CN110572525B - Self-adaptive communication echo cancellation method for voice communication - Google Patents
Self-adaptive communication echo cancellation method for voice communication Download PDFInfo
- Publication number
- CN110572525B CN110572525B CN201911043026.3A CN201911043026A CN110572525B CN 110572525 B CN110572525 B CN 110572525B CN 201911043026 A CN201911043026 A CN 201911043026A CN 110572525 B CN110572525 B CN 110572525B
- Authority
- CN
- China
- Prior art keywords
- current time
- signal
- vector
- tap weight
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M9/00—Arrangements for interconnection not involving centralised switching
- H04M9/08—Two-way loud-speaking telephone systems with means for conditioning the signal, e.g. for suppressing echoes for one or both directions of traffic
- H04M9/082—Two-way loud-speaking telephone systems with means for conditioning the signal, e.g. for suppressing echoes for one or both directions of traffic using echo cancellers
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Filters That Use Time-Delay Elements (AREA)
- Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)
Abstract
A self-adaptive communication echo cancellation method for voice communication mainly comprises the following steps: A. echo cancellation; B. tap weight vector update: b1 residual signal square sequence E2(n) calculating; b2, calculating the weighted median residual error to obtain the weighted median residual error sigma (n) of the current time n; b3, calculating an M estimation value to obtain an M estimation function value psi (n) of a residual error at the current time n; b4, calculating a zero attraction vector to obtain a zero attraction factor of a tap weight coefficient, thereby forming a zero attraction vector F (n) of the current time n, B5, updating the filter tap weight vector, introducing an M estimation function value and the zero attraction vector into an updating formula, and obtaining a filter tap weight vector W (n +1) of the next time n + 1; C. and (6) repeating. The method has the advantages of high convergence rate, low steady-state error and good impact resistance effect.
Description
Technical Field
The invention relates to a self-adaptive echo cancellation method in voice communication.
Background
When a call (voice communication) is carried out, a sound signal is reflected back to a signal source through time delay or deformation to form an echo, and the quality of the voice call is seriously influenced by the echo phenomenon. For example, when a call is made, because the speaker and the microphone are located in the same space, the local near-end microphone receives the far-end speech from the local speaker and transmits the far-end speech back, which causes the far-end speaker to hear his own voice, resulting in a degraded quality of the call. This phenomenon is widely present in voice communication systems such as satellite communication, hands-free telephones, teleconferencing systems, and the like. There is a need to suppress echo signals, remove their effects and improve voice call quality by taking effective measures. The self-adaptive echo cancellation technology has the advantages of low cost, high convergence speed and small echo residual error, and is widely applied to voice communication. The adaptive echo cancellation technique for voice communication achieves the purpose of echo cancellation by estimating the echo signal and subtracting the estimated value of the echo from the near-end signal.
The common practice of the adaptive echo cancellation method is to sample a near-end microphone to obtain a near-end signal with echo at the current moment, subtract an estimated value of the echo signal from the near-end microphone to obtain an error signal at the current moment, and then send the error signal at the current moment back to the far end; the square of the difference (error) between the estimated value of the filter and the near-end signal is the minimum, and the square is used as a cost function to carry out iterative computation, so that the adaptive elimination of the echo is realized. Since the adaptive echo cancellation system is usually a sparse system, the length of the response system can reach hundreds of symbols, but only a few effective factors are nonzero coefficients, so that the convergence rate is slow and the echo cancellation performance is low. Moreover, when the noise of the far-end input signal cannot be ignored, the traditional least mean square algorithm will generate biased estimation; the convergence rate of the algorithm is low, the steady-state error is large, and the echo cancellation effect is poor. In addition, when there is impulse noise, the "error signal" is large, and the tap weight vector of the filter will generate a large update, resulting in an increase of steady-state error and a slow convergence rate.
Disclosure of Invention
The invention aims to provide an adaptive communication echo cancellation method for voice communication. The method can still realize the unbiased estimation of the signal when the far-end signal contains noise, and has higher convergence speed and low steady-state error when the impulse interference exists, and the echo cancellation effect is good.
