JPH10303787A - Device and method for echo canceling - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reinforce proof against the disturbances, such as double talk or the like while converging speed is kept by setting the upper limit value or the lower limit value of a permission range, when the correction quantity of an impulse response estimation value deviates from the permissible range. SOLUTION: A subtracter 8 subtracts a pseudo-echo signal (r) from a transmission signal (y) and outputs the difference signal (e). A ΔH generation circuit 9 operates the correction quantity of the impulse response estimation value from the difference signal (e) and a reception signal (x) stored in an X register 5. A ΔH chip part 21 calculates the upper limit value and the lower limit value of the permission range of correction quantity, based on the number of the reception signals (x) stored in the X register 5, and a loop gain. When correction quantity deviates from the permissible range, the upper limit value or the lower limit value of the permissible range is decided as the correction quantity of the impulse response estimation value. An adder 22 corrects the impulse response estimation value by adding the correction quantity to the impulse response estimation value stored in an H-register 6.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an echo canceling apparatus and an echo canceling method for canceling an echo superimposed on a transmission signal when a part of a reception signal leaks to a transmission side.

[0002]

2. Description of the Related Art FIG.
FIG. 1 is a configuration diagram showing a conventional echo canceling device disclosed in Japanese Patent Application Publication No. 3 (1993) -207, wherein 1 is an input terminal of a received signal x (i), 2 is an output terminal of a received signal x (i), and 3 is a transmission signal y (i) input terminal, 4 is an output terminal of residual signal e (i), 5 is received signal x (i), x (i) for the past M samples.
−1),..., X (i−M + 1), and 6 is an estimated impulse response h ₀ of the echo path for the same number of samples as the received signal stored by the X register 5.
(I), h ₁ (i),..., H _M−1 (i) are stored in the H register, and 7 is the received signal x (i) stored in the X register 5 and stored in the H register 6. The convolution integrator calculates the pseudo echo signal r (i) by performing convolution integration with the estimated impulse response value h _k (i). The convolution integrator 8 converts the pseudo echo signal r (i) from the transmission signal y (i).
Is a subtractor that subtracts the above and cancels the echo.

Further, reference numeral 9 denotes a correction amount Δh _{0 of} an impulse response estimated value from a residual signal e (i) obtained as a result of subtraction by the subtracter 8 and each received signal x (i) stored by the X register 5.
(I), △ h ₁ (i),..., △ h _M-1 (i) The △ H generation circuit 10 calculates the impulse response estimated value h _k (i) stored by the H register 6 On the other hand, it is an adder for adding the correction amount Δh _k (i) of the impulse response estimated value.

Next, the operation will be described. First, at the sampling time i, the reception signal x
When (i) is input, the X register 5 stores the received signal x (i), while receiving the oldest received signal,
That is, the received signal x (i-M) M samples before (the sampling time i-M) before the received signal x (i) is discarded, and the received signals of the past M samples are held. On the other hand, in the H register 6, the impulse response estimated values h _{0 for the} same number of samples (for M samples) as the received signal stored in the X register 5 are stored.
(I), h ₁ (i),..., H _M−1 (i) are stored. In the initial stage, a zero value is usually given as an impulse response estimated value.

When the convolution integrator 7 stores the received signal in each register, the received signal x (i) stored in the X register 5 and the received signal are stored in the H register 6 as described below. By performing convolution integration with the estimated impulse response value h _k (i), a pseudo echo signal r (i) is calculated.

[0006]

(Equation 1)

When the convolution integrator 7 outputs the pseudo echo signal r (i), the subtractor 8 receives the received signal x
A part of (i) leaks to the transmission side and cancels the echo superimposed on the transmission signal y (i), so that a pseudo echo signal r (i) is obtained from the transmission signal y (i) input from the input terminal 3. ) Is subtracted, and the residual signal e (i) is output. e (i) = y (i) -r (i) (2)

Thus, when the pseudo echo signal r (i) completely matches the echo signal superimposed on the transmission signal y (i), the echo is canceled from the transmission signal y (i), but the pseudo signal is removed. If the echo signal r (i) does not match the echo signal, the impulse response estimated values h ₀ (i), h ₁ (i),..., H _{M− 1} are used to improve the accuracy of echo cancellation.
(I) needs to be modified. Therefore, the ΔH generation circuit 9
By using the learning identification method as shown below,
From the residual signal e (i) output from the subtracter 8 and each received signal x (i) stored in the X register 5, etc., the correction amount Δh ₀ (i), Δh ₁ (i) of the impulse response estimated value , ...
・, △ h _M-1 (i) is calculated.

[0009]

(Equation 2)

Here, the loop gain α in equation (3)
Is given as 0 <α ≦ 1, but it is known that given such a value, it is possible to estimate the impulse response of a stable echo path. The amount of echo cancellation and the convergence speed of the echo canceling device have a trade-off relationship with the value of the loop gain α. That is, when the loop gain α is set to a large value (a value close to 1), the convergence speed increases, but the amount of echo cancellation decreases, and
Resistance to double talk (a situation in which both parties of the telephone talk at the same time) will be weakened. Conversely, if the loop gain α is set to a small value (a value close to 0), the amount of echo cancellation can be increased and the resistance to double talk can be increased, but the convergence speed decreases.

[0011] Then, the adder 10, △ when H generation circuit 9 calculates the correction amount of the impulse response estimate △ h _k (i), the impulse response estimate is stored by the H register 6 h _k (i) The amount of correction of the impulse response estimate △
The impulse response estimation value h _k (i) is corrected by adding h _k (i) respectively. h _k (i + 1) = h _k (i) + △ h _k (i) (4) where k = 0, 1,..., M−1

Thus, according to the conventional example, the impulse response estimated value h _k (i) is corrected according to the change of the received signal x (i) and the residual signal e (i), and thus the echo canceling is performed. Although the accuracy can be improved, since the value of the loop gain α is set to a constant value, the characteristics of the echo canceling device cannot be appropriately changed according to the convergence state of the echo. Therefore, an echo canceling device that changes the value of the loop gain α according to the convergence state of the echo has been developed, and such a technology is disclosed in Japanese Patent Publication No. 2-62975.

FIG. 12 is a block diagram showing an echo canceling device disclosed in Japanese Patent Publication No. 2-62975. In the drawing, reference numeral 11 denotes a loop gain control circuit for changing a loop gain α. FIG. 13 is a configuration diagram showing details of the loop gain control circuit 11, and FIG.
12 is a squaring circuit for calculating the square of the residual signal e (i), 1
Reference numeral 3 denotes a moving average filter, an integrator, and the like, and a linear filter for outputting the signal level of the residual signal e (i).
Is a clip circuit that outputs a loop gain α corresponding to the signal level of the residual signal e (i).

Next, the operation will be described. Except for the point that the loop gain control circuit 11 is provided, it is the same as the conventional example shown in FIG. 11, so only the operation of the loop gain control circuit 11 will be described. That is, the loop gain control circuit 11
When the residual signal e (i) is output from the subtractor 8, the square value of the residual signal e (i) is passed through the linear filter 13, so that the signal level of the residual signal e (i) Ask for.

When the signal level of the residual signal e (i) is obtained, the clipping circuit 1 of the loop gain control circuit 11
4 outputs a loop gain α corresponding to the signal level of the residual signal e (i). That is, when the signal level of the residual signal e (i) is large, since the echo has not yet converged, the characteristics are changed in a direction to increase the loop gain α to increase the convergence speed. Conversely, the residual signal e (i)
When the signal level is low, the convergence of the echo has progressed, and the characteristic is changed in the direction of decreasing the loop gain α to increase the amount of echo cancellation.

