JPH07226699A - Echo canceler - Google Patents

Abstract

PURPOSE:To more exactly obtain the conformity of the echo path with a pseudo echo path. CONSTITUTION:In a power ratio arithmetic part 51, the ratio ER(n)=Pz(n)/Pe(n) of the power Pz(n) of a microphone output z(n) and the power Pe(n) of an error signal e(n) is calculated. In a reverse peak hold circuit 52, ipk(n)=min[ER (n), alpha.ipk(n-t)] (ipk (n): the output of a circuit 52, alpha=1.0001, min [a, b]: a function outputting either smaller one of a and b) is calculated. The larger the output ipk(n) of the circuit 52 is, the smaller the insertion loss of an adaptive loss controller 26 is.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an echo control device for eliminating or suppressing echo signals which are a cause of howling and an auditory obstacle in a two-wire to four-wire conversion system and a voice communication system.

[0002]

2. Description of the Related Art FIG. 5A is a schematic diagram of a voice communication system.
The transmitted voice uttered by the transmitter 10 is sequentially supplied to the reception speaker 15 through the transmission microphone 11, the transmission signal amplifier 12, the transmission line 13a, and the reception signal amplifier 14, and the reproduced voice from the speaker 15 is received. It is transmitted to person 16. On the other hand, the voice uttered by the listener 16 is the microphone 17 for transmission, the transmission signal amplifier 18, the transmission path 13b, the reception signal amplifier 19,
The signal is transmitted to the speaker 10 through the receiving speaker 21 in sequence. Unlike the conventional telephone call system, this voice call system does not need to have a handset, so it has the advantage of being able to make a call while working, or to realize a natural face-to-face call, It is widely used in communication conferences, videophones, and loud-speakers.

On the other hand, as a drawback of this voice communication system, the existence of echo is a problem. That is, the voice transmitted from the speaker 15 to the receiving side in FIG. 5A is received by the microphone 17 as indicated by the arrow 22, and is reproduced to the transmitting side via the amplifier 18, the transmission path 13b, the amplifier 19, and the speaker 21. . For the speaker 10, this phenomenon is a reverberation phenomenon in which the voice uttered by the speaker 10 is reproduced from the speaker 21, and is called an acoustic echo or the like. This reverberation phenomenon causes adverse effects such as call disturbance and discomfort in a loud voice communication system. Further, the sound reproduced from the speaker 21 is received by the microphone 11 and forms a closed loop of the signal. When the gain of 20 loops is larger than 1, the howling phenomenon occurs and the call becomes impossible.

In order to overcome the above problems of the voice communication system, a loss control device and an echo canceller are used. FIG. 5B is a schematic diagram showing an example of the loss control device,
The same parts as those in FIG. 5A are assigned the same numbers. Also,
For simplicity, the amplifier is omitted, and only the receiver 16 side in FIG. 5A is shown. The transmission line 13a is connected to the loss control circuit 23.
From the microphone 17 and the output signal z (n) of the microphone 17 (n is a parameter indicating time) are input, and the transmission / reception state is determined based on their magnitudes. As a method of determining the transmission / reception state, for example, each short-time power Px of the reception signal x (n) and the output signal z (n)
(N) and Pz (n) are calculated and their sizes are compared.
Here, assuming that the echo signal y (n) passing through the arrow 22, that is, the echo path is smaller than the reception signal x (n), when there is no transmission signal s (n) uttered by the receiver 16, Px (n)> Pz (n) is established, and at this time, it is determined that the receiving state. On the other hand, the transmission signal s above a certain level
If (n) exists, Px (n) <Pz (n),
At this time, it is determined to be in the transmitting state. Then, if it is determined to be in the receiving state, the loss is inserted into the transmitting side loss device 24 inserted into the output side of the microphone 17. As a result, the speaker 15
The reverberant signal y (n) that wraps around from and is received by the microphone 17 is attenuated by the loss device 24 and the reverberant phenomenon is reduced. On the other hand, if it is determined to be in the transmitting state, the loss of the transmitting-side loss device 24 is set to 0 (dB), and the loss is inserted into the receiving-side loss device 25 inserted before the speaker 15. As a result, the transmission signal received by the microphone 17 is transmitted without being attenuated. Further, by inserting the loss into the receiving side, the closed loop gain can be maintained at 1 or less, and the howling phenomenon can be prevented.

