JP6356087B2 - Echo canceling apparatus, method and program - Google Patents

Description

  In the present invention, M (where M is an integer of 1 or more) speakers and one or more microphones are arranged in a common sound field, and when a received signal is reproduced from the speakers, the microphones are connected to the microphones via an echo path. The present invention relates to a technique for canceling an acoustic echo that wraps around (hereinafter simply referred to as “echo”), and more particularly to a technique for canceling an echo in a loudspeaker communication system such as a video conference system.

  The received signal is reproduced by the speaker, and the sound is picked up by the microphone to generate an echo. If it is transmitted as it is, problems such as trouble of telephone conversation and discomfort arise. Further, howling occurs when the volume of the speaker or microphone is high, making it impossible to make a call. In particular, such a problem becomes conspicuous in the voice call system.

  In order to solve this problem, there is an echo canceller that cancels echoes using an adaptive filter as a prior art. Non-Patent Document 1 is known as a conventional multi-channel echo cancellation method. A conventional multi-channel echo canceling apparatus 80 will be described with reference to FIG.

Speaker ₂ 1, ..., _{2 M} and the microphone ₃ 1, ..., _{3 N} are arranged in a common sound field, speaker ₂ 1, ..., respectively, from _{2 M} received signals _{x 1 (k), ...,} x M (k ), The echo canceling unit 8 _n in the multi-channel echo canceling device 80 cancels the playback sound that wraps around the microphone 3 _n via the M echo paths h _mn (k). However, M is an integer greater than or equal to 1, N is an integer greater than or equal to 1, m = 1, ..., M and n = 1, ..., N. Multi-channel echo canceller 80, receiving terminal ₁ 1, ..., and _{1 M,} transmitter terminal ₄ 1, ..., ₄ and _N, the microphone ₃ 1, ..., _{3 N} and are connected, the received signal x ₁ (k), ..., x _M (k) and the collected sound signals y ₁ (k), ..., y _N (k) are inputted, and the transmission signals u ₁ (k), ..., u _N (k) are respectively inputted. Output to the transmitting terminals 4 ₁ ,..., 4 _N. The multi-channel echo cancellation apparatus 80 includes N echo cancellation units 8 ₁ ,..., 8 _N , and the echo cancellation unit 8 _n includes an echo prediction unit 81, a subtraction unit 82, and an echo path estimation unit 83. . In FIG. 1, y _n (k) is y (k), u _n (k) is u (k), and h _1n (k),..., H _Mn (k) are h ₁ (k),. , h _M (k). Can also perform the same processing for the sound signals picked up from the other microphones, it is only necessary arranged in parallel configuration of the echo canceling portion 8 _n of FIG. 1, the following will be described with reference to FIG.

In the echo canceling unit 8 _n , the echo prediction unit 81 filters the received signals x ₁ (k),..., X _M (k) with an adaptive filter to generate a predicted echo signal y ′ (k). In the subtracting unit 82, a difference (hereinafter referred to as “error signal”) u (k) between the collected sound signal y (k) and the predicted echo signal y ′ (k) is obtained and output as a transmission signal. Further, the echo path estimation unit 83 sequentially estimates the echo path from the error signal u (k) and the received signals x ₁ (k),..., X _M (k), and this estimation result (filter coefficient h of the adaptive filter) '(k)) is copied to the echo prediction unit 81. When the echo path is estimated accurately, the echo component contained in the collected sound signal y (k) and the predicted echo signal y '(k) are almost equal, and the error signal u (k) contains almost no echo. It will not be.

  However, in a situation where a multi-channel echo canceller is actually used, it is not always possible to sufficiently cancel the echo, and a residual echo may occur, resulting in a deterioration in the speech quality. This is because the echo path is constantly fluctuating due to human movement and the like, and the echo path estimation by the adaptive filter is not completed instantaneously. This is because the estimation of the echo path can be slightly disturbed in the double talk state.

  Further, when the received signal is multi-channel, since the correlation between the received signals is high, the estimated echo path may not always match the true echo path even if the echo is canceled. Therefore, when the speaker changes and the cross-correlation between the received signals changes, the residual echo can suddenly increase (see Non-Patent Document 1).

  In order to realize a comfortable loud voice call, it is necessary to quickly reduce the residual echo regardless of the number of channels of the received signal and the conversation state in a state where the echo path estimation and cancellation by the adaptive filter is not sufficient. Non-Patent Document 2 is known as a method of estimating transfer characteristics from a received signal to a residual echo at high speed and subtracting the residual echo from an error signal in order to reduce the residual echo regardless of the number of channels and the conversation state. In this method, the transfer characteristic is estimated by using the correlation between the received signal and the error signal for each frequency, thereby speeding up the estimation and suppressing the estimated fluctuation caused by signals other than the residual echo. Since the amplitude and phase are estimated with respect to the transfer characteristic and residual echo, the present invention can be applied regardless of the number of channels. In addition, since residual echo is eliminated by subtraction, distortion of transmitted sound quality can be reduced even during double talk.

  In Non-Patent Document 2, the residual echo needs to be obtained with high accuracy. However, since the residual echo is estimated from the reception signal and error signal of a limited time length (short time interval), the estimation variation is larger than when the time length is sufficiently long, and the residual echo is estimated larger. May end up.

  In order to improve the quality of transmission, it is necessary to improve the estimation accuracy of residual echo even in the above situation. For this purpose, Patent Literature 1 proposes a method for correcting the residual echo estimation value.