The clipping circuit 14 prevents the coefficient updating operation from stopping when the loop gain α becomes too small, and makes the estimation operation of the impulse response unstable when the loop gain α becomes too large. In order to prevent this, the loop gain α is clipped as shown below. α _min ≦ α ≦ α _max (5)

[0017]

Since the conventional echo canceling apparatus is configured as described above, when the loop gain α is set to a constant value, the characteristic of the echo canceling depends on the convergence state of the echo. Cannot be changed. For example, the loop gain α is set to a large value (a value close to 1).
When the value is set to, the amount of echo cancellation at the time of occurrence of disturbance becomes small, and a problem that the resistance to double talk is weakened occurs. Conversely, if the loop gain α is set to a small value (a value close to 0), there is a problem that the convergence speed is reduced.

On the other hand, the loop gain α is changed to the residual signal e (i).
When changing according to the signal level of the echo, the characteristic of echo cancellation can be changed to some extent according to the convergence state of the echo, but when the loop gain α is changed, the impulse response estimation values for all elements of the impulse response are changed. since that would result in changing the correction amount △ h _k (i) uniformly, there may not be suitable as a correction amount of the impulse response estimates for the individual elements △ h _k (i), necessarily accurately echo There was a problem that could not be erased.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problem, and has a practical convergence speed.
It is an object of the present invention to provide an echo canceling apparatus and an echo canceling method that have high resistance to disturbances such as double talk.

[0020]

According to the first aspect of the present invention, an echo canceling apparatus calculates a correction amount of an impulse response estimated value from a subtraction result of the echo canceling means and a received signal stored by the signal storing means. If the amount of correction of the impulse response estimation value deviates from the allowable range set in accordance with the number of received signals stored by the signal storage means and the loop gain, the upper limit value or the lower limit value of the permissible range is impulsed. This is the correction amount of the response estimation value.

According to a second aspect of the present invention, there is provided an echo canceling apparatus for calculating a ratio between the power of a received signal and the power of a transmitted signal, and a loss estimating means for estimating a loss of an echo path based on the calculation result. And an allowable range updating means for updating the allowable range in accordance with the echo path loss estimated by the loss estimating means.

According to a third aspect of the present invention, there is provided an echo canceling apparatus for calculating a ratio between a power of a received signal and a power of a residual signal, and an echo amount estimating means for estimating an echo canceling amount based on the calculation result. And an allowable range updating unit for updating the allowable range in accordance with the echo cancellation amount estimated by the cancellation amount estimation unit.

According to a fourth aspect of the present invention, there is provided an echo canceling apparatus comprising: a counting means for counting the number of times the correction amount of the estimated impulse response value deviates from the allowable range; and an allowable range in accordance with the number counted by the counting means. And an allowable range updating means for updating the value.

According to a fifth aspect of the present invention, there is provided an echo canceling apparatus comprising: a call monitoring means for determining whether or not a transmission line is in use; An integrating means for integrating the above time and an allowable range updating means for updating the allowable range according to the time integrated by the integrating means are provided.

According to a sixth aspect of the present invention, there is provided an echo canceling apparatus for determining whether a transmission line is in use or not, and permitting updating of an allowable range based on power of a received signal and a transmission signal. Update determination means for determining whether or not to perform the update, integration means for integrating the time during which the update determination means determines that the update is permitted while the transmission line is in use, and integration performed by the integration means. And an allowable range updating means for updating the allowable range in accordance with the time.

According to a seventh aspect of the present invention, there is provided an echo canceling apparatus, comprising: a call monitoring means for determining whether or not a transmission line is in use; While updating the allowable range according to the counting means and the number of times counted by the counting means,
When the transmission line is switched from the non-use state to the use state, a permissible range updating unit for expanding the permissible range is provided.

In the echo canceling apparatus according to the present invention, the upper limit value and the lower limit value of the allowable range of the impulse response are multiplied by weighting values that exponentially attenuate according to the sample value of the impulse response. Things.

According to a ninth aspect of the present invention, the correction amount of the impulse response estimation value for each element of the impulse response is determined every several samples.

According to a tenth aspect of the present invention, in the echo canceling method, a correction amount of the impulse response estimation value is calculated from a subtraction result obtained by subtracting the pseudo echo signal from the transmission signal and the reception signal, and the impulse response estimation value is calculated. If the correction amount deviates from the allowable range set in accordance with the number of received signals and the loop gain, the upper limit or lower limit of the allowable range is used as the correction amount of the impulse response estimated value. is there.

The echo canceling method according to the present invention calculates the ratio between the power of the received signal and the power of the transmitted signal, estimates the loss of the echo path based on the calculation result, and calculates the loss of the echo path. The allowable range is updated according to the loss.

According to a twelfth aspect of the present invention, in the echo canceling method, the ratio between the power of the received signal and the power of the residual signal is calculated, and the amount of echo cancellation is estimated based on the calculation result. The allowable range is updated according to.

According to a thirteenth aspect of the present invention, in the echo canceling method, the number of times the correction amount of the impulse response estimation value deviates from the allowable range is counted, and the allowable range is updated according to the number. .

In the echo canceling method according to the fourteenth aspect of the present invention, it is determined whether or not the transmission line is in use, and when the transmission line is in use, the power of the received signal exceeds a certain level. The time is integrated, and the allowable range is updated according to the time.

According to a fifteenth aspect of the present invention, in the echo canceling method, it is determined whether or not the transmission line is in use and whether or not to permit the updating of the allowable range based on the power of the reception signal and the transmission signal. Is determined, the time during which it is determined that the update of the allowable range is permitted while the transmission line is in use is integrated, and the allowable range is updated according to the time. .

The echo canceling method according to the sixteenth aspect of the present invention determines whether or not the transmission line is in use and counts the number of times the correction amount of the estimated impulse response value deviates from an allowable range. While the allowable range is updated according to the number of times, when the transmission line switches from the non-use state to the use state, the allowable range is expanded.

In the echo canceling method according to the present invention, the upper limit value and the lower limit value of the permissible range of the impulse response are respectively multiplied by a weight value that attenuates exponentially according to the sample value of the impulse response. Things.

In the echo canceling method according to the present invention, the correction amount of the impulse response estimation value for each element of the impulse response is determined every several samples.

[0038]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below. Embodiment 1 FIG. FIG. 1 is a block diagram showing an echo canceling apparatus according to Embodiment 1 of the present invention.
1 is an input terminal of the reception signal x (i), 2 is a reception signal x
(I) an output terminal, 3 is an input terminal of a transmission signal y (i),
4 is an output terminal of the residual signal e (i), 5 is a received signal x (i), x (i-1),.
-M + 1) (signal storage means), 6
Is the estimated impulse response value h ₀ of the echo path for the same number of samples as the received signal stored by the X register 5.
(I), an H register (estimated value storage means) for storing h ₁ (i),..., H _M-1 (i), 7 is a received signal x (i) and the like stored by the X register 5 The convolution integrator (echo cancellation means) for calculating the pseudo echo signal r (i) by performing convolution integration with the impulse response estimation value h _k (i) stored by the H register 6, and the transmission signal y ( A subtractor (echo cancellation means) for subtracting the pseudo echo signal r (i) from i) and canceling the echo.

Reference numeral 9 denotes a correction amount Δh ₀ of the impulse response estimated value from the residual signal e (i) as a result of the subtraction by the subtracter 8 and each received signal x (i) stored by the X register 5.
(I), △ h ₁ (i),..., Mh _M-1 (i) △ H generation circuit (correction amount determining means), 21 is an impulse response calculated by △ H generation circuit 9 If the correction amount Δh _k (i) of the estimated value deviates from the allowable range set according to the number M of received signals stored by the X register 5 and the loop gain α, the upper limit of the allowable range Value @h
_The △ H clip portion (correction amount determining means), where _max (i) or the lower limit を h _min (i) is the correction amount △ h _k (i) of the impulse response estimated value, is determined by the △ H clip portion 21. The correction amount Δh _k (i) of the estimated impulse response
Impulse response estimate h _k stored by register 6
An adder (estimated value correcting means) for adding to (i). FIG. 2 is a flowchart showing an echo canceling method applied by the echo canceling apparatus according to the first embodiment of the present invention.

