Echo cancellation system and method

Abstract

The invention provides an echo cancellation system and method. The method comprises the following steps: processing an obtained near-end signal and a far-end signal through a frequency domain sequence module to obtain a near-end frequency domain sequence and a far-end frequency domain sequence; a reference signal is preset through a double-talk detection module, a coherence coefficient is obtained by combining a near-end frequency domain sequence, a far-end frequency domain sequence and a double-talk detection algorithm, and different threshold values are selected for a first filter and a second filter for double-talk detection to obtain a first double-talk detection result and a second double-talk detection result; corresponding filtering parameters are updated in the filtering process of the two filters through a filtering module, and a two-stage linear filtering result is output; and adjusting corresponding filters in different states based on two-stage linear filtering results to eliminate echoes. Compared with the prior art in which different filters are converged in different states by obtaining the double-talk detection result of the filter, updating the filter coefficient and outputting the filter result, the method can effectively solve the problem of steady-state imbalance and improve the echo elimination effect.

Description

Technical Field

[0001] The present invention relates to the field of information communication technology, and in particular to an echo cancellation system and method. Background Art

[0002] Usually, when a mobile phone or other mobile terminal is in speaker mode, an echo will be generated. Echo cancellers generally use adaptive linear filters to eliminate echoes. Common adaptive filter coefficient update algorithms include least mean square (NLMS), recursive least squares (RLS), Kalman filter, etc. In echo cancellation, there is often a double talk problem. When double talk occurs, the filter coefficients will diverge if they continue to be updated. Another problem is the nonlinear distortion problem. When the nonlinear distortion of the echo is large, the filter is also prone to divergence. The conventional method is to use a double talk detection (DTD) module to detect single talk and double talk. When double talk is detected, the filter coefficients stop updating.

[0003] The Speex solution uses dual adaptive filters; one of the filters is the main filter, which uses a variable step adaptive algorithm to update the coefficients, and the other is the backup filter. The topology is shown in Figure 4 This solution can avoid the filter divergence during double talk to a certain extent, but the steady-state misalignment increases significantly when the nonlinear distortion of the echo is relatively large, and the convergence speed is slow when the echo path changes.

[0004] Therefore, it is an urgent problem to be solved by those skilled in the art to provide an echo cancellation system and method that effectively solves the above problems. Summary of the invention

[0005] The object of the present invention is to provide an echo cancellation system which has a simple structure, is safe, effective, reliable and easy to operate, can reduce the steady-state misalignment of the filter coefficients in the two situations of double talk state and strong nonlinear distortion of the echo in echo cancellation, and ensures a faster convergence speed when the echo path changes.

[0006] Based on the above objectives, the technical solution provided by the present invention is as follows: An echo cancellation system, comprising: A frequency domain sequence module, used for processing the near-end signal and the far-end signal respectively to obtain a near-end frequency domain sequence and a far-end frequency domain sequence; A double-talk detection module, configured to obtain a coherence coefficient according to a preset reference signal, the near-end frequency domain sequence, the far-end frequency domain sequence, and a double-talk detection algorithm; The double-talk detection module is further used to select different thresholds for the first filter and the second filter in the filter module to perform double-talk detection according to the coherence coefficient, so as to obtain a first double-talk detection result and a second double-talk detection result; The filtering module includes: the first filter and the second filter, wherein the first filter is used to update a first filtering coefficient according to the first double-talk detection result during the filtering process, and output the first filtering result to the second filter; the second filter is used to update a second filtering coefficient according to the first filtering result and the second double-talk detection result during the filtering process, so as to obtain a two-stage linear filtering result; The filtering module is also used to converge the second filter in the first state to eliminate the echo based on the two-stage linear filtering results; in the second state, adjust the learning rate of the first filter according to the second filter coefficient to converge the first filter to eliminate the echo.

[0007] An echo cancellation method, implemented based on the echo cancellation system as described above, comprises the following steps: Processing the near-end signal and the far-end signal respectively to obtain a near-end frequency domain sequence and a far-end frequency domain sequence; Acquire a coherence coefficient according to a preset reference signal, the near-end frequency domain sequence, the far-end frequency domain sequence, and a double-talk detection algorithm; Selecting different thresholds for the first filter and the second filter to perform double talk detection according to the coherence coefficient to obtain a first double talk detection result and a second double talk detection result; During the filtering process, a first filter coefficient is updated according to the first double-talk detection result, and the first filter result is output to the second filter; the second filter is used to update the second filter coefficient according to the first filter result and the second double-talk detection result during the filtering process to obtain a two-stage linear filtering result; Based on the two-stage linear filtering results, in the first state, the second filter is converged to eliminate the echo; in the second state, the learning rate of the first filter is adjusted according to the second filter coefficient to converge the first filter to eliminate the echo.