As described above, the loss control device 26
It is possible to reduce the echo phenomenon by using
The insertion loss amount may be fixed in order to keep the configuration simple. In this case, a large value (for example, 20 dB) is given as the value of the insertion loss amount assuming all situations. However, if the inserted loss amount exceeds 10 dB, the beginning or end of the voice is cut off due to a time delay in the transmission / reception determination, which causes a problem that the communication quality deteriorates.

[0006] Therefore, the acoustic connection in the actual use situation
Adaptive loss control that adaptively controls the loss according to the total amount
A device has been proposed. FIG. 6A shows an adaptive loss control device.
FIG. The same numbers are assigned to the same parts as in FIG. 5B.
Granted. Passes the loss device 25 to the insertion loss amount determination circuit 27
Signal L after _RFrom x (n) and microphone 17
Output signal z (n) is input and based on these levels
Sound between speaker 15 and microphone 17
Estimate the coupling amount G and determine the insertion loss amount according to the estimated amount.
Set. For example, the predicted value G ′ of the acoustic coupling amount is L_Rx
(N) and z (n) short time power PL_Rx (n) and P
It can be calculated as follows using z (n).

G '= Pz (n) / PL _R x (n) If this predicted coupling amount G'is larger than 1 (0 dB),
The insertion loss amount determining circuit 27 calculates a loss amount such that the gain between x (n) and L _T z (n) becomes 1 or less. For example, when G'is 4 (6 dB), the input signal x
In order to keep the open loop gain between (n) and the transmission signal L _T z (n) at 1 (0 dB) or less, the insertion loss amount is 0.5.
(6 dB in power) or less. Further, when the acoustic coupling amount G changes and G ′ becomes 2 (3 dB), the insertion loss amount is set to 0.7 (3 dB in power) or less. The insertion loss amount thus determined is transferred to the loss control circuit 23. The loss control circuit 23 determines the loss device 2 based on the above-described judgment.
The amount of insertion loss is given to 4, 25.

As described above, in the adaptive loss control device, the insertion loss amount is determined according to the magnitude of the acoustic coupling amount G of the echo path, and the loss is inserted to minimize the deterioration of the communication quality. It is possible to hold down. However, it is difficult to know an accurate value because the acoustic coupling amount G is estimated for a short time with high power.

As described above, the echo phenomenon can be reduced by using the adaptive loss control device.
Under the condition that the acoustic coupling amount G is large, there remains a problem that the call quality is deteriorated. The echo canceller described below has been introduced in recent years as a new technique that does not cause such a problem. FIG. 6B shows an example of a conventional echo canceller 31. In the echo canceller 31, first, the echo path estimation circuit 32 estimates the impulse response (echo path transmission characteristic) of the echo path 22, and the estimated value h
(N) ′ is transferred to the pseudo echo path 33. Next, in the pseudo echo path 33, the convolution operation of the estimated impulse response h (n) ′ and the received signal x (n) is executed to execute the pseudo echo signal y.
(N) 'is synthesized. Then, in the subtractor 34, the pseudo echo signal y is output from the output signal z (n) of the microphone 17.
Subtract (n) '. If the impulse response of the echo path 22 is well estimated, the echo signal y (n) and the pseudo echo signal y (n) 'are almost equal to each other, and as a result of the subtraction by the subtractor 34, The echo signal y (n) contained in the microphone output z (n) is deleted. Note that the echo signal y (n) and the signal s (n) cannot be distinguished only by looking at the output z (n) of the microphone 17, and therefore z (n) may also be referred to as the echo signal.

Here, the pseudo echo path 33 needs to follow the temporal change of the characteristic h (n) of the echo path 22. Therefore, the echo path estimation circuit 32 estimates the impulse response of the echo path 22 using an adaptive algorithm. This estimation operation is a receiving state, that is, s (n) ≈0, and z (n) ≈y.
It is executed when it can be regarded as (n). In the listening state,
The output of the subtractor 34, that is, the error signal e (n) can be regarded as the cancellation residual y (n) -y (n) 'of the echo signal y (n). In the following description, this receiving state is assumed.
The adaptive algorithm is the received signal x (n) and the error signal e.
It is an algorithm that uses (n) to determine the estimated value h (n) 'of the impulse response so that the power of e (n) is minimized, and the LMS method, learning identification method, ES method, etc. are known. There is.