Next, the operation will be described. First, at the sampling time i, the reception signal x
When (i) is input, the X register 5 stores the received signal x (i), while receiving the oldest received signal,
That is, the received signal x (i-M) that is M samples before the received signal x (i) (sampling time i-M) is discarded, and the received signals for the past M samples are held (step ST).
1). On the other hand, in the H register 6, impulse response estimated values h ₀ (i), h ₁ (i),..., H for the same number of samples (for M samples) as the received signal stored in the X register 5 are stored.
_M-1 (i) is stored, but in the initial stage, a zero value is usually given as an impulse response estimated value.

When the received signal and the like are stored in each register, the convolution integrator 7 stores the received signal x (i) and the like stored in the X register 5 and the H register 6 as described below. The pseudo echo signal r (i) is calculated by performing convolution integration with the estimated impulse response value h _k (i) (step ST2).

[0042]

(Equation 3)

When the convolution integrator 7 outputs the pseudo echo signal r (i), the subtractor 8
A part of (i) leaks to the transmission side and cancels the echo superimposed on the transmission signal y (i), so that a pseudo echo signal r (i) is obtained from the transmission signal y (i) input from the input terminal 3. ) Is subtracted, and the residual signal e (i) is output (step S).
T3). e (i) = y (i) -r (i) (7)

Thus, when the pseudo echo signal r (i) completely matches the echo signal superimposed on the transmission signal y (i), the echo is eliminated from the transmission signal y (i), but the pseudo echo signal is removed. If the echo signal r (i) does not match the echo signal, the impulse response estimated values h ₀ (i), h ₁ (i),..., H _{M− 1} are used to improve the accuracy of echo cancellation.
(I) needs to be modified. Therefore, the ΔH generation circuit 9
By using the learning identification method as shown below,
From the residual signal e (i) output from the subtracter 8 and each received signal x (i) stored in the X register 5, etc., the correction amount Δh ₀ (i), Δh ₁ (i) of the impulse response estimated value , ...
, △ h _M-1 (i) is calculated (step ST4).

[0045]

(Equation 4)

Here, the loop gain α in the equation (8)
Is given as 0 <α ≦ 1, but it is known that given such a value, it is possible to estimate the impulse response of a stable echo path. The amount of echo cancellation and the convergence speed of the echo canceling device have a trade-off relationship with the value of the loop gain α. That is, when the loop gain α is set to a large value (a value close to 1), the convergence speed increases, but the amount of echo cancellation decreases, and
The resistance to double talk will be weakened. Conversely, if the loop gain α is set to a small value (a value close to 0), the amount of echo cancellation can be increased and the resistance to double talk can be increased, but the convergence speed decreases.

Then, the △ H clip unit 21 outputs the correction amount △ h _k (i) of the impulse response estimated value from the △ H generation circuit 9.
Is output, the allowable range of the correction amount, that is, the upper limit value Δh of the allowable range is determined based on the number M of the received signals stored in the X register 5 and the loop gain α as shown below.
Calculate _max (i) and lower limit value Δh _min (i). Δh _max (i) = 0.5 · K · (α / M) (9) Δh _min (i) = − 0.5 · K · (α / M) (10) Where K is a positive constant

Then, the ΔH clip unit 21 corrects the impulse response estimated value Δh _k (i) as shown below.
Is determined to be within the allowable range, and if the value deviates from the allowable range, the upper limit value Δh _max (i) or the lower limit value Δh _min (i) of the allowable range is corrected by the impulse response estimated value. △ h
_k (i) is determined (step ST5). Δh _min (i) <Δh _k (i) <Δh _max (i) (11) In the case of Δh _k (i) <Δh _min (i) Δh _k (i) ← △ h _min (i) · △ h _k (i)> △ h _max (i) Δh _k (i) ← △ h _max (i)

Here, the correction amount of the impulse response estimated value △
It is determined whether or not h _k (i) is within an allowable range. When the value deviates from the allowable range, the upper limit value Δh _max (i) or the lower limit value Δh _min (i) of the allowable range is set to an estimated impulse response value. The reason for determining as the correction amount Δh _k (i) is to increase the resistance to disturbances such as double talk while ensuring a practical convergence speed. The average value of the absolute value of the correction amount Δh _k (i) of the impulse response estimated value, that is, the expected value E (| △ h _k (i) |) is expressed as shown below, and is expressed by Expression (9) and Compared with equation (10), the term of (α / M) is common and the term of (Pe / Px) is different (0.5 K in equations (9) and (10)). E (| △ h _k (i) |) = (Pe / Px) · (α / M) (12) where Px is the power of the received signal x (i) Pe is the power of the echo signal

However, (Pe / Px) is the product of the echo path loss and the echo cancellation amount. For example, in the case of echo cancellation of a telephone line, the worst echo path loss is 0.5.
(6 dB in decibel value), and the echo cancellation amount is about 1 (0 dB in decibel value) unless converged, so the maximum value of (Pe / Px) is about 0.5 × 1 = 0.5 Becomes Therefore, although the term of (Pe / Px) can be replaced by 0.5, when the term of (Pe / Px) is replaced by 0.5, K in equations (9) and (10) can be replaced. As long as is set to 1, the values match the expressions (9) and (10), so that the upper limit value {h _max (i) and the lower limit value} of the allowable range
When h _min (i) is calculated from Expressions (9) and (10), the correction amount Δh _k (i) of the impulse response estimated value is approximately _obtained.
Can be made to match the expected value E (△ h _k (i)). As a result, it is possible to increase the resistance to disturbances such as double talk while ensuring a practical convergence speed. It is. However, the correction amount Δh _k (i) of the estimated impulse response fluctuates instantaneously, and the expected value E
(△ h _k (i)), the upper limit value Δh
_max (i) and the lower limit Δh _min (i) are set to the expected value E (Δh
_k (i)), and in practice, the value of K is set to about 2 or 3.

Then, the adder 22 calculates the ΔH clip unit 2
When the correction amount Δh _k (i) of the impulse response estimation value is output from 1, H is set to prepare for the next echo cancellation.
Impulse response estimate h _k stored by register 6
Correction amount of the impulse response estimation value Δh _k for (i)
The estimated impulse response value h _k (i) is corrected by adding (i) (step ST6). h _k (i + 1) = h _k (i) + △ h _k (i) (13) where k = 0, 1,..., M−1

FIG. 3 shows a computer simulation of the convergence characteristics of the echo canceling apparatus. Curve A shows a curve obtained when the loop gain α is set to 0.2 in the conventional echo canceling apparatus shown in FIG. B shows the case where the loop gain α is set to 0.1 in the conventional echo cancellation apparatus shown in FIG. 11, and the curve C shows the loop gain α of 0.2 and K is 2 in the echo cancellation apparatus of the first embodiment. Shows the amount of echo cancellation when set. However, when the time is 0 second, all the stored values of the H register 6 are set to 0, and a disturbance signal of the same level as the received signal x (i) is applied to the transmitting side for 0.1 second after the lapse of 2 seconds, and the double talk state is obtained. I have to. Further, the characteristic of the echo path is changed at the time when 4 seconds have elapsed.

As is clear from FIG. 3, the curve A has a high convergence speed, but the amount of echo cancellation at the time of occurrence of double talk is greatly deteriorated. Curve B has a smaller degradation of the amount of echo cancellation when double talk occurs than curve A, but has a slower convergence speed. On the other hand, the curve C has a slightly lower convergence speed than the curve A, but is faster than the curve B. The curve C has a smaller amount of echo cancellation when double talk occurs than the curves A and B.