[0008] Preferably, the processing of the near-end signal and the far-end signal separately to obtain the near-end frequency domain sequence and the far-end frequency domain sequence is specifically as follows: Acquire the near-end signal and the far-end signal; The near-end signal and the far-end signal are respectively subjected to frame division, mixed windowing and M-point discrete Fourier transform in sequence to obtain the near-end frequency domain sequence and the far-end frequency domain sequence.

[0009] Preferably, the obtaining of the coherence coefficient according to the preset reference signal, the near-end frequency domain sequence, the far-end frequency domain sequence and the double-talk detection algorithm is specifically: ; in, and are the near-end frequency domain sequence and the far-end frequency domain sequence respectively, is the conjugate operator symbol, is the frequency domain cross-correlation value between the near-end signal and the far-end signal, is the cross power spectrum of the near-end signal and the far-end signal, is the near-end signal autopower spectrum, is the far-end signal autopower spectrum, is the coherence coefficient for each frequency point.

[0010] Preferably, the selecting different thresholds for the first filter and the second filter to perform double talk detection according to the coherence coefficient to obtain a first double talk detection result and a second double talk detection result is specifically: , , in, and They are the first double-talk test result and the second double-talk test result, and are different thresholds, and , define 1 as single talk and 0 as double talk.

[0011] Preferably, the filtering process of the first filter is specifically as follows: , , ; in, is the reference signal frequency domain sequence, is the first filter coefficient, is the order of the first filter, is the first filtering result.

[0012] Preferably, in the filtering process, the first filtering coefficient is updated according to the first double-talk detection result, specifically: ， ， + ， in, is the reference signal autocorrelation power spectrum, is the sliding average coefficient, is the coefficient update step size, The learning rate parameter calculated for the previous frame.

[0013] Preferably, the second filter filtering process and the second filter coefficient updating process are specifically: , ， ， ， ， ， ， ， ， ， ， = ； in, is the filtered estimated echo, is the filter order, is the preset maximum value of the smoothing coefficient, is the sliding average coefficient, is the second filter output, is the cross power spectrum between the first filter output and the far-end signal, The autopower spectrum of the far-end signal, is the linear regression coefficient between the first filter output and the far-end signal, ie, the second filter coefficient.

[0014] Preferably, adjusting the learning rate of the first filter according to the second filter coefficient is specifically: Define the coefficient vector of the second filter , is the conjugate transpose parameter, and the first filter learning rate parameter is obtained ,as follows: , in, is a constant less than 1, are the coefficients of the first filter.

[0015] The echo cancellation system provided by the present invention processes the acquired near-end signal and far-end signal through a frequency domain sequence module to obtain a near-end frequency domain sequence and a far-end frequency domain sequence; presets a reference signal through a double-talk detection module, combines the near-end frequency domain sequence and the far-end frequency domain sequence and a double-talk detection algorithm, obtains a coherence coefficient, and selects different thresholds for the first filter and the second filter to perform double-talk detection, and obtains the first and second double-talk detection results; updates the corresponding filter parameters in the filtering process of the two filters through a filtering module, and outputs a two-stage linear filtering result; based on the two-stage linear filtering result, in the first state, converges the second filter, and in the second state, adjusts the learning rate of the first filter, converges the first filter, and eliminates the echo. Compared with the prior art, the present invention obtains the double-talk detection result of the filter, updates the filter coefficient, outputs the filtering result, and converges different filters in different states, which can effectively solve the steady-state imbalance and improve the echo elimination effect.

[0016] The present invention also provides an echo cancellation method, which solves the same technical problem as the system, belongs to the same inventive concept, and should have the same beneficial effects, so it will not be described in detail here. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A schematic diagram of the structure of an echo cancellation system provided by an embodiment of the present invention; Figure 2 A flowchart of an echo cancellation method provided by an embodiment of the present invention; Figure 3 A flowchart of step S1 provided in an embodiment of the present invention; Figure 4 A topological structure diagram of the existing Speex provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] The embodiments of the present invention are written in a progressive manner.