JP 2012-227466 A

M.M.Sondhi, D.R.Morgan, and J.L.Hall, “Stereophonic Acoustic Echo Cancellation-An Overview of the Fundamental Problem”, IEEE Signal Processing Letters, AUGUST 1995, vol.2, no.8, pp.148-151 Satoshi Emura and Yoichi Haneda, "Multi-channel echo cancellation using multi-stage echo estimation", Proc. Of the Acoustical Society of Japan, 2010, pp.717-719

  However, even if the power of the estimated residual echo (estimated value) is equal to the power of the error signal output from the adaptive filter, the effect of the residual echo cancellation process may be weak. This is when the phase error of the corrected residual echo estimation value is not small, so that the difference signal power does not decrease even if the residual echo estimation value is subtracted from the adaptive filter output signal. One reason is that the reverberation time assumed by the model of the residual echo cancellation processing is several tens of ms, which is set to be considerably shorter than the reverberation time of the actual room (several hundred ms). Another reason is that when the adaptive filter learning progresses and the part corresponding to the above several tens of ms is well erased by the adaptive filter, the influence of the difference in the assumed reverberation time becomes more prominent. It is.

  The object of the present invention is to provide an echo cancellation technique that can reduce the residual echo in the state (the situation where the residual echo cancellation processing is less effective because the phase error of the residual echo estimation value is not small). It is to be. In the present invention, the residual echo means all echo components included in the collected sound signal, and means an echo component remaining in the error signal after performing echo cancellation by an adaptive filter on the collected sound signal. It is a concept that means not only the echo component but also all echo components included in the collected sound signal when the echo cancellation by the adaptive filter is not performed.

  In order to solve the above problems, according to one aspect of the present invention, an echo canceller reproduces a received signal from a speaker, in which one or more speakers and one or more microphones are arranged in a common sound field. In this case, the echo that goes around the microphone via the echo path is eliminated. The echo canceller uses a frequency domain sound pickup signal that is a signal obtained from the first sound pickup signal picked up by the microphone and a frequency domain reception signal that is a frequency domain signal obtained from the reception signal. A residual echo estimator that calculates the residual echo estimate in consideration of the phase and amplitude of the residual echo contained in the collected sound signal, and a residual echo cancellation suppressor that eliminates and suppresses the residual echo estimate from the frequency domain collected signal The residual echo cancellation suppression unit subtracts the residual echo estimate from the frequency domain collected signal to obtain a difference, and the smaller the difference, the higher the ratio of suppressing the residual echo from the frequency domain collected signal. Reduce the rate of echo cancellation.

  In order to solve the above problems, according to another aspect of the present invention, an echo canceling method reproduces a received signal from a speaker by arranging one or more speakers and one or more microphones in a common sound field. When this occurs, the echo that goes around the microphone via the echo path is eliminated. The echo canceling method uses a frequency domain sound pickup signal that is a signal obtained from the first sound pickup signal picked up by the microphone and a frequency domain reception signal that is a frequency domain signal obtained from the reception signal. Considering the phase and amplitude of the residual echo contained in the collected sound signal, a residual echo estimation step for obtaining a residual echo estimate value, and a residual echo cancellation suppression step for eliminating and suppressing the residual echo estimate value from the frequency domain collected signal In the residual echo cancellation suppression step, the residual echo estimation value is subtracted from the frequency domain collected signal to obtain a difference, and the smaller the difference, the higher the ratio of suppressing the residual echo from the frequency domain collected signal. Reduce the rate of echo cancellation.

  The echo cancellation technique according to the present invention has an effect that the residual echo can be reduced as compared with the conventional technique in a situation where the effect of the residual echo cancellation process is weak because the phase error of the residual echo estimation value is not small.

The figure which shows the structural example of the conventional multichannel echo cancellation apparatus 80. FIG. The figure which shows distribution of cancellation / suppression of a residual echo. 1 is a diagram illustrating a configuration example of an echo cancellation apparatus 100. FIG. The figure which shows the example of the processing flow of the echo cancellation apparatus. The figure which shows the structural example of 16A of residual echo estimation parts. The figure which shows the structural example of the residual echo cancellation suppression part 169. FIG. The figure which shows the example of the processing flow of the residual echo estimation part 16A. The figure which shows the structural example of the input-output correlation coefficient calculation part 163. The figure which shows the example of the processing flow of the residual echo cancellation suppression part 169. FIG. 1 is a diagram illustrating a configuration example of an echo canceller 200. FIG. The figure which shows the example of the processing flow of the echo cancellation apparatus. Diagram illustrating an exemplary configuration of the echo cancellation unit 28 _n. It shows an example of a process flow of the echo canceling portion 28 _n.

  Hereinafter, embodiments of the present invention will be described. In the drawings used for the following description, constituent parts having the same function and steps for performing the same process are denoted by the same reference numerals, and redundant description is omitted. In the following description, the symbol “^” or the like used in the text should be described immediately above the character immediately before, but it is described immediately after the character due to restrictions on text notation. In the formula, these symbols are written in their original positions. Further, the processing performed for each element of a vector or matrix is applied to all elements of the vector or matrix unless otherwise specified.

<Points of first embodiment>
In the residual echo cancellation of Patent Document 1, the phase and amplitude of the residual echo are estimated, and the residual echo is canceled by subtraction. On the other hand, there is a technique called echo suppression in which only the amplitude of the residual echo is estimated and the signal is attenuated by multiplication at each frequency by an amount corresponding to the amplitude (Reference Document 1).
(Reference 1) Japanese Patent Laid-Open No. 11-331046

  In the present embodiment, residual echo cancellation by this subtraction is combined with echo suppression by multiplication. Specifically, when the difference between the power of the collected sound signal and the power of the residual echo estimation value is small, the mixed mode of echo cancellation and suppression is entered. As described above, when the power difference is small, there is a high possibility that the phase error is not small, and even if the residual echo estimation value is subtracted by canceling the residual echo when the phase error is not small, the residual echo is often hardly reduced. . . On the other hand, when the power difference is small, the possibility of being in the double talk state is low, and the possibility of distortion of the transmission sound quality due to echo suppression is low. Therefore, in the mixed mode, the smaller the power difference, the lower the echo cancellation distribution and the higher the echo suppression distribution, thereby reducing the residual echo than before.