As is clear from the above, the first embodiment
According to the residual signal e (i) and the received signal x correction amount of the impulse response estimates from (i) △ calculates the h _k (i), the correction amount of the impulse response estimate △ h _k (i) Deviates from the upper limit △ h _max (i) or the lower limit △ h _min (i) of the allowable range, the upper limit △ h _max (i) or the lower limit △ h _min (i) of the allowable range is impulsed. Since the correction amount of the response estimation value is set to Δh _k (i), there is an effect that the immunity to disturbance such as double talk can be enhanced while a practical convergence speed is secured.

Embodiment 2 FIG. 4 is a configuration diagram showing an echo canceling apparatus according to Embodiment 2 of the present invention.
In the figure, the same reference numerals as those in FIG. 23 is a received signal x (i)
Px (i) and the power Py (i) of the transmission signal y (i)
, And an echo path loss estimator (loss estimating means) 24 for estimating the estimated loss value L (i) of the echo path based on the calculation result. Range setting unit (tolerance range updating unit) that updates the tolerance range according to the estimated loss value L (i)
It is.

Next, the operation will be described. Except that the echo path loss estimating unit 23 and the allowable range setting unit 24 are provided, the configuration is the same as that of the first embodiment. Therefore, only the operation of the echo path loss estimating unit 23 and the allowable range setting unit 24 will be described. That is, the echo path loss estimating unit 23 first determines the power Px (i) of the received signal x (i) (the received signal x (i)
After calculating the short-time average level of (i) and the power Py (i) of the transmission signal y (i) (short-time average level of the transmission signal y (i)), the power Px (i) for the power Py (i) ) Is calculated, and a preset estimated loss value L (i) of the echo path (in the initial stage, usually,
0.5 is set) and Px (i) / Py (i) are compared.

Then, the echo path loss estimating section 23 calculates that the estimated loss value L (i) of the echo path is Px (i) / Py.
If smaller than (i), the estimated loss value L (i) of the echo path is increased by a small amount in order to increase the convergence speed by expanding the allowable range. L (i + 1) = (1 + δ) · L (i) (14) where δ is a positive constant

Conversely, the estimated loss value of the echo path L (i)
Is larger than Px (i) / Py (i), the estimated loss value L (i) of the echo path is reduced by a small amount in order to reduce the allowable range and increase the resistance to double talk. L (i + 1) = (1−δ) · L (i) (15)

However, the received signal x (i) is silent, that is,
When the power Px (i) of the received signal x (i) is equal to or less than a certain value, the estimated loss value L (i) of the echo path is obtained as described above.
Is determined, the estimation accuracy deteriorates. In this case, the previous value is used as it is without updating the loss estimated value L (i). Even if the estimated loss value L (i) of the echo path is estimated at the time of occurrence of double talk, the estimation accuracy is degraded, but generally speaking, both parties talk alternately by telephone. The time rate of occurrence of double talk is lower than the time during which the received signal x (i) is a sound,
In reality, the estimation accuracy does not deteriorate.

Then, the allowable range setting unit 24 sends the estimated loss value L (i) of the echo path from the echo path loss estimation unit 23.
Is output, as shown below, the upper limit value of the allowable range △ h _max according to the estimated loss value L (i) of the echo path.
(I) and the lower limit value Δh _min (i) are updated. Δh _max (i) = L (i) · K · (α / M) (16) Δh _min (i) = − L (i) · K · (α / M) (17) )

Here, when the expressions (16) and (17) are compared with the expressions (9) and (10), the expression (9) or the like in which “0.5” is replaced by L (i) is obtained. It turns out that it is. That is, in the first embodiment, “0.5” is used as the maximum possible value of the echo path loss, but in the second embodiment, a value close to the actual loss of the echo path (the loss estimated value L Using (i)), the upper limit value of the allowable range Δh
_max (i) and the lower limit value Δh _min (i) are updated.

As is apparent from the above, this embodiment 2
According to the above, the ratio between the power Px (i) of the received signal x (i) and the power Py (i) of the transmitted signal y (i) is calculated,
Based on the calculation result, the estimated loss value L of the echo path
(I) is estimated, and a loss estimated value L (i) of the echo path is estimated.
Upper limit {h _max (i) and lower limit _} of the allowable range according to
Since h _min (i) is updated, even if disturbance such as double talk occurs when the loss of the echo path is large, it is possible to suppress the deterioration of the echo cancellation amount.

Embodiment 3 FIG. 5 is a configuration diagram showing an echo canceling apparatus according to Embodiment 3 of the present invention.
In the figure, the same reference numerals as those in FIG. 25 is a received signal x (i)
Px (i) and the power Pe (i) of the residual signal e (i)
Is calculated, and the echo cancellation amount C is calculated based on the calculation result.
The echo cancellation residual power estimating unit (cancellation amount estimating unit) 26 for estimating (i) updates the allowable range in accordance with the echo cancellation amount C (i) estimated by the echo cancellation residual power estimating unit 25. A range setting unit (permissible range updating unit).

Next, the operation will be described. Since the present embodiment is the same as the first embodiment except that the echo cancellation residual power estimation unit 25 and the allowable range setting unit 26 are provided, only the operations of the echo cancellation residual power estimation unit 25 and the allowable range setting unit 26 are performed. explain. That is, the echo cancellation residual power estimation unit 2
5 is the power Px (i) of the received signal x (i) (the short-time average level of the received signal x (i)) and the residual signal e (i)
After calculating the power Pe (short-time average level of the residual signal e (i)), the ratio Px (i) / Pe (i) of the power Px (i) to the power Pe (i) is calculated and set in advance. The amount of echo cancellation C (i) (normally set to a zero value in the initial stage) is compared with Px (i) / Pe (i).

Then, the echo canceling residual power estimating unit 25
Means that the echo cancellation amount C (i) is Px (i) / Pe (i)
If smaller, the echo cancellation amount C (i) is increased by a small amount in order to increase the allowable range and increase the convergence speed. C (i + 1) = (1 + δ) · C (i) (18)

On the contrary, when the echo cancellation amount C (i) is Px
If (i) / Pe (i), the echo canceling amount C (i) is reduced by a small amount in order to reduce the allowable range and increase the resistance to double talk. C (i + 1) = (1−δ) · C (i) (19)

However, the received signal x (i) is silent, that is,
If the echo cancellation amount C (i) is estimated as described above when the power Px (i) of the received signal x (i) is equal to or less than a certain value, the estimation accuracy deteriorates. Without updating C (i), the previous value is used as it is. When double talk occurs, the amount of echo cancellation C
Even if (i) is estimated, the estimation accuracy is degraded, but generally speaking, both telephones usually talk alternately. Therefore, during the time when the received signal x (i) is a sound, On the other hand, the time rate of occurrence of double talk is low, and the estimation accuracy does not actually deteriorate.

When the echo canceling amount C (i) is output from the echo canceling residual power estimating unit 25, the allowable range setting unit 26 calculates the echo canceling amount C (i) as described below.
The upper limit value Δh _max (i) and the lower limit value Δh _min (i) of the allowable range are updated according to (i). Δh _max (i) = C (i) · K · (α / M) (20) Δh _min (i) = − C (i) · K · (α / M) (21) )

Here, when the expressions (20) and (21) are compared with the expressions (9) and (10), the expression (9) or the like in which “0.5” is replaced by C (i) is obtained. It turns out that it is. That is, in the first embodiment, “0.5” is used as the worst value of the total echo cancellation amount obtained by combining the echo path and the echo canceling device.
Is the product of the conceivable maximum value of the echo path loss and the amount of echo cancellation when no echo cancellation is performed by the echo canceller.) In the third embodiment, a value close to the actual amount of echo cancellation Using the (echo cancellation amount C (i)), the upper limit value Δh _max (i) and the lower limit value Δh of the allowable range
_min (i) is updated.