Next, a typical adaptive algorithm will be described. When a digital FIR filter is used as the pseudo echo path 33, since the impulse response of the filter and the filter coefficient match, the estimated value h (n) 'of the echo path impulse response will be referred to as a filter coefficient. In the adaptive algorithm, the estimation of h (n) 'is done in the iterative form. That is, the filter coefficient h (n + 1) 'at time n + 1 is calculated by the following equation, which is obtained by modifying the filter coefficient h (n)' at time n.

H (n + 1) ′ = h (n) ′ + αΔ (n) (1) However, h (n) ′ = (h1 (n) ′, h2 (n) ′, ..., hL
(N) ′) ^T : filter coefficient (vector representing impulse response of pseudo echo path 33 at time n) Δ (n): vector representing modification direction of coefficient α: step size (parameter representing modification size of coefficient : Scalar amount) L: Number of taps (number of filter coefficients) ^T : Transposition of vector n: Discretization time Δ (n) differs depending on the algorithm. For example, in the LMS method, Δ (n) = e (n ) X (n) (2) In the learning identification method, Δ (n) = e (n) X (n) / (X (n) ^T X (n)) (3). However, e (n): error signal (= y (n) -y (n) ') y (n)' = h (n) ' ^T X (n) X (n) = (x (n), x (n-1), ..., x (n-L + 1)) ^T : Received signal vector Here, the characteristic of the pseudo echo path 32 is close to the characteristic of the true echo path 22, and the pseudo echo signal y (n) ' Is converged when it becomes almost equal to the echo signal y (n).

In various adaptive algorithms, the step size parameter α is usually set to 1. The step size parameter α is an amount that influences the state of convergence. When the value is 1, the convergence speed becomes maximum, and when it is set to a value smaller than 1, the convergence speed becomes slower, but the final error value Is smaller than when 1. For this reason, a variable step size parameter type adaptive algorithm has been proposed in which the step size parameter α is set to 1 in the initial state, and when it converges to some extent, the step size parameter α is reduced. In such an algorithm, it is necessary to correctly judge the state of convergence.

Further, the algorithm for estimating the characteristic of the echo path 22 is as follows when the transmission signal of the near-end speaker is present.
This is regarded as an error and the pseudo echo path 32 is set in the wrong direction. Therefore, it is necessary to detect the presence (double talk) of the reception signal x (n) and the transmission signal s (n) of the near-end speaker to stop the estimation of the echo path in the double talk state.

As a method for solving the problem of stopping the adaptive estimation during double talk, an FG / BG (foreground / background) method is proposed, in which the echo path is satisfactorily estimated without explicitly detecting the table talk. Has been done. FIG. 7A shows an echo canceller 36 configured using the FG / BG method, and the same parts as those in FIG. 6B are assigned the same numbers. In the FG / BG echo canceller 36, the echo path estimation circuit 32 first estimates the impulse response of the echo path 22 and transfers the estimated value hb (n) ′ to the pseudo echo path 33 on the BG side. Next, in the pseudo echo path 33 on the BG side, a convolution operation of hb (n) 'and the received signal x (n) is executed to synthesize the pseudo echo signal yb (n)'.
Then, the subtractor 34 subtracts the pseudo echo signal yb (n) 'from the output signal z (n) of the microphone 17.
If the echo path impulse response is well estimated,
The echo signal y (n) and the pseudo echo signal yb (n) 'are almost equal. If the characteristic of the pseudo echo path 33 on the BG side is close to the characteristic of the true echo path 22, the pseudo echo path 33 on the BG side is obtained.
Is transferred to the coefficient of the pseudo echo path 37 on the FG side. In general, the fact that the pseudo echo path 33 on the BG side approaches the characteristics of the true echo path 22 means that the difference eb (n) and z (between the microphone output signal z (n) and the pseudo echo signal yb (n) '. This is done by comparing the power with n). Here, the power is a time integral value of a signal, and when handling a discretized signal, for example, Px (n) = Σx ² (n−i) (Σ is i
= 0 to k−1). Here, the integration time is shown. The transfer of the coefficients is performed as follows.
That is, when the input signal x (n) is greater than or equal to the set threshold value in the input determination circuit 38, the power comparison circuit 39 determines that the power of the error signal eb (n) is smaller than the power of z (n) to some extent or more. And the error comparison circuit 41 outputs B
When it is determined that the power of the G-side error signal eb (n) is smaller than the power of the FG-side error signal ef (n) (difference between the output of the FG-side pseudo echo path 37 and z (n)), The estimated coefficient hb (n) ′ is the FG-side pseudo echo path coefficient hf.
It is considered that the impulse response of the actual echo path 22 is better simulated than that of (n) ', and therefore the coefficient hb (n)' of the BG side pseudo echo path 33 is set to the FG side pseudo echo path 37.
Transfer to. Coefficient hf (n) 'of the FG side pseudo echo path 37
Is updated only when the above condition is satisfied, so if the echo path estimation circuit 32 on the BG side makes an erroneous estimation during double talk, the above condition is not satisfied and the BG coefficient is transferred to the FG side. And the good echo cancellation before double talk is retained. The pseudo echo signal yf (n) 'is subtracted from z (n) by the subtractor 42 from the FG side pseudo echo path 37,
The subtraction output ef (n) is output to the transmission line 13b.