An example of this process will be described with reference to FIG. Power Y (f, j) ² and the residual echo estimate Y ^ ₂ (f, j) power Y ^ ₂ (f, j) of the horizontal axis in FIG. 2 is a frequency domain sound pickup signal Y (f, j) ² Is divided by the power Y (f, j) ² of the frequency domain collected signal Y (f, j), and the vertical axis represents the share ratio R _mxc (f, j) of residual echo cancellation. Frequency domain sound collection signal Y (f, j) power Y (f, j) ² and residual echo estimate Y ^ ₂ (f, j) power Y ^ ₂ (f, j) of the ^second difference is smaller situation

Detect with. However,

It is. Here, p_hyb_range_upper is a threshold value for detection, and a value is set in a range of −20 to −10 dB. MAX (a, b) is a function that returns the larger value of a and b. When the residual echo is estimated to be larger (the power Y ^ _{2 of the} residual echo estimate Y ^ ₂ (f, j) (f, j) power Y (f, j) of ² frequency domain sound pickup signal Y (f, j) ² or more time), function J returns 0.

In situations where the difference is small, the residual echo cancellation share ratio R _mxc (f, j) is

Calculate with Here, the parameter p_cancel_allotted_min is a ratio when the share of residual echo cancellation is minimized, and a value in the range of 0 to 1 is set.

Residual echo suppression shares the remaining (1-R _mxc (f, j)). The amount of suppression is converted to amplitude

Echo suppression gain G _mxs (f, j)

Can be used.

  Hereinafter, a configuration for realizing the above-described processing will be described.

<First embodiment>
<Echo canceling apparatus 100>
FIG. 3 shows an example of a functional block diagram of the echo cancellation apparatus 100 according to the first embodiment, and FIG. 4 shows a processing flow thereof. The echo cancellation apparatus 100 according to the first embodiment will be described with reference to FIGS. 3 and 4. M speakers ₂ 1, ..., _{2 M} and N microphones ₃ 1, ..., _{3 N} are arranged in a common sound field, speaker ₂ 1, ..., respectively, from _{2 M} received signals x ₁ (k), .., X _M (k) is reproduced, the echo canceller 100 cancels the reproduced sound (echo) that wraps around the microphone via M × N echo paths h _mn (k). More specifically, the residual echo canceling unit 16 _n in the echo canceling apparatus 100 cancels the reproduced sound (echo) that wraps around the microphone 3 _n via the M echo paths h _mn (k). The echo canceling apparatus 100 includes receiving terminals 1 ₁ ,..., 1 _M for all M channels on the receiving side, transmitting terminals 4 ₁ ,..., 4 _N for all N channels on the transmitting side, and microphones 3 ₁ ,. 3 _N is connected, and the received signal x ₁ (k),..., X _M (k) and the collected sound signal y ₁ (k),..., Y _N (k) are input, and the transmitted signal v _{1 (k), ..., v N} (k) , respectively transmitter terminals 4 _1, ..., and outputs a 4 _N.

The echo canceller 100 includes N residual echo cancelers 16 ₁ ,..., 16 _N.

<Residual echo canceller _16n >
The residual echo canceling unit 16 _n is connected to all M channel receiving terminals 1 ₁ ,..., 1 _{M on} the receiving side, one channel transmitting terminal 4 _n on the transmitting side, and a microphone 3 _n . M-channel received signals x ₁ (k),..., X _M (k) and 1-channel sound pickup signal y _n (k) are input, and 1-channel transmitted signal v _n (k) is transmitted to transmission terminal 4. output to _n . In each figure, y _n (k) is y (k), v _n (k) is v (k), and h _1n (k), ..., h _Mn (k) are h ₁ (k), respectively. , ..., h _M (k). In each figure, only the processing unit of the nth channel will be described. Similar processing can be performed on the collected sound signals from other microphones, and the configuration of the processing units of the n-th channel only needs to be arranged in parallel.

The residual echo canceling unit 16 _n includes M frequency domain transforming units 161 ₁ ,..., 161 _M , a frequency domain transforming unit 162, a residual echo estimating unit 16 A, a residual echo canceling suppressing unit 169, and a time domain transforming unit. 168.

  The residual echo estimation unit 16A includes an input / output correlation coefficient calculation unit 163, an input / output transfer characteristic estimation unit 164, a residual echo prediction unit 165, and a residual echo correction unit 166 (see FIG. 5).

  The residual echo erasure suppression unit 169 includes an erasure suppression distribution control unit 1691, an erasure distribution setting unit 1692, a subtraction unit 1693, and a suppression unit 1694 (see FIG. 6).

<Frequency Domain Transformer 161 ₁ ,..., 161 _M and Frequency Domain Transformer 162>
As shown in FIGS. 3 and 4, the frequency domain converter 161 1, _..., 161 _M each received signal x ₁ (k), ..., as input x _M (k), which in a short time interval per frequency domain received signals X ₁ is a signal in the frequency domain (f, j), ..., it is converted to X _M (f, j), and outputs (s161). Similarly, the frequency domain conversion unit 162 receives the collected sound signal y (k) collected by the microphone 3 _n , and the frequency domain collected signal Y (f, j), which is a frequency domain signal for each short time interval. And output (s162). Hereinafter, the collected sound signal y (k) is also referred to as a first collected sound signal y (k) in order to distinguish it from a second collected sound signal u (k) described later.

A case will be described in which each signal is set to 1 frame = 2 L samples, the L / D samples are blocked, and the L / D samples are shifted to create a frame. However, L is an integer greater than or equal to 1, D is an integer which can divide L, j represents a frame number, and is time k = jL / D. f represents a frequency number. For example, f corresponds to a discrete point (frequency bin) obtained by dividing half of the sampling frequency f _s into L equal parts, and f = 0, 1,..., L−1, and f = 0 is Corresponding to frequency 0, f = 1 corresponds to frequency (1 / L) f _s / 2,..., F = L-1 corresponds to ((L-1) / L) f _s / 2.

  The conversion to the frequency domain is performed by, for example, FFT (Fast Fourier transform) or DFT (discrete Fourier transform), and it is preferable to set L to a power of 2 in order to simplify and speed up the calculation. For example, L = 64 to 1024, D = 2 to 8 and the like. The frame length (number of samples included in one frame) may be set so as to correspond to 10 ms to 20 ms.

