier transform on the far-end signal to obtain far-end frequency domain information, perform Fourier transform on the first near-end signal to obtain first near-end frequency domain information, and perform Fourier transform on the second near-end signal to obtain second near-end frequency domain information, filter the far-end frequency domain information using the filter weight coefficient updated in the previous frame to obtain an echo spectrum, subtract the echo spectrum from the first near-end frequency domain information to obtain a first initial error spectrum, and subtract the echo spectrum from the second near-end frequency domain information to obtain a second initial error spectrum; a centroid fitting module, configured to calculate a first cross-correlation coefficient between the far-end frequency domain information and the first initial error spectrum, and a second cross-correlation coefficient between the far-end frequency domain information and the second initial error spectrum, perform centroid fitting on the first cross-correlation coefficient to obtain a first initial suppression gain factor of the first near-end signal, perform centroid fitting on the second cross-correlation coefficient to obtain a second initial suppression gain factor of the second near-end signal, and multiply the first initial error spectrum by the first initial suppression gain factor to obtain a first error spectrum, and multiply the second initial error spectrum by the second initial suppression gain factor to obtain a second error spectrum; The residual echo suppression module is used to perform residual echo suppression on the first error spectrum and the second error spectrum by using an adaptive zero-pole echo canceller based on the following formula to obtain a target spectrum signal: in, represents the first error spectrum, represents the second error spectrum, represents the second branch spectrum signal, represents the spectrum signal of the first branch, delay represents the delay compensation operation, represents the delay compensation of the first branch, represents the delay compensation of the second branch. The value range of α is 0≤α≤1, the value range of β is 0≤β≤1, and α+β=1. represents the weight coefficient of the adaptive zero-pole echo canceller, represents the cross-correlation power spectrum coefficient of the differential signal, represents the second branch autocorrelation power spectrum, Represents the target spectrum signal.

7. An electronic device, It is characterized in that The electronic device comprises: A dual microphone array and at least one processor and a memory, wherein the memory is communicatively connected to the processor, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described in any one of claims 1-5.

8. The electronic device according to claim 7, It is characterized in that The dual microphone array is an end-fire configuration.

9. A non-volatile computer-readable storage medium, It is characterized in that The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by an electronic device, the electronic device executes the method according to any one of claims 1 to 5.

Images