The technical scheme adopted by the invention for realizing the aim is that the self-adaptive communication echo cancellation method for the voice communication comprises the following steps:
A. echo cancellation
A1, remote signal acquisition
Sampling a signal transmitted from a far end to obtain a discrete value x (n) of a far end input signal at the current moment n; input signals x (n), x (n-1),. and x (n-L +1) from a current time n to a time n-L +1 are combined to form an adaptive filter input vector x (n) at the current time n, x (n) ([ x (n), x (n-1),. and x (n-L +1)]T(ii) a Wherein, L is 512, which represents the number of filter taps, and T represents the transposition operation;
a2 echo signal estimation
The vector X (n) of the input signal at the current time n is passed through an adaptive filter to obtain the output value of the adaptive filter, namely the estimated value y (n) of the echo signal,
y(n)＝XT(n)W(n)
where w (n) is the tap weight vector of the adaptive filter at the current time n, w (n) ═ w1(n),w2(n),...,wl(n),...,wL(n)]T，wl(n) is the first tap weight coefficient of the adaptive filter, and the initial value of W (n) is a zero vector;
a3 echo cancellation
Sampling a near-end microphone to obtain a near-end signal d (n) with echo at the current time n, subtracting an estimated value y (n) of the echo signal from the near-end microphone to obtain an error signal e (n) at the current time n, wherein e (n) is d (n) -y (n), and sending the error signal e (n) at the current time n back to the far end;
B. tap weight vector update
B1 calculation of residual signal squared sequence
From the current time N to the time N-Nw+1 residual signals e (N), e (N-1),. -, e (N-N)w+1) to obtain the current time N to the time N-NwThe square e of the residual signal between +12(n),e2(n-1),...,e2(n-Nw+1) to obtain a residual signal squared sequence E in the estimation window of the current time n2(n)，
E2(n)＝[e2(n),e2(n-1),...,e2(n-Nw+1)]
Wherein N iswThe length of the estimation window is in a value range of 5-9;
b2 calculation of quantized residual in weighting
Estimating the squared sequence E of the residual signal in the window from the current time n2(n) calculating a weighted median localization residual σ (n) at the current time n:
wherein λ is a forgetting factor, and the value range thereof is usually 0.800-0.999, c is a weighting coefficient, and c is 1.483(1+ 5/(N)w-1)), med (·) represents an operation taking an intermediate value, the initial value of σ (n) being zero;
b3, calculation of M estimated value
According to the weighted median localization residual sigma (n) of the current time n, obtaining an M estimation threshold value xi, xi-2.576 sigma (n) of the current time; the residual M estimate function value ψ (n) for the current time n is then calculated by the following equation:
b4 calculation of zero attraction vector
By the l-th tap weight coefficient w of the filter at the current time nl(n) approximating l0Norm limit calculation to obtain the zero attraction factor f of the ith tap weight coefficient at the current time nl(n)：
Wherein beta is a zero attraction parameter, the value of beta is 5-50, and sgn (·) represents symbolic operation;
zero attraction factor f of all tap weight coefficients of the current time n1(n),f2(n),...,fl(n),...,fL(n) a zero attraction vector f (n) at the current time n, f (n) ═ f1(n),f2(n),...,fl(n),...,fL(n)]；
B5 updating of filter tap weight vector
The filter tap weight vector W (n +1) for the next time instant n +1 is updated from:
mu represents the step length of the filter, the value range of mu is 0-1, and | | · | | represents the Euclidean norm; gamma ray1Is the ambient noise variance, γ, of the near-end signal d (n)2Is the ambient noise variance of the far-end input signal x (n); rho is a balance parameter and has a value range of 1 multiplied by 10-5～1×10-2；
C. Repetition of
Let n be n +1, repeat the procedure of step A, B until the call is ended.
Compared with the prior art, the invention has the beneficial effects that:
firstly, the invention introduces an M estimation method, so that the tap weight vector is updated by using the M estimation function value psi (n) of the error instead of directly using the error signal e (n). Weighting and carrying out position conversion processing in a time window on an error signal e (n) to obtain a dynamic weighted position conversion residual error and an error M estimation threshold value; in the updating process, when the error signal e (n) exceeds the M estimation threshold value of the error, the impact noise exists in the system, the M estimation function value psi (n) of the error returns to zero, and thus the tap weight vector is not updated at the current moment, so that the interference of the impact noise is effectively avoided, and the impact resistance effect is good; when the error signal e (n) is less than or equal to the M estimation threshold value of the error, judging that the system has no impact noise, and the M estimation function value psi (n) of the error is equal to the error signal e (n); the filter can well follow the change of the error, the steady state error is small, the convergence rate is high, and the echo cancellation effect is good.