[0021] The embodiment of the present invention provides an echo cancellation system and method, which mainly solves the technical problem of steady-state misalignment of filter coefficients in the two situations of double talk state and strong nonlinear distortion of echo in echo cancellation in the prior art.

[0022] An echo cancellation system, comprising: A frequency domain sequence module, used for processing the near-end signal and the far-end signal respectively to obtain a near-end frequency domain sequence and a far-end frequency domain sequence; A double-talk detection module, used to obtain a coherence coefficient according to a preset reference signal, a near-end frequency domain sequence and a far-end frequency domain sequence and a double-talk detection algorithm; The double talk detection module is further used to select different thresholds for the first filter and the second filter in the filter module according to the coherence coefficient to perform double talk detection, so as to obtain a first double talk detection result and a second double talk detection result; The filtering module includes: a first filter and a second filter, wherein the first filter is used to update a first filtering coefficient according to a first double-talk detection result during a filtering process, and output the first filtering result to the second filter; and the second filter is used to update a second filtering coefficient according to the first filtering result and the second double-talk detection result during a filtering process, so as to obtain a two-stage linear filtering result. The filtering module is also used to converge the second filter in the first state to eliminate the echo based on the two-stage linear filtering results; in the second state, adjust the learning rate of the first filter according to the second filter coefficient to converge the first filter to eliminate the echo.

[0023] The echo cancellation system provided by the present invention processes the acquired near-end signal and far-end signal through a frequency domain sequence module to obtain a near-end frequency domain sequence and a far-end frequency domain sequence; presets a reference signal through a double-talk detection module, combines the near-end frequency domain sequence and the far-end frequency domain sequence and a double-talk detection algorithm to obtain a coherence coefficient, and selects different thresholds for a first filter and a second filter to perform double-talk detection to obtain first and second double-talk detection results; updates corresponding filtering parameters in the filtering process of the two filters through a filtering module, and outputs a two-stage linear filtering result; based on the two-stage linear filtering result, in a first state, converges the second filter, and in a second state, adjusts the learning rate of the first filter, converges the first filter, and eliminates the echo.

[0024] In this embodiment, the first filter is specifically selected as an NLMS filter, and the second filter is specifically selected as an SLR filter; Among them, the LMS algorithm stands for the least mean square algorithm, which is a linear adaptive filtering algorithm. It does not require the calculation of correlation functions or matrix inversion calculations. Due to the simplicity of the basis (small amount of calculation), the LMS algorithm has become a reference standard for other linear adaptive filtering algorithms. LMS generally includes two basic processes: filtering process: a. Calculate the response of the linear filter output to the input signal; b. Generate an estimated error by comparing the output result with the expected response. Adaptive process: Automatically adjust the filter parameters according to the estimated error; NLMS is the Normalized LMS algorithm, which is an improved version of the LMS algorithm. In the LMS algorithm, the adjustment amount of the tap weight is proportional to the tap input vector. When the tap input vector is relatively large, LMS will encounter the problem of gradient noise amplification. To overcome this problem, the NLMS algorithm can be used. It is called "Normalized" because it uses the squared Euclidean norm of the tap input vector to normalize the tap weight adjustment amount.

[0025] The SLR filter is a filter based on the linear regression algorithm structure. The Stage Linear Regression (SLR) algorithm is a method that can achieve fast filter convergence. Considering that the energy of the echo is concentrated in a short continuous period of time, a smaller order can be used in the frequency domain filter, and the filter coefficients can be iteratively solved by hierarchical linear regression using continuous reference signals.

[0026] An echo cancellation method, based on the above-mentioned echo cancellation system, comprises the following steps: S1. Processing the near-end signal and the far-end signal respectively to obtain a near-end frequency domain sequence and a far-end frequency domain sequence; S2. Obtaining the coherence coefficient according to the preset reference signal, the proximal frequency domain sequence and the distal frequency domain sequence and the double-talk detection algorithm; S3. Select different thresholds for double-talk detection for the first filter and the second filter according to the coherence coefficient to obtain the first double-talk detection result and the second double-talk detection result; S4. During the filtering process, the first filter coefficient is updated according to the first double-talk detection result, and the first filter result is output to the second filter; the second filter is used to update the second filter coefficient according to the first filter result and the second double-talk detection result during the filtering process to obtain a two-stage linear filtering result; S5. Based on the two-stage linear filtering results, in the first state, converge the second filter to eliminate the echo; in the second state, adjust the learning rate of the first filter according to the second filter coefficient to converge the first filter to eliminate the echo.