However, in the FG / BG method, even if the characteristic of the pseudo echo path 33 on the background side is different from that of the true echo path 22, the transfer determination circuit 43 satisfies the above condition and an erroneous BG pseudo echo echo is generated. The coefficient of the path 33 may be transferred to the pseudo echo path 37 on the FG side. This occurs because the power of the error signal e (n) becomes small even when the coefficient of the pseudo echo path is incorrect, particularly when the input signal x (n) is a vowel having a sinusoidal waveform.

Further, the echo canceller 31 described above.
Also, the FG / BG echo canceller 36 has a problem in that a sufficient echo canceling amount cannot be obtained with respect to the change of the echo path 22 in the initial stage or during use, resulting in howling. The way to solve this problem is to do initial learning or
Alternatively, when the echo canceller does not converge, it is conceivable that the loss control device 28 inserts a loss to the extent that howling does not occur. In the following, an example in which the echo canceller and the adaptive loss controller described above are combined is shown in FIGS. 7B and 8. In FIG. 7B, the echo canceller 3
An adaptive loss control device 26 is inserted between 1 and the transmission lines 13a and 13b. FIG. 8 shows an example in which the adaptive loss control device 26 is inserted between the echo canceller 36 using the FG / BG method and the transmission lines 13a and 13b. In these figures, the same parts as those in FIGS. 6A, 6B and 7A are designated by the same reference numerals. In these devices, the adaptive loss control device 26 inserts a loss corresponding to the acoustic coupling amount between the speaker 15 including the echo canceller 31 or 36 and the microphone 17. Therefore, when the echo canceling amount of the echo canceller 31 or 36 is small in the initial stage,
If the echo loss is suppressed by increasing the insertion loss amount of the adaptive loss control device 26 and the echo signal is removed to some extent by the echo canceller 31 or 36, the adaptive loss control device 2
The amount of insertion loss of No. 6 can be reduced.

As described above, it can be understood that howling suppression, which is a problem in the echo canceller when the echo canceller is small, can be solved by combining the echo canceller and the adaptive loss controller. However, if the echo canceller does not correctly determine how much echo has been removed, for example, the insertion loss amount may change below the insertion loss amount required to prevent howling.

[0019]

As described above, in the conventional echo canceller, the pseudo echo path may not simulate the true echo path even when the error signal becomes smaller than the echo signal. is there. In such a case, in the variable step size parameter type adaptive algorithm, since it is determined that the variable step size parameter has converged even if it has not converged, there is a problem that the step size parameter is reduced and the convergence becomes slow. . Further, in the FG / BG echo canceller, an erroneous transfer occurs in which a coefficient on the BG side that has not converged is transferred to a coefficient on the FG side. Further, in the device in which the echo canceller and the adaptive loss control device are combined, the insertion loss amount may be changed to be equal to or less than the insertion loss amount necessary for preventing howling. In order to solve such a problem, an object of the present invention is to provide an echo canceller in which a pseudo echo path correctly simulates a true echo path.

[0020]

According to the invention of claim 1, the ratio of the power of the echo signal z (n) and the power of the error signal e (n) is calculated, and the calculated inverse ratio is calculated. The peak value is detected by the inverse peak hold means, and is output as a signal indicating the matching of the pseudo echo path with the echo path. According to the invention of claim 2, the invention of claim 1 is provided with an adaptive loss controller, and the loss amount thereof is controlled by the output of the inverse peak hold means.