Secondly, zero attraction factor (vector) is added in the tap weight vector updating formula, and when the tap weight coefficient is close to zero
Then, the corresponding zero attraction factor β sgn (w) is subtracted from the tap weight vector update formulal(n))-β2wl(n), realizing zero attraction (fast zero return) of the tap weight coefficient close to zero in the sparse system of voice communication, and the tap weight coefficient not close to zero, wherein the corresponding zero attraction factor is zero, namely, the zero attraction item is not subtracted in the tap weight vector updating formula, and the non-zero tap weight coefficient is better preserved; therefore, the problem of low speed of a sparse system with a large number of zero elements, namely voice communication, is solved better, and the convergence rate of the algorithm is improved.
The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
Drawings
FIG. 1 is a diagram of a remote Gaussian signal used in simulation experiments of the present invention.
FIG. 2 is a diagram of the remote colored signals used in the simulation experiments of the present invention.
FIG. 3 is a normalized steady state detuning curve for the comparison method and the method of the present invention at the input of a remote Gaussian signal.
FIG. 4 is a normalized steady state detuning curve for the comparative method and the inventive method at the input of a far-end colored signal.
Detailed Description
Examples
In a specific embodiment of the present invention, an adaptive communication echo cancellation method for voice communication includes the following steps:
A. echo cancellation
A1, remote signal acquisition
Sampling a signal transmitted from a far end to obtain a discrete value x (n) of a far end input signal at the current moment n; input signals x (n), x (n-1),. and x (n-L +1) from a current time n to a time n-L +1 are combined to form an adaptive filter input vector x (n) at the current time n, x (n) ([ x (n), x (n-1),. and x (n-L +1)]T(ii) a Wherein, L is 512, which represents the number of filter taps, and T represents the transposition operation;
a2 echo signal estimation
The vector X (n) of the input signal at the current time n is passed through an adaptive filter to obtain the output value of the adaptive filter, namely the estimated value y (n) of the echo signal,
y(n)＝XT(n)W(n)
where w (n) is the tap weight vector of the adaptive filter at the current time n, w (n) ═ w1(n),w2(n),...,wl(n),...,wL(n)]T，wl(n) is the first tap weight coefficient of the adaptive filter, and the initial value of W (n) is a zero vector;
a3 echo cancellation
Sampling a near-end microphone to obtain a near-end signal d (n) with echo at the current time n, subtracting an estimated value y (n) of the echo signal from the near-end microphone to obtain an error signal e (n) at the current time n, wherein e (n) is d (n) -y (n), and sending the error signal e (n) at the current time n back to the far end;
B. tap weight vector update
B1 calculation of residual signal squared sequence
From the current time N to the time N-Nw+1 residual signals e (N), e (N-1),. -, e (N-N)w+1) to obtain the current time N to the time N-NwThe square e of the residual signal between +12(n),e2(n-1),...,e2(n-Nw+1) to obtain a residual signal squared sequence E in the estimation window of the current time n2(n)，
E2(n)＝[e2(n),e2(n-1),...,e2(n-Nw+1)]
Wherein N iswThe length of the estimation window is in a value range of 5-9;
b2 calculation of quantized residual in weighting
Estimating the squared sequence E of the residual signal in the window from the current time n2(n) calculating a weighted median localization residual σ (n) at the current time n:
wherein λ is a forgetting factor, and the value range thereof is usually 0.800-0.999, c is a weighting coefficient, and c is 1.483(1+ 5/(N)w-1)), med (·) represents an operation taking an intermediate value, the initial value of σ (n) being zero;
b3, calculation of M estimated value
According to the weighted median localization residual sigma (n) of the current time n, obtaining an M estimation threshold value xi, xi-2.576 sigma (n) of the current time; the residual M estimate function value ψ (n) for the current time n is then calculated by the following equation:
b4 calculation of zero attraction vector
By the l-th tap weight coefficient w of the filter at the current time nl(n) approximating l0Norm limit calculation to obtain the zero attraction factor f of the ith tap weight coefficient at the current time nl(n)：
Wherein beta is a zero attraction parameter, the value of beta is 5-50, and sgn (·) represents symbolic operation;
zero attraction factor f of all tap weight coefficients of the current time n1(n),f2(n),...,fl(n),...,fL(n) a zero attraction vector f (n) at the current time n, f (n) ═ f1(n),f2(n),...,fl(n),...,fL(n)]；
B5 updating of filter tap weight vector
The filter tap weight vector W (n +1) for the next time instant n +1 is updated from:
mu represents the step length of the filter, the value range of mu is 0-1, and | | · | | represents the Euclidean norm; gamma ray1Is the ambient noise variance, γ, of the near-end signal d (n)2Is the ambient noise variance of the far-end input signal x (n); rho is a balance parameter and has a value range of 1 multiplied by 10-5～1×10-2；
C. Repetition of
Let n be n +1, repeat the procedure of step A, B until the call is ended.