[0027] In step S1 to step S3, by processing the near-end signal and the far-end signal respectively, a near-end frequency domain sequence and a far-end frequency domain sequence are obtained, and the coherence coefficient of each frequency point is calculated and obtained by combining the double talk detection (DTD) algorithm; according to the coherence coefficient of each frequency point, different thresholds are selected for the NLMS filter and the SLR filter to perform double talk detection, and the corresponding double talk detection result is obtained; In step S4 to step S5, the coefficient of the NLMS filter is updated according to the first double-talk detection result, and the filtering result of the NLMS filter is output to the SLR filter. During the filtering process, the SLR filter updates the coefficient of the SLR filter according to the filtering result of the NLMS filter and the second double-talk detection result to obtain a two-stage linear filtering result; based on the two-stage linear filtering result, when the NLMS convergence efficiency is poor, a large amount of residual echo can be eliminated by fast convergence of the SLR filter; when the SLR filter converges well but the NLMS does not converge, the residual echo can be eliminated by adjusting the learning rate of the first filter so that the NLMS filter converges quickly.

[0028] Preferably, S1 is specifically: A1. Obtain near-end signal and far-end signal; A2. The near-end signal and the far-end signal are respectively framed, mixed and windowed, and subjected to M-point discrete Fourier transform to obtain a near-end frequency domain sequence and a far-end frequency domain sequence.

[0029] In steps A1 to A2, the near-end signal and the far-end signal are collected through a conventional communication system, and the near-end signal and the far-end signal are framed, mixed, windowed, and subjected to M-point discrete Fourier transform to obtain corresponding near-end frequency domain sequences and far-end frequency domain sequences; In this embodiment, before performing frequency domain analysis, the signal is first preprocessed as in the above steps; wherein, in the framing process, there needs to be overlap between two adjacent frames. Because the speech signal is time-varying, the feature changes in a short time range are small, so it can be regarded as a steady state for analysis. However, if it exceeds this short time range, the signal will change, and the corresponding fundamental tones may change between the two adjacent frames. If it is exactly in the middle of two syllables, or it is the transition from the initial consonant to the final, etc., the feature parameters of the two frames may change greatly. However, in order to make the feature parameters change more smoothly, some frames are inserted between two non-overlapping frames to extract the feature parameters, thus forming an overlapping part between adjacent frames.

[0030] Preferably, the coherence coefficient is obtained according to the preset reference signal, the near-end frequency domain sequence, the far-end frequency domain sequence and the double-talk detection algorithm, specifically: ; in, and are the near-end frequency domain sequence and the far-end frequency domain sequence respectively, is the conjugate operator symbol, is the frequency domain cross-correlation value between the near-end signal and the far-end signal, is the cross power spectrum of the near-end signal and the far-end signal, is the near-end signal autopower spectrum, is the far-end signal autopower spectrum, is the coherence coefficient for each frequency point.

[0031] In actual application, the near-end frequency domain sequence and the far-end frequency domain sequence are input into the above-mentioned double-talk detection algorithm formula, and combined with the preset reference signal, the coherence coefficient of each frequency point is obtained.

[0032] Preferably, different thresholds are selected for the first filter and the second filter according to the coherence coefficient to perform double talk detection, so as to obtain a first double talk detection result and a second double talk detection result, specifically: , , in, and They are the first double-talk test result and the second double-talk test result, and are different thresholds, and , define 1 as single talk and 0 as double talk.

[0033] In actual application, after determining the coherence coefficient, different thresholds are selected for the NLMS filter and the SLR filter respectively. The threshold of the NLMS filter is greater than the threshold of the SLR filter. When the double talk detection result is 1, it is defined as single talk, and when the result is 0, it is defined as double talk.

[0034] Preferably, the filtering process of the first filter is specifically as follows: , , ; in, is the reference signal frequency domain sequence, is the first filter coefficient, is the order of the first filter, is the first filtering result.