As is apparent from the above, the third embodiment
According to, the ratio between the power Px (i) of the received signal x (i) and the power Pe (i) of the residual signal e (i) is calculated,
The echo canceling amount C (i) is estimated based on the calculation result, and the upper limit value Δh _max (i) and the lower limit value Δh _min (i) of the allowable range are updated according to the echo canceling amount C (i). Since the convergence of the echo canceling device progresses, the allowable range can be reduced, and even if a disturbance such as double talk occurs,
This has the effect of suppressing the deterioration of the echo cancellation amount.

Embodiment 4 FIG. 6 is a configuration diagram showing an echo canceling apparatus according to Embodiment 4 of the present invention.
In the figure, the same reference numerals as those in FIG. 27 is a clip occurrence counting unit (counting means) for counting the number of times the correction amount Δh _k (i) of the impulse response estimation value deviates from the allowable range, and 28 is allowable according to the number counted by the clip occurrence counting unit 27. An allowable range setting unit (an allowable range updating unit) that updates the range.

Next, the operation will be described. Except for providing the clip occurrence counting unit 27 and the allowable range setting unit 28,
Since the operation is the same as in the first embodiment, only the operations of the clip occurrence counting unit 27 and the allowable range setting unit 28 will be described. That is, the clip occurrence counting unit 27 determines that the correction amount _{ h _k (i) of the estimated impulse response value is the upper limit value of the allowable range _}.
The number of times exceeding h _max (i) or the lower limit value Δh _min (i) is counted. Specifically, the correction amounts △ h ₀ (i), △ h ₁ (i),... Of the impulse response estimation values for M samples
, △ h _M-1 (i) is counted how many samples are clipped by the upper limit value △ h _max (i) or the lower limit value △ h _min (i) of the allowable range.

When the count value is output from the clip occurrence counting section 27, the allowable range setting section 28 considers that the convergence of the echo canceling device has advanced if the count value is "0". Therefore, the permissible range is reduced in order to further suppress the deterioration of the echo cancellation amount when disturbance such as double talk occurs. Δh _max (i + 1) = (1−δ) · Δh _max (i) (22) Δh _min (i + 1) = (1−δ) · Δh _min (i) (23)

Conversely, if the count value output from the clip occurrence counting section 27 is other than "0", it is considered that the convergence of the echo canceling device has not progressed. To enlarge. Δh _max (i + 1) = (1 + δ) · Δh _max (i) (24) Δh _min (i + 1) = (1 + δ) · Δh _min (i) (25)

However, in order to prevent the upper limit value Δh _max (i) and the lower limit value Δh _min (i) of the allowable range from unnecessarily increasing or decreasing, the allowable range setting unit 28 sets the allowable range. Clipping is performed so that the upper limit value Δh _max (i) and the lower limit value Δh _min (i) fall within a certain range. t1 ≦ Δh _max (i) ≦ t2 (26) t1 ≦ −Δh _min (i) ≦ t2 (27) The value of t1 is, for example, a fixed-point operation performed by an echo canceling device. If implemented using
The minimum value that can be realized with the calculation accuracy is selected. Also, t
As the value of 2, a value as shown in Expressions (9) and (10) is selected.

As is apparent from the above, the fourth embodiment
According to the method, the number of times the correction amount of the impulse response estimation value deviates from the allowable range is counted, and the allowable range is updated according to the number. Therefore, as the convergence of the echo canceling device progresses, disturbance such as double talk is caused. There is an effect that the deterioration of the echo cancellation amount at the time of occurrence can be further suppressed.

Embodiment 5 FIG. 7 is a configuration diagram showing an echo canceling apparatus according to Embodiment 5 of the present invention.
In the figure, the same reference numerals as those in FIG. Reference numeral 29 denotes a call connection state monitoring unit (call monitoring means) that monitors signaling transmission paths on the transmission side and the reception side and determines whether a channel (transmission line) to which the echo canceling device is connected is in use. ,
Numeral 30 denotes a received signal power measuring unit (integrating means) for measuring the power Px (i) of the received signal x (i). Numeral 31 denotes that the channel to which the echo canceling device is connected is not in use.
The time count value tim (i) is reset, the channel to which the echo canceling device is connected is in use,
When the power Px (i) of the received signal x (i) is equal to or higher than a certain level, a timer unit (integrating means) for increasing the time count value tim (i) is provided.
An allowable range setting unit (an allowable range updating unit) that updates the upper limit value Δh _max (i) and the lower limit value Δh _min (i) of the allowable range according to (i).

Next, the operation will be described. Except that a call connection state monitoring unit 29, a received signal power measurement unit 30, a timer unit 31, and an allowable range setting unit 32 are provided, the configuration is the same as that of the first embodiment, so that the call connection state monitoring unit 29, the reception signal Power measurement unit 30, timer unit 31, and allowable range setting unit 32
Only the operation will be described.

First, the call connection state monitoring unit 29 monitors the signaling transmission paths on the transmitting side and the receiving side to determine whether or not the channel to which the echo canceling device is connected is in use. On the other hand, received signal power measuring section 30 measures power Px (i) of received signal x (i).

In the case of a telephone line echo canceling device, the timer section 31 and the allowable range setting section 32 change the convergence speed in this case because the impulse response of the echo path changes at the moment when a channel is newly connected. In consideration of the need to increase the speed, the following processing is executed. That is, the timer unit 31 does not need to perform echo cancellation at present unless the channel to which the echo canceling device is connected is in use. Therefore, the timer unit 31 prepares for a newly connected channel and prepares the time count value tim (i ) reset (in this case, as shown below, the upper limit of the allowable range △ h _max (i) the absolute value of and the lower limit value △ h _{mi n} (i) is the maximum value).

Conversely, the channel to which the echo canceling device is connected is in use and the received signal x (i)
When the power Px (i) is equal to or higher than a certain level, the time count value tim (i) is increased as the convergence of the echo canceling device progresses in order to further suppress the deterioration of the amount of echo cancellation when disturbance such as double talk occurs. Let it.

When the time count value tim (i) is output from the timer unit 31, the allowable range setting unit 32
As shown below, the upper limit value Δh _max (i) and the lower limit value Δh of the allowable range are set according to the time count value tim (i).
Update _min (i). _{△ h max (i) = T1} / (tim (i) + T2) ··· (28) △ h min (i) = - T1 / (tim (i) + T2) However ··· (29), T1, T2 Is a positive constant

As apparent from the above, the fifth embodiment
According to the above, it is determined whether or not the channel to which the echo canceling device is connected is in use, and when the channel is in use, the power Px (i) of the received signal x (i) is at a certain level. The above time count value ti
m (i) is integrated, and the upper limit value Δh _max (i) and the lower limit value Δh of the allowable range are determined according to the time count value tim (i).
_{Since min} (i) is updated, it is possible to prevent the echo cancellation amount from being deteriorated when disturbance such as double talk occurs, particularly after the echo cancellation device has sufficiently converged.

Embodiment 6 FIG. FIG. 8 is a configuration diagram showing an echo canceling apparatus according to Embodiment 6 of the present invention.
In the figure, the same reference numerals as those in FIG. 7 indicate the same or corresponding parts, and thus description thereof will be omitted. 33 is a received signal x
(I) a coefficient update control unit (update determination means) for determining whether to permit updating of the allowable range based on the power of the transmission signal y (i) and the residual signal e (i); If the channel to which the device is connected is not in use, the time count value tim (i) is reset, the channel to which the echo cancellation device is connected is in use, and the coefficient update control unit 33 updates the channel. A timer unit (integrating means) that increases the time count value tim (i) when it is determined that permission is given.