<Residual echo estimation unit 16A>
The residual echo estimation unit 16A receives the frequency domain collected signal Y (f, j) and the frequency domain received signal X ₁ (f, j),..., X _M (f, j), and uses these values. Taking into account the phase and amplitude of the residual echo contained in the frequency domain collected signal Y (f, j), the estimated residual echo (hereinafter also referred to as the residual echo estimated value) Y ^ ₂ (f, j) Obtain (s16A) and output. FIG. 7 shows an example of the processing flow of the residual echo estimator 16A. The processing of the residual echo estimator 16A will be described with reference to FIGS.

<Input / output correlation coefficient calculation unit 163>
The input / output correlation coefficient calculation unit 163 receives the frequency domain received signal X ₁ (f, j),..., X _M (f, j) and the frequency domain collected signal Y (f, j) as inputs. using the value, the power spectrum P _mm (f, j) of the m channels of the frequency domain received signal X _m (f, j) and the frequency domain received signal X _m (f, j) of the m channels and the m '(Where m' = 1, ..., M, m ≠ m ') and the cross spectrum P _m'm (f, j) with the frequency domain received signal X _m' (f, j) of the channel The cross spectrum Q _{m ′} (f, j) between the frequency domain received signal X _{m ′} (f, j) of the m′-th channel and the frequency domain collected signal Y (f, j) is obtained and output (s163). ).

Each cross spectrum and power spectrum are values at time k = jL / D. The power spectrum P _mm (f, j) represents the autocorrelation coefficient of the input signal (frequency domain received signal X _m (f, j) of the m-th channel), and the cross spectrum P _m'm (f, j) is the input The correlation coefficient between the signals (the frequency domain received signal X _m (f, j) of the m-th channel and the frequency domain received signal X _{m ′} (f, j) of the m-th channel) is represented. The matrix composed of the power spectrum P _mm (f, j) and the cross spectrum P _m′m (f, j) described above is represented as the correlation coefficient P (f, j) of the input signal as follows.

On the other hand, the cross spectrum Q _{m ′} (f, j) includes an input signal (frequency domain received signal X _{m ′} (f, j) of the mth channel) and an output signal (frequency domain collected signal Y (f, j)). ) And the correlation coefficient between input and output Q (f, j)

It expresses. The input / output correlation coefficient calculation unit 163 will be described with reference to FIG. For example, the input / output correlation coefficient calculation unit 163 includes a power spectrum calculation unit 163a, an inter-received signal cross spectrum calculation unit 163b, and an input / output signal cross spectrum calculation unit 163c.

The power spectrum calculation unit 163a calculates a power spectrum P _mm (f, j) using the frequency domain received signal X _m (f, j) of the m-th channel.

The inter-received signal cross spectrum calculation unit 163b uses the M frequency domain received signals X ₁ (f, j),..., X _M (f, j) to generate the m-th channel frequency domain received signal X _m (f , j) and the frequency domain received signal X _{m ′} (f, j) of the _m′-th channel, a cross spectrum P _m′m (f, j) is calculated.

The input / output signal cross spectrum calculation unit 163c uses M frequency domain received signals X ₁ (f, j),..., X _M (f, j) and the frequency domain collected sound signal Y (f, j). And the cross spectrum Q _{m ′} (f, j) between the M frequency domain received signals X ₁ (f, j),..., X _M (f, j) and the frequency domain collected signal Y (f, j) Is calculated.

For example, P _mm (f, j), P _m′m (f, j), Q _{m ′} (f, j) is the frequency domain received signal X _m (f, j) of the m-th channel at time k = jL / D. It is calculated from the following equations (3), (4), and (5) from j) and the frequency domain sound pickup signal Y (f, j), respectively.

X ^* means the complex conjugate of X, and E [] means the average. As an example of the averaging process,

As described above, a method using a smoothing constant β which takes a processing result of one frame before and a value of 0 to 1, a method of multiplying several past frames by a time constant, and the like can be considered. P _mm (f, j) and Q _{m ′} (f, j) can also be obtained by the same method.

<Input / output transfer characteristic estimation unit 164>
The input / output transfer characteristic estimation unit 164 receives the power spectrum P _mm (f, j), the cross spectrum P _m′m (f, j), and Q _{m ′} (f, j) as input, and uses these values. , M frequency domain received signals X ₁ (f, j), ..., X _M (f, j) and frequency domain sound pickup signal Y (f, j) estimated input / output transfer characteristics G (f, j) = [G ₁ (f, j),..., G _M (f, j)] ^T is estimated for each frequency and output (s164).

  For example, the input / output transfer characteristic estimation unit 164 estimates the input / output transfer characteristic estimated value G (f, j) by the following equation (7).

  For the matrix composed of the power spectrum and cross spectrum, in order to stabilize the inverse matrix calculation, a small constant δ is added to the diagonal component,

It is good.

<Residual echo prediction unit 165>
The residual echo prediction unit 165 receives M frequency domain received signals X ₁ (f, j),..., X _M (f, j) and an estimated value G (f, j) of input / output transfer characteristics as inputs, From these values, a residual echo component included in the frequency domain sound pickup signal Y (f, j) is predicted, and an estimated value Y ^ (f, j) is output (s165).

  For example, the residual echo estimate Y ^ (f, j)

To predict.

  It should be noted that the phase of the residual echo component is taken into account when obtaining the input / output correlation coefficients P (f, j) and Q (f, j) by the equations (3) to (5). Furthermore, when obtaining the estimated value G (f, j) of the input / output transfer characteristic from the input / output correlation coefficient P (f, j), Q (f, j) by the equation (7) or (7 ′), The phase and amplitude of the residual echo component are taken into account, and it can be said that the residual echo estimation unit 16A obtains the residual echo estimated value Y ^ (f, j) in consideration of the phase and amplitude of the residual echo.