[0035] Preferably, during the filtering process, the first filtering coefficient is updated according to the first double-talk detection result, specifically: ， ， + ， in, is the reference signal autocorrelation power spectrum, is the sliding average coefficient, is the coefficient update step size, The learning rate parameter calculated for the previous frame.

[0036] In actual application, during the filtering process of the NLMS filter, the filter coefficient of the NLMS filter is updated according to the first double-talk detection result. After the update is completed, the first filtering result is output. to the SLR filter; In this embodiment, the sliding average coefficient ranges from 0.8 to 0.99.

[0037] Preferably, the second filter filtering process and the process of updating the second filter coefficient are specifically: , ， ， ， ， ， ， ， ， ， ， = ； in, is the filtered estimated echo, is the filter order, is the preset maximum value of the smoothing coefficient, is the sliding average coefficient, is the second filter output, is the cross power spectrum between the first filter output and the far-end signal, The autopower spectrum of the far-end signal, is the linear regression coefficient between the first filter output and the far-end signal, ie, the coefficient of the second filter.

[0038] In actual application, when the SLR filter receives the first filtering result After that, the SLR filter is updated and filtered by double-talk detection. This process is carried out in stages, and the filter order is , filter and update the coefficients as shown in the above formula, and finally get the second filtering result , output two-stage linear filtering results .

[0039] Preferably, in step S5, the learning rate of the first filter is adjusted according to the second filter coefficient, specifically: Define the coefficient vector of the second filter , is the conjugate transpose parameter, and the first filter learning rate parameter is obtained

[0040] , in, is a constant less than 1, are the coefficients of the first filter.

[0041] In actual application, the learning rate when the SLR filter converges well but the NLMS does not converge It has a relatively large value, so it can accelerate the convergence of NLMS filtering.

[0042] In the embodiments provided in the present application, it should be understood that the disclosed methods and systems can be implemented in other ways. The system embodiments described above are merely schematic. For example, the division of modules is only a logical function division. There may be other division methods in actual implementation, such as: multiple modules or components can be combined, or can be integrated into another system, or some features can be ignored, or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0043] In addition, all functional modules in the embodiments of the present invention may be integrated into one processor, or each module may be a separate device, or two or more modules may be integrated into one device; each functional module in the embodiments of the present invention may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0044] A person of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiment can be completed by program instructions and related hardware. The aforementioned program instructions can be stored in a computer-readable storage medium. When the program instructions are executed, the steps of the above method embodiment are executed; and the aforementioned storage medium includes: a mobile storage device, a read-only memory (ROM), a magnetic disk or an optical disk, and other media that can store program codes.

[0045] It should be understood that the use of "system", "device", "unit" and / or "module" in this application is only a method for distinguishing different components, elements, parts, parts or assemblies at different levels. However, if other words can achieve the same purpose, the word can be replaced by other expressions.

[0046] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "a kind" and / or "the" do not refer to the singular, but also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of clearly identified steps and elements, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements. The elements defined by the sentence "includes a..." do not exclude the existence of other identical elements in the process, method, commodity or device that includes the elements.

[0047] In the description of the embodiments of the present application, "plurality" refers to two or more than two.

[0048] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features.

[0049] If a flow chart is used in the present application, the flow chart is used to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, each step can be processed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or a certain step or several steps of operations can be removed from these processes.

[0050] The above is a detailed introduction to an echo cancellation system and method provided by the present invention. The above description of the disclosed embodiments enables professionals and technicians in the field to implement or use the present invention. Various modifications to these embodiments will be obvious to professionals and technicians in the field, and the general principles defined in this article can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown in this article, but will conform to the widest scope consistent with the principles and novel features disclosed in this article.

Claims

1. An echo cancellation system, characterized in that: include: A frequency domain sequence module, used for processing the near-end signal and the far-end signal respectively to obtain a near-end frequency domain sequence and a far-end frequency domain sequence; A double-talk detection module, configured to obtain a coherence coefficient according to a preset reference signal, the near-end frequency domain sequence, the far-end frequency domain sequence, and a double-talk detection algorithm; The double-talk detection module is further used to select different thresholds for the first filter and the second filter in the filter module to perform double-talk detection according to the coherence coefficient, so as to obtain a first double-talk detection result and a second double-talk detection result; The filtering module includes: the first filter and the second filter, wherein the first filter is used to update a first filtering coefficient according to the first double-talk detection result during the filtering process, and output the first filtering result to the second filter; the second filter is used to update a second filtering coefficient according to the first filtering result and the second double-talk detection result during the filtering process, so as to obtain a two-stage linear filtering result; The filtering module is also used to converge the second filter in the first state to eliminate the echo based on the two-stage linear filtering results; in the second state, adjust the learning rate of the first filter according to the second filter coefficient to converge the first filter to eliminate the echo.