Simulation experiment
In order to verify the effectiveness of the method, a simulation experiment is carried out, and a method without introducing M estimation and a zero attraction factor is used as a comparison method to be compared with the method. The updating formula of the filter tap weight vector of the comparison method is as follows:
the far-end signals x (n) of the simulation experiment are the gaussian signal x' (n) with zero mean variance of 1 of fig. 1 and the colored signal x "(n) of fig. 2, respectively. The colored signal x "(n) of fig. 2 is the gaussian signal x '(n) of fig. 1 generated by a first-order autoregressive process x" (n) ═ x' (n) +0.8 × x "(n-1), i.e., the current time value and the previous time of the colored signal are correlated. The number of sampling points of the far-end signal is 40000. The echo channel impulse response vector h is measured in a quiet closed room with the width of 3.75m, the height of 2.5m, the length of 6.25m, the temperature of 20 ℃ and the humidity of 50%, and the impulse response length, namely the number L of filter taps is 512.
Background noise of experiment: the background noise in the far-end signal x (n) is white gaussian noise with zero mean variance of 0.05. The background noise in the near-end signal d (n) is also white gaussian noise with zero mean variance of 0.1, and the near-end signal d (n) also includes impulse noise with an occurrence frequency of 0.02 (the impulse noise is generated by the simulation of bernoulli gaussian signal).
The performance of two different echo cancellation methods is measured by using normalized steady state imbalance (NMSD) in the simulation experiment, and the formula is as follows:
the far-end signal and the corresponding near-end signal are subjected to echo cancellation by using the method and the comparison method of the invention. The values of the parameters of the two methods are shown in table 1.
Table 1 parameter values for the two methods of experiment
The simulation experiment obtains a simulation result by independently operating for 100 times.
FIG. 3 is a normalized steady-state imbalance curve for the comparison method and the method of the present invention when the far-end signal is a far-end Gaussian signal. As can be seen from FIG. 3, when the input signal is Gaussian, the comparison method is stabilized at about-11 dB and the method of the present invention is stabilized at about-17 dB under the condition that the convergence rates are approximately the same; the steady state error of the method of the invention is 6dB lower than that of the contrast method.
FIG. 4 is a normalized steady state offset curve for the comparison method and the method of the present invention when the far end signal is a far end colored signal. It can be seen from fig. 4 that when the input signal is a colored signal, the present invention still achieves better convergence performance, the comparison method is stabilized at about-7.5 dB, the method of the present invention is stabilized at about-16 dB, and the steady-state error of the method of the present invention is 8.5dB lower than that of the comparison method. In addition, in fig. 3 and 4, the curve of the method of the present invention is smoother after convergence, which also indicates that the method has better resistance to impulse noise and better echo cancellation effect.
Claims (1)
1. An adaptive communication echo cancellation method for voice communication, comprising the steps of:
A. echo cancellation
A1, remote signal acquisition
Sampling a signal transmitted from a far end to obtain a discrete value x (n) of a far end input signal at the current moment n; input signals x (n), x (n-1),. and x (n-L +1) from a current time n to a time n-L +1 are combined to form an adaptive filter input vector x (n) at the current time n, x (n) ([ x (n), x (n-1),. and x (n-L +1)]T(ii) a Wherein, L is 512, which represents the number of filter taps, and T represents the transposition operation;
a2 echo signal estimation
The vector X (n) of the input signal at the current time n is passed through an adaptive filter to obtain the output value of the adaptive filter, namely the estimated value y (n) of the echo signal,
y(n)＝XT(n)W(n)
where w (n) is the tap weight vector of the adaptive filter at the current time n, w (n) ═ w1(n),w2(n),...,wl(n),...,wL(n)]T，wl(n) is the first tap weight coefficient of the adaptive filter, and the initial value of W (n) is a zero vector;
a3 echo cancellation
Sampling a near-end microphone to obtain a near-end signal d (n) with echo at the current time n, subtracting an estimated value y (n) of the echo signal from the near-end microphone to obtain an error signal e (n) at the current time n, wherein e (n) is d (n) -y (n), and sending the error signal e (n) at the current time n back to the far end;