According to the third aspect of the present invention, in the first or second aspect of the invention, the echo canceling is performed by the BG / FG method, and the pseudo echo path on the BG side in the BG / FG method is in good agreement with the echo path. The output of the inverse peak hold means is used for the determination. According to the invention of claim 4, claims 1 to 3
In any one of the above aspects, the step size parameter that determines the degree of convergence in the adaptive algorithm for echo path estimation is controlled according to the output of the inverse peak hold means.

[0022]

1 shows an embodiment in which the present invention is applied to a device in which an echo canceller 31 and an adaptive loss controller 26 are combined, and the same parts as those in FIG. 7B are designated by the same reference numerals.
In the present invention, the power ratio ER between the power Pz (n) of the microphone output (echo signal) z (n) and the power Pe (n) of the error signal e (n) in the power ratio calculation unit 51.
(N) = z (n) / Pe (n) is calculated. The reverse peak hold circuit 52 continuously detects the smallest value of the power ratio ER within a certain time, that is, the reverse peak value, that is, the moving reverse peak value is detected. That is, the inverse peak hold circuit 52 performs the following calculation.

Ipk (n) = min [ER (n), β ·
ipk (n-1)] where ipk (n) is an inverse peak value and β is an inverse damping constant of 1
As described above, the value is very close to 1, and depends on the convergence speed of the algorithm used and the number of taps of the adaptive filter used, but the value should be as close as possible to the convergence speed of the algorithm. 1.000 in the commonly used learning identification method
About 1 ± 0.00005 is good. min [a, b] is a
Is a function that compares b with b and outputs the smaller value.
The output of the inverse peak hold circuit 52 is used as the pseudo echo path 3
The judgment of convergence is 3. That is, it is determined that the larger the output of the inverse peak hold circuit 52, the better the convergence state.

FIG. 2 is a graph showing the effect of the present invention, showing the state of convergence. The horizontal axis represents time, and the vertical axis represents a value indicating the closeness between the pseudo echo path 33 and the true echo path 22. The larger the value, the closer the value is. This is the result of the computer simulation, and the dotted line 54 indicates the microphone output signal z (n) and the error signal e.
The ratio of each power Py (n) and Pe (n) to (n) (10
log (Pz (n) / Pe (n))),
It is a value that has been conventionally used for the determination of convergence. However, for the integration time when calculating the power, a short value is used here because there is a time lag between the determination of convergence and the actual convergence state when the integration time is long. Solid line 55
Is an inverse peak hold value ipeka (n) calculated by the inverse peak hold circuit 52. However, the reverse subtraction constant β
Is set to 1.0001, and the minimum value of ipeka (n) is 5d.
B. Further, the alternate long and short dash line 56 shows a true convergence state between the echo path 22 and the pseudo echo path 33. That is, since the impulse response h (n) of the echo path 22 is known because it is a simulation, the degree of approximation of the impulse response of the pseudo echo path 33 to this is shown.
^{_{og 10 h i (n) 2}} / ((h i (n) -h
_i (n) ′)) ² (Σ is i = 1 to L).

The arrow 57 indicates the time when the echo path 22 changes at that time. From FIG. 2, it can be seen that in the case of the conventional power comparison (curve 54), the value representing the convergent state may be large even when the pseudo echo path 33 is not close to the true echo path 22. For example, if the convergence is determined to be 10 dB or more, an incorrect determination may be made. On the other hand, when the reverse peak hold value (curve 55) is used, the variation is small, and the pseudo echo path 33 is actually close to the echo path 22 (curve 56). For example, after the arrow 57, it does not actually converge, and the curve 55 is 10 dB or less.
4 has a portion exceeding 10 dB, and there is a risk of misjudgment of convergence. Therefore, it is more accurate to evaluate the true convergence of the echo path 22 and the pseudo echo path 33 by using the inverse peak hold value to evaluate the convergence than the conventional power comparison. You can see that it shows.

In the present invention, since the reverse peak hold value is used to judge the convergence state, the reverse peak hold circuit 5
In No. 2, since the convergence state in the initial stage is poor and the minimum value is raised at the expected convergence speed, the faster convergence is discarded as an error. For this reason, erroneous calculations that converged very quickly are reduced, and variations are reduced. For vowels, although it may not appear to have converged, it may appear to converge early, but if it does not converge at the beginning of a vowel, it is retained and calculated as a value close to an accurate value. Therefore, it does not make an erroneous judgment.

In the reverse peak hold circuit 52, for example, 10 l
ogipk (n) is calculated, and if it is, for example, 10 dB or more, it is determined that the echo signal has been eliminated,
In the adaptive insertion control device 26, the inverse peak hold circuit 5 is arranged so that the loss amount is reduced by 10 dB from the initial value.
The insertion loss amount determining circuit 27 determines the insertion loss amount in accordance with the output of 2.

This first application of the present invention to an apparatus combining an echo canceller 31 and an adaptive loss controller 26
In this embodiment, it is possible to calculate a value close to the value actually converged, as is clear from FIG. The adaptive loss control device 26 is
The insertion loss amount can be reduced as the echo cancellation amount obtained by the echo canceller 31 increases, but in the conventional power comparison of z (n) and e (n),
Even if the echo cancellation is not actually performed, it may be determined that the echo cancellation is obtained by calculation, and the insertion loss amount may be reduced more than necessary. On the other hand, in the present invention, the value calculated by the inverse peak hold circuit 52 (curve 55) is close to the actually converged value (curve 56). Therefore, by sending this result to the insertion loss amount determination circuit 27, an accurate insertion loss can be obtained. The amount can be determined.

In this embodiment, the convergence judgment according to the present invention is applied to a device in which the adaptive loss control device 26 and the echo canceller 31 are combined, but the combination method is adaptive loss control device 26 and FG. The same can be applied to a device in which the / BG type echo canceller 36 is combined.
Next, FIG. 3 shows an example in which the present invention is applied as a second embodiment to the echo path estimation adaptive algorithm of the echo canceller.
In the figure, the same parts as those in FIG. 6B are given the same numbers. LMS, learning identification method, projection algorithm, RLS
In various adaptive algorithms such as, the step size parameter is set to 1 in order to speed up the convergence speed in the stage where it does not converge, and when it converges to some extent, the step size parameter is decreased in order to reduce the final steady elimination amount. A variable-step-size parameter type adaptive algorithm has been proposed for reducing. In this variable step size parameter type algorithm, the step size parameter is made variable depending on the degree of convergence, so the convergence state must be calculated accurately. According to the present invention, as shown in FIG.
Whether B has converged can be calculated close to the actual value. Therefore, in this embodiment, the output of the inverse peak hold circuit 52 is input to the step size parameter variable circuit 58. For example, in FIG. 2, the output of the inverse peak hold circuit 52, that is, the step size parameter is 1 when the convergence value is 10 dB or less,
0.5, 15d when the convergence value is between 10 and 15dB
If it is B or more, it is set to 0.1. The step size parameter controlled in this way is used as the step size of the adaptive algorithm in the estimation circuit 32. As can be seen from the curve 55, in the present invention, the step size can be changed in a state where there is little variation with time. On the other hand, in the conventional determination of the degree of convergence, since the converged value due to time is different from the actual converged state (curve 56), the step size parameter is reduced to 0.1 or the like even if the converged value is not actually converged. It may happen.

Next, as a third embodiment of the present invention, FG
The case of application to the echo canceller using the / BG method will be described. FIG. 4 shows an example thereof, and the same numbers are given to the portions common to FIGS. 7A and 1. Conventional FG / B
In the G method, as described above, the transfer condition that the pseudo echo path on the BG side is close to the true echo path is defined as the error signal eb.
It is determined whether or not the power of (n) becomes smaller than the power of the echo signal z (n) to some extent. On the other hand, when the present invention is applied to the FG / BG system, as described with reference to FIG. 7A, the input signal x (n) is not less than the threshold value and the power of the error signal eb (n) on the BG side is FG side. Is smaller than the power of the error signal ef (n), and the transfer condition that the pseudo echo path 33 on the BG side is close to the true echo path 22 is determined by whether the output of the inverse peak hold circuit 52 is equal to or more than a predetermined value. As described above with reference to FIG. 2, the inverse peak hold value can more accurately represent the difference between the echo path and the pseudo echo path than the power comparison. Therefore, the pseudo echo path 3 on the BG side that occurs during conventional power comparison
The problem that the coefficient of the BG side pseudo echo path 33 is transferred to the FG side pseudo echo path 37 when 3 is different from the true echo path 22 is solved by the comparison using the inverse peak hold value.

The estimation circuit 32 shown in FIGS. 1 and 4 is also shown in FIG.
Step size parameter variable circuit 58 described above
And the step size may be changed by the output of the inverse peak hold circuit 52.

[0032]

As described above, the present invention evaluates how close the pseudo echo path is to the true echo path by using the inverse peak hold value of the power ratio between the echo signal and the error signal. It can be performed as a value closer to the actual value than the conventional simple comparison with power.

When this is applied to a device in which the echo canceller and the adaptive loss controller are combined, it is possible to calculate how much echo is canceled by the echo canceller with a value close to the actual value, and insert it. The loss amount can be changed accurately, which leads to an improvement in call quality. Also, in the echo path estimation adaptive algorithm that makes the step size parameter variable according to the state of echo cancellation, the step size parameter can be changed accurately by using the inverse peak hold value for the judgment of the echo canceller. In the stage where the size parameter type adaptive algorithm is not converged, which is the purpose, the step size parameter is set to 1 in order to accelerate the convergence speed. It is possible to maximize the effect of reducing the size parameter.

Further, when applied to the echo canceller using the FG / BG method, which has been difficult to calculate correctly, it is erroneous B that occurred in the conventional power comparison.
Since the coefficient on the G side is not transferred to the FG side, the call quality is improved.

[Brief description of drawings]

FIG. 1 is a block diagram showing a first embodiment in which the present invention is applied to a device in which an adaptive loss control device and an echo canceller are combined.

FIG. 2 is a diagram showing a change state of an amount indicating various types of convergence for explaining the effect of the present invention.

FIG. 3 is a block diagram showing a second embodiment in which the present invention is applied to a variable step size parameter type algorithm.

FIG. 4 is a block diagram showing a third embodiment in which the present invention is applied to an FG / BG type echo canceller.

FIG. 5A is a schematic diagram of a voice communication system, and B is a block diagram showing a conventional loss insertion device.

FIG. 6A is a block diagram showing a conventional adaptive loss control apparatus, and B is a block diagram showing a conventional echo canceller.

7A is a block diagram showing a conventional FG / BG echo canceller, and FIG. 7B is a block diagram showing a device in which a conventional adaptive loss control device and an echo canceller are combined.

FIG. 8 is a block diagram showing an apparatus in which a conventional adaptive loss control apparatus and an FG / BG echo canceller are combined.

Front Page Continuation (72) Inventor Masafumi Tanaka 1-1-6 Uchisaiwaicho, Chiyoda-ku, Tokyo Nihon Telegraph and Telephone Corporation

Claims (4)

[Claims] 1. A transfer characteristic of the echo path is estimated by an estimating means from a signal sent to the echo path and an error signal, the estimated transfer characteristic is set to a pseudo echo path, and the transfer characteristic is sent to the pseudo echo path. In the echo canceling device, which generates a pseudo echo signal through the signal, subtracts the pseudo echo signal from the echo signal whose transmission signal has passed through the echo path by subtraction means, and uses the rest as the error signal, Means for calculating the ratio between the power and the power of the error signal, and inverse peak hold means for outputting the moving inverse peak value of the calculated ratio as a signal indicating the coincidence of the pseudo echo path with the echo path. And an echo canceller characterized by being provided. 2. The loss given to the transmission signal and the loss given to the error signal are controlled in reverse according to the transmission signal and the echo signal, and the control amount is set to the echo of the pseudo echo path. An adaptive loss control means for adaptively controlling according to the matching with the road is provided, and the adaptive control of the control amount is performed based on the output of the inverse peak hold means. The echo canceller according to Item 1. 3. The output signal is passed through a second pseudo echo path to obtain a second pseudo echo signal, the second pseudo echo signal is subtracted from the echo signal, and a second error signal is used as an output signal. BG / FG which detects a state where the 2 error signal is larger than the error signal and the pseudo echo path is in good agreement with the echo path, and transfers the characteristics of the pseudo echo path to the second pseudo echo path.
The echo canceller according to claim 1 or 2, wherein the output of the inverse peak hold means is used for determining the coincident state in the BG / FG method. 4. The step size parameter for determining the degree of convergence in the adaptive algorithm for echo path estimation in the estimation means is controlled according to the output of the inverse peak hold means. The echo canceller described in.