<Residual echo correction unit 166>
The residual echo correction unit 166 receives the frequency domain sound pickup signal Y (f, j) and the residual echo estimated value Y ^ (f, j) as input, and uses the residual echo estimated value Y ^ (f, j). ) To obtain a corrected residual echo estimated value Y ₂ ^ (f, j) and output it (s166). The corrected residual echo estimated value Y ₂ ^ (f, j) can be obtained by the following equation, for example.

However, T is the number of degrees of freedom of estimation of each spectrum, and in the input / output correlation coefficient calculation unit 163, the power spectrum P _mm (f, j), the cross spectrum P _m′m (f, j), Q _{m '} This is the number of frames when calculating (f, j). An appropriate value is set for T before use or after setting the number M of received signal channels so that T-2M> 0. When the ratio η (f, j) <0 is obtained as a result of Expression (11), η (f) = 0 is used instead in Expression (12).

A storage unit (not shown) may store a correspondence between the estimated coherence value γ ^ ² (f) and the ratio η (f) defined by the equation (11). With such a configuration, the calculation time of Expression (11) can be shortened. That is, the residual echo correction unit 166 calculates Equations (9) and (10) using the frequency domain sound pickup signal Y (f, j) and the residual echo estimation value Y ^ (f, j), and provides coherence. The estimated value γ ^ ² (f) is obtained, the ratio η (f) corresponding to the estimated value γ ^ ² (f) obtained from the storage unit (not shown) is extracted, and the residual echo estimated value Y ^ (f, j) is obtained. Multiplying (see equation (12)), a corrected residual echo estimated value Y ₂ ^ (f, j) may be obtained and output. In other words, M and T are constants known in advance, and the ratio η (f) can be regarded as a function of an estimated value γ ^ ² (f) that takes a value between 0 and 1. That is, the ratio η (f) is regarded as a function of the estimated value γ ^ ² (f), and a table can be created by calculating in advance. In actual signal processing, this table is subtracted to obtain the ratio η (f), so that η (f) can be obtained efficiently without calculating √.

<Residual echo cancellation suppressor 169>
The residual echo cancellation suppression unit 169 receives the corrected residual echo estimation value Y ₂ ^ (f, j) and the frequency domain collected signal Y (f, j), and receives the frequency domain collected signal Y (f, j). The corrected residual echo estimated value Y ₂ ^ (f, j) is deleted from the input signal, suppressed (s169), and the transmission signal V (f, j) in the frequency domain is obtained and output. The difference is obtained by subtracting the corrected residual echo estimated value Y ₂ ^ (f, j) from the frequency domain collected signal Y (f, j), and the smaller the difference, the more the frequency domain collected signal Y (f , j), the ratio of suppressing the residual echo is increased and the ratio of canceling the residual echo is decreased. For example, the echo suppression gain is G _mxs (f, j), the ratio of canceling the residual echo is the sharing ratio R _mxc (f, j), and the transmission signal V (f, j) is

Asking. With such a configuration, the distribution of echo cancellation and echo suppression can be set appropriately in order to cancel the residual echo. An example of setting the echo suppression gain G _mxs (f, j) and the sharing ratio R _mxc (f, j) will be described in the erasure suppression distribution control unit 1691 described later. FIG. 9 shows an example of the processing flow of the residual echo cancellation suppressing unit 169. The processing of the residual echo cancellation suppressing unit 169 will be described with reference to FIGS. 6 and 9.

<Erasure Suppression Distribution Control Unit 1691>
When the difference between the power of the frequency domain collected signal Y (f, j) and the estimated residual echo estimated value Y ₂ ^ (f, j) is small, the mixed mode of echo cancellation and suppression is entered. In the mixed mode, the smaller the difference, the lower the echo cancellation distribution and the higher the echo suppression distribution.

An example of this process will be described with reference to FIG. Frequency domain sound collection signal Y (f, j) power Y (f, j) ² and residual echo estimate Y ₂ ^ (f, j) power Y ₂ ^ (f, j) of the ^second difference is smaller situation

Detect with. However, p_hyb_range_upper is a threshold value for detection, and a value is set in the range of -20 to -10 dB. MAX (a, b) is a function that returns the larger value of a and b. When the residual echo is estimated larger than the collected sound signal, the function J returns 0.

The share ratio of residual echo cancellation R _mxc (f, j) is

And is set in the erasure distribution setting unit 1692. The remaining (1-R _mxc (f, j)) is applied to residual echo suppression, and the amount of suppression is | Y ₂ ^ (f, j) | (1-R _mxc (f, j)) in terms of amplitude become. As an example of echo suppression gain,

This is set in the suppression unit 1694.

  The parameter p_cancel_allotted_min is the minimum ratio for canceling the residual echo, and a value in the range of 0 to 1 is set.

Also, when the difference between the collected signal power and the residual echo signal power is not small, that is, p_cancel_allotted_min ≦ J (Y (f, j), Y ^ ₂ (f, j)), R _mxc (f, j) = Set to 1. At this time, G _mxs (f, j) = 1, and only residual echo cancellation is valid.

That is, the erasure suppression distribution control unit 1691 receives the frequency domain sound pickup signal Y (f, j) and the residual echo estimated value Y ^ ₂ (f, j), obtains respective powers, and obtains the frequency domain sound pickup signal. Y (f, j) power Y (f, j) ² from the residual echo estimate Y ^ ₂ (f, j) power _{Y ^ 2 (f, j)} 2 the pulling difference of (Y (f, j) ² -Y ^ ₂ (f, j) ² ), and when the difference (Y (f, j) ² -Y ^ ₂ (f, j) ² ) is smaller than the predetermined threshold p_hyb_range_upper, the sharing ratio R _mxc ( f, j)

When the difference (Y (f, j) ² -Y ^ ₂ (f, j) ² ) is greater than or equal to a predetermined threshold p_hyb_range_upper, the sharing ratio R _mxc (f, j) is set to 1, and the sharing ratio R _mxc ( f, j) is output to the erasure distribution setting unit 1692 (s1691). Further, the erasure suppression distribution control unit 1691 calculates the frequency domain collected signal Y (f, j), the residual echo estimated value Y ^ ₂ (f, j), and the sharing ratio R _mxc (f, j ) To obtain an echo suppression gain G _mxs (f, j) and output it to the suppression unit 1694.

<Erase distribution setting unit 1692>
The erasure distribution setting unit 1692 receives the residual echo estimation value Y ^ ₂ (f, j) and the sharing ratio R _mxc (f, j), and the product Y ^ ₂ (f, j) R _mxc (f, j ) Is obtained (s1692) and output. This process corresponds to a process for setting the residual echo cancellation ratio.

<Subtraction unit 1693>
The subtracting unit 1693 receives the frequency domain sound _pickup signal Y (f, j) and the product Y ^ ₂ (f, j) R _mxc (f, j), and _{produces the} product from the frequency domain sound _pickup signal Y (f, j). _Subtract Y ^ ₂ (f, j) _Rmxc (f, j) to _find the difference {Y (f, j) -Y ^ ₂ (f, j) _Rmxc (f, j)} (s1693) and output To do. This process corresponds to a residual echo cancellation process.

<Suppression unit 1694>
The suppressor 1694 receives the difference {Y (f, j) −Y ^ ₂ (f, j) R _mxc (f, j)} and the echo suppression gain G _mxs (f, j), and the product G _mxs (f , j) {Y (f, j) -Y ^ ₂ (f, j) R _mxc (f, j)} is obtained (s1694), and this product is used as the transmission signal V (f, j) of the frequency meter. ,Output. This process corresponds to the residual echo suppression process.

  Therefore, the residual echo cancellation suppression unit 169 obtains the transmission signal V (f, j) by the following equation (14), for example.

<Time domain conversion unit 168>
As shown in FIG. 3 and FIG. 4, the time domain conversion unit 168 receives the frequency domain transmission signal V (f, j), converts this signal into a time domain signal v (k), and converts this signal v (k). It is output as an output value of the echo canceller 100 (s168). Note that the time domain conversion unit 168 may use a time domain conversion method corresponding to the frequency domain conversion method used in the frequency domain conversion units 161 _m and 162.

<Effect>
With such a configuration, the residual echo power is equal to the collected signal power, but the residual echo is suppressed more than before even in the situation where the residual echo cancellation processing is less effective because the phase error of the corrected residual echo estimate is not small. be able to.

<Modification>
In the first embodiment, the case where M> 1 is mainly described. However, M = 1 may be used. In this case, the input / output correlation coefficient calculation unit 163 cross-spectrums the frequency domain received signal X _m (f, j) of the m-th channel and the frequency domain received signal X _{m ′} (f, j) of the m′-th channel. There is no need to find P _m'm (f, j). The input / output transfer characteristic estimation unit 164 uses the power spectrum P ₁₁ (f, j) and the cross spectrum Q ₁ (f, j), and the frequency domain received signal X ₁ (f, j) and the frequency domain sound pickup signal. An estimated value G (f, j) of input / output transfer characteristics with Y (f, j) is estimated for each frequency and output.

  The residual echo estimator 16A may have other configurations as long as the residual echo estimation value is obtained in consideration of the phase and amplitude of the residual echo included in the frequency domain sound pickup signal Y (f, j). . For example, the residual echo estimation unit 16A does not include the residual echo correction unit 166, and may use the output value (estimated value Y ^ (f, j)) of the residual echo prediction unit 165 as the output value of the residual echo estimation unit 16A. Good.

The point of this embodiment is that the residual echo can be suppressed by combining residual echo cancellation and echo suppression even in a situation where the effect of residual echo cancellation is weak because the phase error of the residual echo estimation value is not small. is there. Therefore, the residual echo cancellation unit 16n may include at least the residual echo estimation unit 16A and the residual echo cancellation suppression unit 169, and other configurations (for example, M frequency domain conversion units 161 ₁ ,..., 161 _M , The frequency domain transform unit 162 and the time domain transform unit 168) are not necessarily included.

<Second embodiment>
A description will be given centering on differences from the first embodiment.

<Echo canceling apparatus 200>
An echo canceling apparatus 200 according to the second embodiment will be described with reference to FIGS. 10 and 11. The echo cancellation apparatus 200 includes N echo cancellation units 28 ₁ ,..., 28 _N and N residual echo cancellation units 26 ₁ ,..., 26 _N , and the echo cancellation unit 28 precedes the residual echo cancellation unit 26 _{n. n} is provided.

<Echo canceling unit 28 _n >
The echo cancellation unit 28 _n, receiving terminal ₁ 1, ..., _{1 M} and the residual echo cancellation unit 26 _n, which is connected to the microphone _{3 n} is the received signal _{x 1 (k), ...,} x M ( k) and the first sound pickup signal y _n (k) are input, and the second sound pickup signal u _n (k) of one channel is output to the residual echo canceling unit 26 _n . Note that an error signal obtained by eliminating the echo component from the first sound collection signal is referred to as a second sound collection signal for convenience.

The echo canceller 28 _n filters the received signals x ₁ (k),..., X _M (k) with an adaptive filter, generates a predicted echo signal y ′ (k), and further picks up the sound with the microphone 3 _n The difference between the first collected signal y (k) and the predicted echo signal y ′ (k) is obtained as the second collected signal u (k), and the second collected signal u (k) and the received signal x ₁ (k ),..., X _M (k) and the filter coefficient h ′ (k) of the adaptive filter is updated (s28).

Details will be described below with reference to FIGS. 12 and 13. The echo erasure unit 28 _n includes an echo prediction unit 281, a subtraction unit 282, and an echo path estimation unit 283.

To illustrate the processing of the echo canceling portion 28 _n, first described the relationship between the received signal and the first voice collecting signal. If the impulse response of the echo path from the speakers 2 ₁ ,..., 2 _M to the microphone 3 _n is h ₁ ,..., H _M (k) and the length is L ₁ , the received signal x ₁ (k),. , x _M (k) and the first collected sound signal y (k) have the following relationship:

M-th channel impulse response h _m and received signal x _m
h _m = [h _m (0)… h _m (L ₁ -1)] ^T (22)
x _m = [x _m (0)… x _m (L ₁ -1)] ^T (23)
As a vector, the relationship between the received signal x ₁ (k),..., X _M (k) and the first collected signal y (k) is described as follows.

y (k) = h ₁ ^T x ₁ (k) +… + h _M ^T x _M (k) (24)
However, T represents transposition.

<Echo Prediction Unit 281>
The echo prediction unit 281 inputs the received signals x ₁ (k),..., X _M (k) to the predicted echo path by the adaptive filter, generates a predicted echo signal y ′ (k), and outputs it (s281). The echo prediction unit 281 includes an adaptive filter, and the characteristic of the adaptive filter is controlled by an echo path estimation unit 283 described later so that the error signal of the subtraction unit 282 in the reception state is minimized.

For example, the filter coefficient of the adaptive filter of the mth channel is
h ' _m = [h' _m (0)… h ' _m (L _E -1)] ^T (25)
And the predicted echo signal
y '(k) = h' ₁ ^T x ₁ (k) +… + h ' _M ^T x _M (k) (26)
Is generated. However, L _E represents a tap length of the adaptive filter. The echo prediction unit 281 outputs the generated predicted echo signal y ′ (k) to the subtraction unit 282. For example, the tap length of the adaptive filter is often set to about 100 to 300 ms.

<Subtraction unit 282>
The subtraction unit 282 receives the first collected sound signal y (k) and the predicted echo signal y ′ (k), subtracts the predicted echo signal y ′ (k) from the first collected sound signal y (k), and The collected sound signal u (k) is obtained (s282).

u (k) = y (k) -y '(k) (27)
And it outputs the obtained second sound pickup signal u (k) of the frequency domain transform section 262 of the echo path estimation unit 283 residual echo canceling portion 26 _n.

<Echo path estimation unit 283>
The echo path estimation unit 283 receives the second collected sound signal u (k) and the received signals x ₁ (k),..., X _M (k) as input, and uses them to use the filter coefficient h ′ (k ) Is updated and output (s283). When the Normalized Least Mean Square algorithm (NLMS algorithm) is used as the coefficient correction method for the adaptive filter, the filter coefficient is updated by the following equation (28).

h ' _m (k + 1) = h' _m (k) + μu (k) x _m (k) (28)
Where μ is the step size,

Determined by. Note that μ ₀ is a parameter that is controlled based on the power of the input signal and is preset to a value of 0 to 1 in order to perform stable estimation. The echo path estimation unit 283 copies the updated filter coefficient h ′ (k + 1) and outputs it to the echo prediction unit 281. The filter coefficient updating method is not limited to the above-described method, and other updating methods may be used.

<Residual echo canceller 26 _n >
The processing performed using the first collected sound signal y _n (k) in the residual echo canceling unit 16 _n of the first embodiment is performed in the above-described second collected sound signal u _n (k) in the residual echo canceling unit 26 _n . To do. For example, the frequency domain conversion unit 262 converts the second collected sound signal u (k) into a frequency domain signal U (f, j), and using this signal, the residual echo estimation unit 26A and the residual echo cancellation suppression unit 269 In FIG. The processing performed by the residual echo estimation unit 26A is the same as that in the first embodiment, but the estimated residual echo value U ^ ₂ (f, j) is the first collected sound signal y _n (k). It is not the residual echo estimation value included, but the residual echo estimation value included in the second collected sound signal u _n (k). The residual echo canceling unit 26 _n erases not the residual echo component included in the first sound pickup signal y _n (k) but the residual echo component included in the second sound pickup signal u _n (k).

<Effect>
With such a configuration, the same effect as that of the first embodiment can be obtained. When there is no large fluctuation in the echo path, the echo canceling unit 28 _n in the previous stage can estimate the echo path with high accuracy, so that the transmission quality is improved. In addition, when the echo path greatly fluctuates, the residual echo component can be canceled by the subsequent residual echo canceller 26 _n until the estimation of the echo path performed in the echo canceler 28 _n is stabilized. Therefore, compared to a device that performs echo cancellation using only an adaptive filter (for example, the multi-channel echo cancellation device 80 in FIG. 1), high transmission quality can be maintained throughout the echo path stabilization and fluctuation.

<Modification>
In this embodiment, the adaptive filter is updated using signals in the time domain (received signals x ₁ (k),..., X _M (k) and the second collected sound signal u _n (k)). The adaptive filter may be updated using a signal in the domain or wave number domain (see Reference 2).
(Reference document 2) JP2013-255155A

In this case, the frequency domain signals (X ₁ (f, j),..., X _M (f, j) and U _n (f, j)) obtained in the calculation process of the echo canceling unit 28 _n are converted into time domain signals. The residual echo canceling unit 26 _n may be directly output to the residual echo canceling unit 26 _n without being converted, in which case the residual echo canceling unit 26 _n does not include the frequency domain converting units 161 ₁ , ..., 161 _M and the frequency domain converting unit 262. In addition, since the calculation cost of the adaptive filter is high, the echo canceling process of the echo canceling unit 28 _n is performed at some frequencies (for example, frequencies 300 Hz to 3.4 kHz or 100 Hz to 7 kHz having a strong auditory influence). only performed, for residual echo cancellation processing of the residual echo canceling portion 26 _n may be configured to perform at all frequencies. with such a configuration, efficiently, thereby improving the transmission quality.

<Program and recording medium>
In addition, various processing functions in each device described in the above embodiments and modifications may be realized by a computer. In that case, the processing contents of the functions that each device should have are described by a program. Then, by executing this program on a computer, various processing functions in each of the above devices are realized on the computer.

  The program describing the processing contents can be recorded on a computer-readable recording medium. As the computer-readable recording medium, for example, any recording medium such as a magnetic recording device, an optical disk, a magneto-optical recording medium, and a semiconductor memory may be used.

  The program is distributed by selling, transferring, or lending a portable recording medium such as a DVD or CD-ROM in which the program is recorded. Further, the program may be distributed by storing the program in a storage device of the server computer and transferring the program from the server computer to another computer via a network.

  A computer that executes such a program first stores, for example, a program recorded on a portable recording medium or a program transferred from a server computer in its storage unit. When executing the process, this computer reads the program stored in its own storage unit and executes the process according to the read program. As another embodiment of this program, a computer may read a program directly from a portable recording medium and execute processing according to the program. Further, each time a program is transferred from the server computer to the computer, processing according to the received program may be executed sequentially. Also, the program is not transferred from the server computer to the computer, and the above-described processing is executed by a so-called ASP (Application Service Provider) type service that realizes the processing function only by the execution instruction and result acquisition. It is good. Note that the program includes information provided for processing by the electronic computer and equivalent to the program (data that is not a direct command to the computer but has a property that defines the processing of the computer).

  In addition, although each device is configured by executing a predetermined program on a computer, at least a part of these processing contents may be realized by hardware.

<Other variations>
The present invention is not limited to the above-described embodiments and modifications. For example, the various processes described above are not only executed in time series according to the description, but may also be executed in parallel or individually as required by the processing capability of the apparatus that executes the processes. In addition, it can change suitably in the range which does not deviate from the meaning of this invention.

  The frequency domain collected signal in the claims is a signal in the frequency domain obtained from the first collected signal collected by the microphone, and the first collected signal itself collected by the microphone or the first collected sound signal. This is a concept including a second sound collection signal obtained as a difference between the signal and the predicted echo signal. However, when the first sound pickup signal itself or the second sound pickup signal is a signal in the time domain, the signal is converted into a signal in the frequency domain. Furthermore, a signal that has been devised so that the cross-correlation of the multi-channel received signal changes with respect to the first or second collected sound signal (for example, a signal loaded with noise, half-wave rectification, delay variation, level variation) Or a second sound pickup signal obtained as a difference between the signal obtained by performing the above-described contrivance on the first sound pickup signal and the predicted echo signal. Good.

Claims (8)

One or more speakers and one or more microphones are arranged in a common sound field, and when an incoming signal is reproduced from the speakers, an echo canceller that cancels echoes that enter the microphone via an echo path. ,
Using the frequency domain collected signal that is a signal obtained from the first collected signal collected by the microphone and the frequency domain received signal that is a frequency domain signal obtained from the received signal, the frequency domain collected signal is used. Considering the phase and amplitude of the residual echo contained in the signal, a residual echo estimation unit for obtaining a residual echo estimate,
Canceling the residual echo estimation value from the frequency domain collected signal, including a residual echo cancellation suppression unit to suppress,
The residual echo cancellation suppression unit obtains a difference by subtracting the residual echo estimation value from the frequency domain collected signal, and increases the ratio of suppressing the residual echo from the frequency domain collected signal as the difference is smaller. Reduce the rate of echo cancellation,
Echo canceler. The echo canceller of claim 1 ,
The residual echo cancellation suppressor is
The frequency index is f, the frame index is j, the frequency domain collected signal is Y (f, j), the residual echo estimate is Y ^ ₂ (f, j), and the smallest residual echo is eliminated. The ratio is p_cancel_allotted_min, and the difference Y is obtained by subtracting the power Y ^ ₂ (f, j) ² of the residual echo estimated value from the power Y (f, j) ^{2 of the} frequency domain collected signal. If it is smaller than p_hyb_range_upper, set the sharing ratio R _mxc (f, j)

And when the difference is equal to or greater than a predetermined threshold p_hyb_range_upper, an erasure suppression distribution control unit that _sets the sharing ratio R _mxc (f, j) to 1 is included
Echo canceler. The echo canceller of claim 1 or claim 2 ,
The residual echo cancellation suppressor is
The frequency index is f, the frame index is j, the echo suppression gain is G _mxs (f, j), the frequency domain collected signal is Y (f, j), and the residual echo estimate is Y ^ ₂ ( f, j),
The ratio of canceling the residual echo is the sharing ratio R _mxc (f, j), and the transmission signal V (f, j) is

Asking,
Echo canceler. A echo canceller of 請 Motomeko 3,
The echo suppression gain G _mxs (f, j) is

Is,
Echo canceler. The echo canceller according to any one of claims 1 to 4,
The received signal is filtered by an adaptive filter to generate a predicted echo signal, and a difference between the first collected sound signal collected by the microphone and the predicted echo signal is obtained as a second collected sound signal. An echo canceler that updates a filter coefficient of the adaptive filter based on the sound signal and the received signal, and
Using the second sound collection signal in the frequency domain as the frequency domain sound collection signal,
Echo canceler. An echo canceling method in which one or more speakers and one or more microphones are arranged in a common sound field, and when an incoming signal is reproduced from the speakers, an echo that goes around the microphone via an echo path is canceled. ,
Using the frequency domain collected signal that is a signal obtained from the first collected signal collected by the microphone and the frequency domain received signal that is a frequency domain signal obtained from the received signal, the frequency domain collected signal is used. A residual echo estimation step for obtaining a residual echo estimate in consideration of the phase and amplitude of the residual echo contained in the signal;
Canceling the residual echo estimate from the frequency domain collected signal, and a residual echo cancellation suppressing step of suppressing,
The residual echo cancellation suppression step obtains a difference by subtracting the residual echo estimation value from the frequency domain collected signal, and the smaller the difference, the higher the ratio of suppressing the residual echo from the frequency domain collected signal. Reduce the rate of echo cancellation,
Echo cancellation method. The echo cancellation method of claim 6,
The received signal is filtered by an adaptive filter to generate a predicted echo signal, and a difference between the first collected sound signal collected by the microphone and the predicted echo signal is obtained as a second collected sound signal. An echo cancellation step of updating a filter coefficient of an adaptive filter based on the sound signal and the received signal, and
Using the second sound collection signal in the frequency domain as the frequency domain sound collection signal,
Echo cancellation method.   A program for causing a computer to function as the echo canceling apparatus according to any one of claims 1 to 5.

Images