2. An echo cancellation method, implemented based on the echo cancellation system according to claim 1, characterized in that: The steps include: Processing the near-end signal and the far-end signal respectively to obtain a near-end frequency domain sequence and a far-end frequency domain sequence; Acquire a coherence coefficient according to a preset reference signal, the near-end frequency domain sequence, the far-end frequency domain sequence, and a double-talk detection algorithm; Selecting different thresholds for the first filter and the second filter to perform double talk detection according to the coherence coefficient to obtain a first double talk detection result and a second double talk detection result; During the filtering process, a first filter coefficient is updated according to the first double talk detection result, and the first filter result is output to the second filter; The second filter is used to update a second filter coefficient according to the first filtering result and the second double-talk detection result during the filtering process to obtain a two-stage linear filtering result; Based on the two-stage linear filtering results, in the first state, the second filter is converged to eliminate the echo; in the second state, the learning rate of the first filter is adjusted according to the second filter coefficient to converge the first filter to eliminate the echo.

3. The echo cancellation method according to claim 2, characterized in that: The processing of the near-end signal and the far-end signal respectively to obtain the near-end frequency domain sequence and the far-end frequency domain sequence is specifically as follows: Acquire the near-end signal and the far-end signal; The near-end signal and the far-end signal are respectively subjected to frame division, mixed windowing and M-point discrete Fourier transform in sequence to obtain the near-end frequency domain sequence and the far-end frequency domain sequence.

4. The echo cancellation method according to claim 3, characterized in that: The obtaining of the coherence coefficient according to the preset reference signal, the near-end frequency domain sequence, the far-end frequency domain sequence and the double-talk detection algorithm is specifically: ； in, and are the near-end frequency domain sequence and the far-end frequency domain sequence respectively, is the conjugate operator symbol, is the frequency domain cross-correlation value between the near-end signal and the far-end signal, is the cross power spectrum of the near-end signal and the far-end signal, is the near-end signal autopower spectrum, is the far-end signal autopower spectrum, is the coherence coefficient for each frequency point.

5. The echo cancellation method according to claim 4, characterized in that: The step of selecting different thresholds for the first filter and the second filter according to the coherence coefficient to perform double talk detection to obtain a first double talk detection result and a second double talk detection result is specifically: , , in, and They are the first double-talk test result and the second double-talk test result, and are different thresholds, and , define 1 as single talk and 0 as double talk.

6. The echo cancellation method according to claim 5, characterized in that: The filtering process of the first filter is specifically as follows: ， ， ； in, is the reference signal frequency domain sequence, is the first filter coefficient, is the order of the first filter, is the first filtering result.

7. The echo cancellation method according to claim 6, characterized in that: In the filtering process, the first filtering coefficient is updated according to the first double-talk detection result, specifically: ， ， + ， in, is the reference signal autocorrelation power spectrum, is the sliding average coefficient, is the coefficient update step size, The learning rate parameter calculated for the previous frame.

8. The echo cancellation method according to claim 7, characterized in that: The filtering process of the second filter and the process of updating the second filter coefficient are specifically as follows: , ， ， ， ， ， ， ， ， ， ， = ； in, is the filtered estimated echo, is the filter order, is the preset maximum value of the smoothing coefficient, is the sliding average coefficient, is the second filter output, is the cross power spectrum between the first filter output and the far-end signal, The autopower spectrum of the far-end signal, is the linear regression coefficient between the first filter output and the far-end signal, ie, the coefficient of the second filter.

9. The echo cancellation method according to claim 8, characterized in that: The learning rate of the first filter is adjusted according to the second filter coefficient, specifically: Define the coefficient vector of the second filter , is the conjugate transpose parameter, and the first filter learning rate parameter is obtained ,as follows: ， in, is a constant less than 1, are the coefficients of the first filter.

Images