Next, the operation will be described. Since the configuration other than the coefficient update control unit 33 and the timer unit 34 is the same as that of the fifth embodiment, only the operations of the coefficient update control unit 33 and the timer unit 34 will be mainly described. That is, the coefficient update control unit 33 receives the reception signal x (i), the transmission signal y (i), and the residual signal e (i), and determines whether the reception signal x (i) is in a sound / non-speech state and in a double talk state. Detect state. Then, when the received signal x (i) is in a sound state and not in a double talk state, a signal indicating that coefficient updating is permitted is output to the timer unit 34. In other cases, a signal indicating that the coefficient update is prohibited is output to the timer unit 34.

In the case of an echo canceling device for a telephone line, the timer unit 34 and the allowable range setting unit 32 change the impulse response of the echo path at the moment when a channel is newly connected. In consideration of the need to increase the speed, the following processing is executed. That is, if the channel to which the echo canceling device is connected is not in use, the timer unit 34 does not need to perform echo canceling at present, so that the time count value tim (i ) reset (in this case, as shown below, the upper limit of the allowable range △ h _max (i) the absolute value of and the lower limit value △ h _{mi n} (i) is the maximum value).

Conversely, the channel to which the echo canceling device is connected is in use, and the coefficient update control unit 3
3 permits the update, the time count value tim is set so as to further suppress the deterioration of the echo cancellation amount due to the occurrence of disturbance such as double talk as the convergence of the echo cancellation device progresses.
(I) is increased.

When the time count value tim (i) is output from the timer unit 34, the allowable range setting unit 32
As shown below, the upper limit value Δh _max (i) and the lower limit value Δh of the allowable range are set according to the time count value tim (i).
Update _min (i). _{△ h max (i) = T1} / (tim (i) + T2) ··· (30) △ h min (i) = - T1 / (tim (i) + T2) ··· (31)

As is apparent from the above, the sixth embodiment
According to the above, it is determined whether or not the channel to which the echo canceling device is connected is in use, and the power of the received signal x (i), the transmitted signal y (i), and the residual signal e (i) is reduced. A count value tim that determines whether to permit the updating of the allowable range based on the count value, and determines that the updating of the allowable range is permitted when the channel is in use.
(I) is integrated, and the upper limit value Δh _max (i) and the lower limit value Δh of the allowable range are calculated according to the time count value tim (i).
_{Since min} (i) is updated, it is possible to prevent the echo cancellation amount from being deteriorated when disturbance such as double talk occurs, particularly after the echo cancellation device has sufficiently converged.

Embodiment 7 FIG. 9 is a configuration diagram showing an echo canceling apparatus according to Embodiment 7 of the present invention.
In the figure, the same reference numerals as those in FIG. 35 updates the upper limit value Δh _max (i) and the lower limit value Δh _min (i) of the permissible range in accordance with the number of times counted by the clip occurrence counting unit 27, while the channel to which the echo canceling device is connected is updated. An allowable range setting unit (allowable range updating unit) that expands the allowable range when switching from non-use to use.

Next, the operation will be described. In the fourth embodiment, the allowable range setting unit 28 always sets the upper limit value Δh _max (i) and the lower limit value Δh _min (i) of the allowable range according to the number of times counted by the clip occurrence counting unit 27. Although the updating process has been described, the following process may be added.

That is, the allowable range setting section 35 normally sets
Similarly to the above-described fourth embodiment, the upper limit value of the allowable range Δh _max according to the number of times counted by the clip occurrence counting unit 27.
(I) and the lower limit value Δh _min (i) are updated, but when the call connection state monitoring unit 29 indicates that the channel to which the echo cancellation device is connected has been switched from non-use to use. The absolute values of the upper limit value Δh _max (i) and the lower limit value Δh _min (i) of the allowable range may be increased.

Thus, according to the seventh embodiment,
The upper limit of the allowable range at the moment when the channel is newly connected.
Since the absolute values of h _max (i) and the lower limit △ h _min (i) temporarily increase, when the echo canceling device is not converged, that is, when the impulse response of the echo path changes, the convergence speed The effect is that the allowable range can be updated so that the speed becomes faster.

Embodiment 8 FIG. FIG. 10 is a block diagram showing an echo canceling apparatus according to Embodiment 8 of the present invention. In the figure, the same reference numerals as those in FIG. 4 denote the same or corresponding parts, and a description thereof will be omitted. Reference numeral 36 denotes an exponential weighting value multiplying unit (allowable range updating means) for multiplying the upper limit value and the lower limit value of the allowable range of the impulse response by a weight value that exponentially attenuates according to the sample value k of the impulse response.

Next, the operation will be described. In the second embodiment, the ΔH clip unit 21 includes the allowable range setting unit 24.
, The correction amount Δh _k (i) of the impulse response estimation value is determined based on the upper limit value Δh _max (i) and the lower limit value Δh _min (i) of the allowable range set by Upper limit value {h _kmax (i) and lower limit value _} of the allowable range multiplied by the weight value by the shape weight value multiplication unit 36.
h _kmin (i) modification of the impulse response estimate based on △ h _k (i) may be determined.

That is, the exponential weighting value multiplying unit 36 determines that the amplitude of the impulse response of the acoustic echo path between the microphone and the speaker is generally attenuated exponentially. Upper limit value {h _max (i) and lower limit value} of the allowable range set by the setting unit 24
h _min (i) is multiplied by a weight value that exponentially attenuates according to the sample value k of the impulse response. _{Δh kmax} (i) = A · _{Δh max} (i) · e− ^{B · k} (32) _{Δh kmin} (i) = − A · _{Δh min} (i) · e− ^{B · k} · (33) where k = 0, 1,..., M-1A and B are positive constants

As is clear from the above, this embodiment 8
According to the above, since the upper limit value and the lower limit value of the allowable range of the impulse response are each multiplied by a weight value that exponentially decays according to the sample value of the impulse response,
In particular, there is an effect that the convergence speed when canceling the acoustic echo between the microphone and the speaker can be increased.

Embodiment 9 FIG. In the eighth embodiment, the echo canceling apparatus shown in FIG.
6, the exponential weighting value multiplying unit 36 may be applied to the echo canceling device in FIG. 5 to increase the convergence speed when canceling the acoustic echo between the microphone and the speaker. The effect that can be achieved.

Embodiment 10 FIG. In Embodiment 8 described above,
Although the exponential weighting value multiplying unit 36 is applied to the echo canceling device of FIG. 4, the exponential weighting value multiplying unit 36 may be applied to the echo canceling device of FIG. This has the effect of increasing the convergence speed when canceling the acoustic echo.

Embodiment 11 FIG. In the first to tenth embodiments, the ΔH generation circuit 9 outputs the sample time i
For each impulse response estimate 応 答 h ₀
(I), △ h ₁ (i),..., △ h _M−1 (i) are calculated, but the correction amount of the impulse response estimation value for each element of the impulse response is changed every few samples. Calculation (block processing) may be performed.

That is, the △ H generation circuit 9 sets the block size to b and sets k = k at the sample time i.
.., L + J × b, k for every b samples, the correction amount of the impulse response estimated value △ h ₀ (i), △ h ₁ (i),.・ ・ 、
Δh _M-1 (i) is calculated.

[0102]

(Equation 5)

Here, L is a b-ary counter that counts up by one at each sample time i, and J is the maximum integer less than M / b. Note that the ΔH clip unit 21 calculates the correction amount Δh _k of the impulse response estimated value as described below.
Clip (i). Δh _min (i) <Δh _k (i) <Δh _max (i) (35) where k = L, L + b,..., L + J × b

As is clear from the above, the first embodiment
According to No. 1, since the correction amount Δh _k (i) of the impulse response estimation value for each element of the impulse response is determined every several samples, the impulse response estimation value h _k (i) at a certain sample k is It will be updated only once for b samples, and at most △ h after b samples
_max (i) (without block processing, a maximum of b × Δh _max (i) after b samples)
Will only increase). Therefore, performing the block process on the impulse response estimation value is equivalent to reducing the allowable range, and has an effect of increasing the resistance to disturbances such as double talk.