B. tap weight vector update
B1 calculation of residual signal squared sequence
From the current time N to the time N-Nw+1 residual signals e (N), e (N-1),. -, e (N-N)w+1) to obtain the current time N to the time N-NwThe square e of the residual signal between +12(n),e2(n-1),...,e2(n-Nw+1) to obtain a residual signal squared sequence E in the estimation window of the current time n2(n)，
E2(n)＝[e2(n),e2(n-1),...,e2(n-Nw+1)]
Wherein N iswThe length of the estimation window is in a value range of 5-9;
b2 calculation of quantized residual in weighting
Estimating the squared sequence E of the residual signal in the window from the current time n2(n) calculating a weighted median localization residual σ (n) at the current time n:
wherein, λ is forgetting factor, the value range is 0.800-0.999, c is weighting coefficient, c is 1.483(1+ 5/(N)w-1)), med (·) represents an operation taking an intermediate value, the initial value of σ (n) being zero;
b3, calculation of M estimated value
According to the weighted median localization residual sigma (n) of the current time n, obtaining an M estimation threshold value xi, xi-2.576 sigma (n) of the current time; the residual M estimate function value ψ (n) for the current time n is then calculated by the following equation:
b4 calculation of zero attraction vector
By the l-th tap weight coefficient w of the filter at the current time nl(n) approximating l0Norm limit calculation to obtain the zero attraction factor f of the ith tap weight coefficient at the current time nl(n)：
Wherein beta is a zero attraction parameter, the value of beta is 5-50, and sgn (·) represents symbolic operation;
zero attraction factor f of all tap weight coefficients of the current time n1(n),f2(n),...,fl(n),...,fL(n) a zero attraction vector f (n) at the current time n, f (n) ═ f1(n),f2(n),...,fl(n),...,fL(n)]；
B5 updating of filter tap weight vector
The filter tap weight vector W (n +1) for the next time instant n +1 is updated from:
mu represents the step length of the filter, the value range of mu is 0-1, and | | · | | represents the Euclidean norm; gamma ray1Is the ambient noise variance, γ, of the near-end signal d (n)2Is the ambient noise variance of the far-end input signal x (n); rho is a balance parameter and has a value range of 1 multiplied by 10-5～1×10-2；
C. Repetition of
Let n be n +1, repeat the procedure of step A, B until the call is ended.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911043026.3A CN110572525B (en) | 2019-10-30 | 2019-10-30 | Self-adaptive communication echo cancellation method for voice communication |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911043026.3A CN110572525B (en) | 2019-10-30 | 2019-10-30 | Self-adaptive communication echo cancellation method for voice communication |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110572525A CN110572525A (en) | 2019-12-13 |
| CN110572525B true CN110572525B (en) | 2021-05-07 |
Family
ID=68786053
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911043026.3A Active CN110572525B (en) | 2019-10-30 | 2019-10-30 | Self-adaptive communication echo cancellation method for voice communication |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110572525B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111028856B (en) * | 2020-01-08 | 2022-01-28 | 西南交通大学 | Echo cancellation method with variable step length |
| CN111243568B (en) * | 2020-01-15 | 2022-04-26 | 西南交通大学 | Convex constraint self-adaptive echo cancellation method |
| CN112866149B (en) * | 2020-12-15 | 2022-01-18 | 北京理工大学 | Adaptive estimation method based on bias compensation block sparse proportional NLMS |
| CN113037661B (en) * | 2021-03-01 | 2022-05-13 | 重庆邮电大学 | Sparse LMS (least mean square) method combining zero attraction punishment and attraction compensation |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107071195A (en) * | 2017-03-21 | 2017-08-18 | 西南交通大学 | The exponential function echo cancel method attracted based on a norm zero |
| CN109040499A (en) * | 2018-08-14 | 2018-12-18 | 西南交通大学 | A kind of adaptive echo cancellation method of shock resistance interference |
| CN109040497A (en) * | 2018-07-24 | 2018-12-18 | 西南交通大学 | A kind of proportional class illumination-imitation projection self-adoptive echo cancel method based on M estimation |
| CN109151237A (en) * | 2018-08-23 | 2019-01-04 | 西南交通大学 | The illumination-imitation projection self-adoptive echo cancel method attracted based on zero |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6996229B2 (en) * | 2003-07-09 | 2006-02-07 | Utah State University | Echo cancellation filter |