[0105]

As described above, according to the first aspect of the present invention, the correction amount of the impulse response estimation value is calculated from the subtraction result of the echo canceling means and the received signal stored by the signal storage means, and the impulse response is calculated. If the correction amount of the response estimation value deviates from the allowable range set according to the number of received signals stored by the signal storage means and the loop gain, the upper limit or the lower limit of the allowable range is estimated by the impulse response estimation. Since the configuration is such that the correction value is used, there is an effect that the resistance to disturbance such as double talk can be enhanced while ensuring a practical convergence speed.

According to the second aspect of the present invention, the loss estimating means for calculating the ratio of the power of the reception signal to the power of the transmission signal and estimating the loss of the echo path based on the calculation result, and the loss estimating means And the allowable range updating means for updating the allowable range in accordance with the estimated loss of the echo path, so that even when disturbance such as double talk occurs when the loss of the echo path is large, There is an effect that deterioration of the echo cancellation amount can be suppressed.

According to the third aspect of the present invention, the elimination amount estimating means for calculating the ratio of the power of the received signal to the power of the residual signal and estimating the echo elimination amount based on the operation result, and the erasure amount Since the apparatus is provided with the allowable range updating means for updating the allowable range in accordance with the echo cancellation amount estimated by the estimating means, the allowable range can be reduced as the convergence of the echo canceling device progresses. ,
As a result, even if disturbance such as double talk occurs, there is an effect that deterioration of the echo cancellation amount can be suppressed.

According to the fourth aspect of the present invention, the counting means for counting the number of times the correction amount of the estimated impulse response value deviates from the allowable range, and the allowable range is updated according to the number of times counted by the counting means. Since the configuration is such that the allowable range updating means is provided, there is an effect that it is possible to further suppress the deterioration of the echo cancellation amount when disturbance such as double talk occurs as the convergence of the echo canceling device progresses.

According to the fifth aspect of the present invention, there is provided a call monitoring means for judging whether or not a transmission line is in use, and a method for determining whether a power of a received signal is equal to or higher than a predetermined level while the transmission line is in use. And an allowable range updating means for updating the allowable range in accordance with the time accumulated by the integrating means, especially after the echo canceling device has sufficiently converged, such as double talk. There is an effect that the amount of echo cancellation can be prevented from deteriorating when a disturbance occurs.

According to the sixth aspect of the present invention, the call monitoring means for determining whether or not the transmission line is in use, and whether or not to permit the updating of the allowable range based on the power of the received signal and the transmitted signal. Update determining means for determining whether or not the transmission line is in use, integrating means for integrating the time during which the update determining means determines that updating is permitted, and Since the configuration is such that the allowable range updating means for updating the allowable range according to the present invention is provided, it is possible to prevent the deterioration of the amount of echo cancellation when a disturbance such as double talk occurs, particularly after the echo cancellation device has sufficiently converged. There is.

According to the seventh aspect of the present invention, the call monitoring means for determining whether the transmission line is in use and the counting means for counting the number of times the correction amount of the estimated impulse response value deviates from the allowable range. And updating the allowable range in accordance with the number of times counted by the counting means, and providing the allowable range updating means for expanding the allowable range when the transmission line switches from the non-use state to the use state. Therefore, when the echo cancellation device is not converged, that is, when the impulse response of the echo path changes, there is an effect that the allowable range can be updated so that the convergence speed is increased.

According to the eighth aspect of the present invention, the upper limit value and the lower limit value of the permissible range of the impulse response are multiplied by the exponentially attenuating weight according to the sample value of the impulse response. In particular, there is an effect that the convergence speed when canceling the acoustic echo between the microphone and the speaker can be increased.

According to the ninth aspect of the present invention, the correction amount of the impulse response estimation value for each element of the impulse response is determined every several samples, so that the result is equivalent to reducing the allowable range. Thus, there is an effect that resistance to disturbances such as double talk can be further enhanced.

According to the tenth aspect, the correction amount of the impulse response estimated value is calculated from the subtraction result obtained by subtracting the pseudo echo signal from the transmission signal and the reception signal,
If the correction amount of the impulse response estimation value deviates from the allowable range set according to the number of received signals and the loop gain, the upper limit or the lower limit value of the allowable range is determined as the correction amount of the impulse response estimation value. Therefore, there is an effect that it is possible to increase resistance to disturbances such as double talk while ensuring a practical convergence speed.

According to the eleventh aspect of the present invention, the ratio between the power of the received signal and the power of the transmitted signal is calculated, and the loss of the echo path is estimated based on the calculation result. In particular, since the allowable range is updated by updating, even if disturbance such as double talk occurs when the loss of the echo path is large, there is an effect that the deterioration of the echo cancellation amount can be suppressed.

According to the twelfth aspect, the ratio between the power of the received signal and the power of the residual signal is calculated, and the echo cancellation amount is estimated based on the calculation result. Since the allowable range is configured to be updated, the allowable range can be reduced as the convergence of the echo canceller progresses, and the echo canceling amount can be reduced by that much even if a disturbance such as double talk occurs. This has the effect of suppressing the deterioration of.

According to the thirteenth aspect of the present invention, the number of times the correction amount of the impulse response estimation value deviates from the allowable range is counted, and the allowable range is updated according to the number. With the progress of convergence, there is an effect that it is possible to further suppress deterioration of the amount of echo cancellation when disturbance such as double talk occurs.

According to the fourteenth aspect of the present invention, it is determined whether or not the transmission line is in use, and while the transmission line is in use, the time during which the power of the received signal is equal to or higher than a certain level is integrated. However, since the allowable range is updated according to the time, it is possible to prevent the deterioration of the echo canceling amount when a disturbance such as double talk occurs, particularly after the echo cancellation device has sufficiently converged. .

According to the fifteenth aspect of the present invention, it is determined whether or not the transmission line is in use and whether or not to permit the updating of the allowable range based on the power of the received signal and the transmitted signal. However, since the transmission line is being used and the time for determining that the allowable range is allowed to be updated is integrated, and the allowable range is updated according to the time, the echo cancellation is particularly performed. After the apparatus has sufficiently converged, there is an effect that it is possible to prevent deterioration of the amount of echo cancellation when disturbance such as double talk occurs.

According to the sixteenth aspect of the present invention, it is determined whether or not the transmission line is in use, and the number of times the correction amount of the estimated impulse response deviates from an allowable range is counted. When the transmission line is switched from the non-use state to the use state, the permissible range is expanded. Therefore, when the echo cancellation device is not converged, that is, when the echo path is When the impulse response changes, there is an effect that the allowable range can be updated so that the convergence speed increases.

According to the seventeenth aspect of the invention, the upper limit value and the lower limit value of the allowable range of the impulse response are multiplied by the exponentially attenuated weighting value in accordance with the sample value of the impulse response. In particular, there is an effect that the convergence speed when canceling the acoustic echo between the microphone and the speaker can be increased.

According to the eighteenth aspect of the invention, the correction amount of the impulse response estimation value for each element of the impulse response is determined every several samples, so that the result is equivalent to reducing the allowable range. Thus, there is an effect that resistance to disturbances such as double talk can be further enhanced.

[Brief description of the drawings]

FIG. 1 is a configuration diagram illustrating an echo cancellation device according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing an echo canceling method applied by the echo canceling apparatus according to the first embodiment of the present invention.

FIG. 3 is a graph showing an echo canceling characteristic of the echo canceling device.

FIG. 4 is a configuration diagram showing an echo canceling device according to a second embodiment of the present invention.

FIG. 5 is a configuration diagram illustrating an echo cancellation device according to a third embodiment of the present invention.

FIG. 6 is a configuration diagram illustrating an echo cancellation apparatus according to a fourth embodiment of the present invention.

FIG. 7 is a configuration diagram illustrating an echo cancellation apparatus according to a fifth embodiment of the present invention.

FIG. 8 is a configuration diagram showing an echo canceling device according to a sixth embodiment of the present invention.

FIG. 9 is a configuration diagram illustrating an echo cancellation device according to a seventh embodiment of the present invention.

FIG. 10 is a configuration diagram illustrating an echo cancellation apparatus according to an eighth embodiment of the present invention.

FIG. 11 is a configuration diagram showing a conventional echo cancellation device.

FIG. 12 is a configuration diagram showing a conventional echo cancellation device.

FIG. 13 is a configuration diagram showing details of a loop gain control circuit.

[Explanation of symbols]

5 X register (signal storing means), 6 H register (estimated value storing means), 7 convolution integrator (echo canceling means), 8 subtractor (echo canceling means), 9 ΔH generating circuit (correction amount determining means), 21 ΔH clip section (correction amount determining means), 22 adder (estimated value correcting means), 23 echo path loss estimating section (loss estimating means), 24, 26, 2
8, 32, 35 Allowable range setting unit (allowable range updating unit), 25 echo cancellation residual power estimating unit (erasing amount estimating unit), 27 clip occurrence counting unit (counting unit), 29
Call connection state monitoring unit (call monitoring unit), 30 received signal power measurement unit (integration unit), 31, 34 timer unit (integration unit), 33 coefficient update control unit (update determination unit), 36
Exponential weighting value multiplying unit (allowable range updating means).

 ────────────────────────────────────────────────── ─── Continued on the front page (72) Inventor Tomoko Ishii 2-3-2 Marunouchi, Chiyoda-ku, Tokyo Mitsubishi Electric Corporation

Claims (18)

[Claims] 1. A signal storage means for storing a plurality of reception signals having different sampling times, and an estimation value storage means for storing an impulse response estimation value of an echo path corresponding to each reception signal stored by the signal storage means. An echo canceling unit that calculates a pseudo echo signal based on the received signal stored by the signal storage unit and the impulse response estimation value stored by the estimation value storage unit, and subtracts the pseudo echo signal from the transmission signal. In the echo canceling apparatus provided with the above, the correction amount of the impulse response estimated value is calculated from the subtraction result of the echo canceling unit and the received signal stored by the signal storage unit, and the correction amount of the impulse response estimated value is stored in the signal storage unit. Deviates from the allowable range set according to the number of received signals stored by the means and the loop gain In this case, the correction amount determining means for setting the upper limit value or the lower limit value of the permissible range to the correction amount of the impulse response estimated value, and the correction amount of the impulse response estimated value determined by the correction amount determining means is stored in the estimated value. An echo canceling device for adding an estimated value added to the impulse response estimated value stored by the means. 2. A loss estimating means for calculating a ratio of a power of a received signal to a power of a transmitting signal and estimating a loss of an echo path based on the calculation result, and a loss of the echo path estimated by the loss estimating means. 2. An echo canceling apparatus according to claim 1, further comprising an allowable range updating means for updating the allowable range according to the condition. 3. An erasure amount estimating means for calculating a ratio between a power of a received signal and a power of a residual signal as a subtraction result of the echo erasing means, and estimating an echo elimination amount based on the operation result.
2. The echo canceling apparatus according to claim 1, further comprising an allowable range updating unit that updates an allowable range in accordance with the amount of echo cancellation estimated by the cancellation amount estimating unit. 4. A counting means for counting the number of times the correction amount of the estimated impulse response value deviates from the allowable range, and an allowable range updating means for updating the allowable range according to the number counted by the counting means. The echo canceling device according to claim 1, wherein: 5. Call monitoring means for determining whether or not a transmission line is in use, and integration means for integrating the time during which the power of a received signal is equal to or higher than a certain level while the transmission line is in use. 2. The echo canceling device according to claim 1, further comprising an allowable range updating unit that updates the allowable range according to the time accumulated by the integrating unit. 6. A call monitoring means for determining whether or not a transmission line is in use, and an update determining means for determining whether or not to permit updating of an allowable range based on power of a received signal and a transmission signal. An accumulating means for accumulating a time during which the update judging means determines that the update is permitted while the transmission line is in use; and updating an allowable range according to the time accumulated by the accumulating means. 2. The echo canceling device according to claim 1, further comprising: an allowable range updating unit that performs the operation. 7. A call monitoring means for determining whether or not the transmission line is in use, a counting means for counting the number of times the correction amount of the estimated impulse response value deviates from an allowable range, and a counting means for counting the number of times. 2. An apparatus according to claim 1, further comprising: an allowable range updating means for updating the allowable range in accordance with the number of times, while increasing the allowable range when the transmission line switches from the non-use state to the use state. Echo cancellation device. 8. The allowable range updating means multiplies an upper limit value and a lower limit value of an allowable range of an impulse response by weighting values that exponentially attenuate according to a sample value of the impulse response. The echo canceling device according to any one of claims 1 to 4. 9. The method according to claim 1, wherein the correction amount determining means determines the correction amount of the impulse response estimation value for each element of the impulse response every several samples. The described echo canceling device. 10. An echo for calculating a pseudo echo signal based on a plurality of reception signals having different sampling times and an impulse response estimation value of an echo path corresponding to each reception signal, and subtracting the pseudo echo signal from a transmission signal. In the cancellation method, a correction amount of the impulse response estimation value is calculated from the subtraction result obtained by subtracting the pseudo echo signal from the transmission signal and the reception signal, and the correction amount of the impulse response estimation value is determined by the number of signals of the reception signal and the loop. If you deviate from the tolerance set according to the gain,
An echo cancellation method, wherein an upper limit value or a lower limit value of the permissible range is set as a correction amount of the impulse response estimation value, and the correction amount of the impulse response estimation value is added to the impulse response estimation value. 11. A method for calculating a ratio between the power of a reception signal and the power of a transmission signal, estimating a loss of an echo path based on the calculation result, and updating an allowable range according to the loss of the echo path. 11. The echo canceling method according to claim 10, wherein: 12. A power ratio of a received signal and a power of a residual signal which is a subtraction result obtained by subtracting a pseudo echo signal from a transmission signal, and an echo cancellation amount is estimated based on the calculation result. 11. The echo canceling method according to claim 10, wherein the allowable range is updated according to the echo cancellation amount. 13. The echo canceling method according to claim 10, wherein the number of times the amount of correction of the estimated impulse response value deviates from the allowable range is counted, and the allowable range is updated according to the number of times. 14. A method for determining whether or not a transmission line is in use, accumulating a time during which the power of a received signal is equal to or higher than a certain level in a state where the transmission line is in use, and permitting according to the time. The echo canceling method according to claim 10, wherein the range is updated. 15. A method for determining whether a transmission line is in use and determining whether to permit updating of an allowable range based on power of a received signal and a transmission signal, and determining whether the transmission line is in use. 11. The echo canceling method according to claim 10, further comprising: accumulating a time during which it is determined that the updating of the allowable range is permitted in the state, and updating the allowable range according to the time. 16. A method for determining whether a transmission line is in use or not, counting the number of times the correction amount of the impulse response estimation value deviates from an allowable range, and updating the allowable range according to the number of times. 11. The echo canceling method according to claim 10, wherein when the transmission line switches from a non-use state to a use state, the permissible range is expanded. 17. The method according to claim 11, wherein an upper limit value and a lower limit value of an allowable range of the impulse response are multiplied by weighting values that exponentially attenuate according to a sample value of the impulse response. The echo canceling method according to any one of the above items. 18. The echo canceling method according to claim 10, wherein the correction amount of the impulse response estimation value for each element of the impulse response is determined every several samples.