| US20140310326A1 (en) * | 2013-04-10 | 2014-10-16 | King Fahd University Of Petroleum And Minerals | Adaptive filter for system identification |
| US9491306B2 (en) * | 2013-05-24 | 2016-11-08 | Broadcom Corporation | Signal processing control in an audio device |
| CN107871510B (en) * | 2017-07-07 | 2019-11-26 | 西南交通大学 | A kind of zero attracts the maximum entropy echo cancel method for becoming core width |
| CN107610714A (en) * | 2017-09-13 | 2018-01-19 | 西南交通大学 | The echo cancel method of the minimum cube absolute value attracted based on a norm zero |
| CN108877824B (en) * | 2018-05-31 | 2022-03-25 | 西南交通大学 | Combined step echo cancellation method with high tracking performance |
| CN109102794A (en) * | 2018-07-24 | 2018-12-28 | 西南交通大学 | M based on convex combination estimates the echo cancel method of proportional class affine projection |
| CN109089004B (en) * | 2018-08-14 | 2020-10-09 | 西南交通大学 | Collective member self-adaptive echo cancellation method based on correlation entropy induction |
| CN109697986B (en) * | 2018-09-19 | 2020-12-18 | 四川大学 | Adaptive bias compensation echo cancellation method based on minimum cubic absolute value |
| CN110191245B (en) * | 2019-07-10 | 2021-06-22 | 西南交通大学 | Self-adaptive echo cancellation method based on time-varying parameters |
-
2019
- 2019-10-30 CN CN201911043026.3A patent/CN110572525B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107071195A (en) * | 2017-03-21 | 2017-08-18 | 西南交通大学 | The exponential function echo cancel method attracted based on a norm zero |
| CN109040497A (en) * | 2018-07-24 | 2018-12-18 | 西南交通大学 | A kind of proportional class illumination-imitation projection self-adoptive echo cancel method based on M estimation |
| CN109040499A (en) * | 2018-08-14 | 2018-12-18 | 西南交通大学 | A kind of adaptive echo cancellation method of shock resistance interference |
| CN109151237A (en) * | 2018-08-23 | 2019-01-04 | 西南交通大学 | The illumination-imitation projection self-adoptive echo cancel method attracted based on zero |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110572525A (en) | 2019-12-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110572525B (en) | Self-adaptive communication echo cancellation method for voice communication | |
| CN109040499B (en) | Adaptive echo cancellation method for resisting impact interference | |
| CN111028856B (en) | Echo cancellation method with variable step length | |
| CN106533500B (en) | A method of optimization Echo Canceller convergence property | |
| CN110191245B (en) | Self-adaptive echo cancellation method based on time-varying parameters | |
| US20110261950A1 (en) | Adaptive filter and echo canceller having this adaptive filter | |
| CN109102794A (en) | M based on convex combination estimates the echo cancel method of proportional class affine projection | |
| CN109040497B (en) | Proportional affine projection self-adaptive echo cancellation method based on M estimation | |
| CN114172490B (en) | Ekblom norm-based robust self-adaptive noise elimination method | |
| CN108877824B (en) | Combined step echo cancellation method with high tracking performance | |
| CN104158512B (en) | The adaptive sparse system identifying method that a kind of shock resistance based on the independent activities factor is disturbed | |
| CN109151237B (en) | Affine projection self-adaptive echo cancellation method based on zero attraction | |
| CN109089004B (en) | Collective member self-adaptive echo cancellation method based on correlation entropy induction | |
| CN107871510B (en) | A kind of zero attracts the maximum entropy echo cancel method for becoming core width | |
| CN104410762B (en) | Sane echo cancelltion method in hands-free speaking system | |
| CN109697986B (en) | Adaptive bias compensation echo cancellation method based on minimum cubic absolute value | |
| CN105721729A (en) | Echo cancellation method based on block-sparse proportionate reuse weight coefficient affine projection | |
| CN111243568B (en) | Convex constraint self-adaptive echo cancellation method | |
| CN110767245B (en) | Voice communication self-adaptive echo cancellation method based on S-shaped function | |
| CN111739548B (en) | Minimum error entropy self-adaptive echo cancellation method for resisting input noise | |
| CN113873090B (en) | Robust estimation affine projection spline self-adaptive echo cancellation method | |
| CN114400016A (en) | Echo cancellation method based on adaptive decorrelation and variable-step proportional M estimation | |
| CN113409806B (en) | Zero-attraction echo cancellation method based on arctangent function | |
| WO1999041848A1 (en) | Methods and apparatus for controlling filter adaption in noise | |
| Sudhir et al. | Acoustic echo cancellation using adaptive algorithms |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |