JPS62130029A - Echo canceller - Google Patents

Abstract

PURPOSE:To attain echo cancellation by stopping the identification of an echo path of an echo canceller when the 1st echo cancellation effect is lowered and starting the estimate of the echo path when the effect of the 2nd echo cancellation is equal to the estimated echo cancellation effect thereby applying high speed tracking to fluctuation. CONSTITUTION:If a nearby talker or an echo path fluctuation exists, the echo cancellation effect of an echo canceller 5 operated by an echo cancellation effect operating means 27 is lowered and larger than the estimated echo cancellation effect operated by the estimated echo cancellation effect operation means 28. A comparison means 30 detects it, a decision means 32 stops identification of the echo path of the canceller 5 immediately, compares the echo cancellation effect of an echo canceller 6 operated by an echo cancellation effect operation means 29 with the estimated echo cancellation effect, and decides it as the fluctuation of the echo path when the both are equal, and a control means starts the estimation of the echo path of the canceller 5. When no nearby talker exists, the canceller 5 simulates the echo path to cancel the echo signal. In other case, it is regarded as a simultaneous talking, the identification of the echo path by the canceller 5 remains stopped and the canceller 6 clears the identified echo path at each prescribed time.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to an echo canceling device for controlling echoes in long-distance circuits such as satellite communications, teleconference systems, and the like.

2. Description of the Related Art In recent years, with the rapid increase in long-distance lines, especially satellite communications, and the practical use of teleconference/systems, clear and natural γf voice transmission has been desired. 3-echo signal U is a phenomenon in which the signal generated by the sender 1 returns to the sender due to reflection etc. at the receiver 11i11. The presence of echoes causes deterioration in call quality. Recently, adaptive echo cancelers that cancel this echo are being put into practical use, but these are remote-operated devices (identifying the echo path based on the far-end speaker's voice). Detect presence (- is required (for example, "How to detect simultaneous two-way calls")
(1989 National Conference of the Telecommunications Society of the Institute of Electronics and Communication Engineers 589))
.

The conventional echo canceling device as described above will be described below with reference to the drawings.

FIG. 7 shows the configuration of a conventional echo canceling device.

In Fig. 7, 1 is the receiving side input terminal, 2 is the receiving side input terminal 11+
11 output terminals, 3 is a transmitting side input terminal, 4 is a transmitting side output terminal, and 6 is an echo canceller provided between each terminal to cancel the echo that returns from the !II output terminal to the transmitting side output terminal. Reference numeral 56 denotes an erased residual signal level detector for detecting the level of the erased residual signal which is the transmission output of the echo cancellers, 57 denotes a received signal level detector, and 68 denotes the erased residual signal level Lsout, the received signal level Ltin, and the threshold used last time. A threshold updating unit calculates the threshold value Kt to be used this time from the value Kt, 59 is a threshold storage unit that stores the threshold value Kt sent from the threshold updating unit, and 60 is a storage unit that stores the threshold value Kt and the erased residual. A receiving side signal level estimator 61 estimates the receiving side signal level from the signal level, and 61 is the estimated receiving side signal level I, tin and the actual receiving side signal L/hell L.
A comparison unit that compares r in and controls whether or not to identify the echo path of the echo canceller.

The operation of the echo canceling device configured as described above will be explained below.

First, the voice signal of the far end speaker is supplied to the echo canceller 6 via the receiving side input terminal 1 as the receiving human power Rin.

Also, the signal Rin is sent from the receiving side output terminal 2 to the near-end speaker as a reception signal Rout.

When there is no near-end speaker signal, the signal Rout enters the transmitter input terminal 3 as the transmitter input signal Sin through the echo path. In the echo canceller 6, a pseudo echo path is created from the sequence of the receiving human power Rin and the transmission output 5out so that the transmission side output signal 5out (also called the cancellation residual signal) becomes zero. In this way, the echo signal is canceled, but if a near-end speaker is present,
The echo path cannot be estimated well. Therefore, it is necessary to detect the presence of a near-end speaker signal and stop estimating the echo path. Therefore, the presence or absence of a near-end speaker is detected from the relationship between the receiving side input signal Rin and the erasure residual signal 5outO, which is the transmitting side output signal. A receiving side signal level Lx in is detected by the receiving side signal level detection section 57 . The erased residual signal level detection section 66 detects the erased residual signal level L8out. In the threshold value updating section 58, the receiving side signal level Lr1
n, erased residual signal level Lgout, and threshold storage section 59
The threshold value Kt used last time is received from the comparator 61, and the signal Adp indicating the stop or continuation of the echo path estimation is received from the comparator 61 to calculate the threshold value Kt' used this time and send it to the threshold storage unit 69. Next, the receiving side signal level estimation unit 6o calculates the estimated value L r i nA of the receiving side signal level from the threshold value Kt' and the erased residual signal level Lsout. The comparison unit 61 compares the estimated receiving side signal level and the actual receiving side signal level Lr1.
When Lr 1nA) Lr in, it is assumed that a near-end speaker exists and estimation of the echo path is stopped.

As for the method of updating the threshold value, when the estimation of the echo path is progressing, the threshold value is increased little by little so as not to exceed the convergence speed of the echo canceller.

If there are only near-end speakers, the threshold value is not updated.

In the case of simultaneous two-way communication or a change in the echo path, the threshold value is decreased by a value proportional to the amount of cancellation attenuation so that the time required for determining that the change in the echo path is not caused by the signal level on the transmitting side does not change significantly. The above judgment time is approximately 3 seconds.

Problems to be Solved by the Invention However, with the above configuration, when holding a conference with a large number of people using a teleconference system, there is a high possibility that two-way simultaneous calls lasting more than 3 seconds will occur, so the determination time will be longer. However, it had the disadvantage of being slow to follow changes in the echo path.

In view of the two-legged problem, the present invention provides an echo canceling device that can quickly determine two-way simultaneous communication and fluctuations in the echo path.

Means for Solving the Problems To achieve this object, the echo canceler of the present invention includes a first echo canceller that cancels the echo signal contained in the input signal on the transmitting side, and a first echo canceller that cancels the echo signal included in the input signal on the receiving side and the transmitting side input signal. first and second filters for extracting signals in a predetermined band of side output signals; first and second thinning means for thinning out the output signals of the first and second filters; from a second echo canceller whose output is a receiving side input and the output of the second thinning means is a transmitting side input, a receiving side input signal, a transmitting side input/output signal, and a transmitting side input/output signal of the second echo canceller. It is comprised of a two-way simultaneous conversation detection means that detects the presence or absence of a near-end speaker and controls the first and second echo cancellers.

The two-way simultaneous call detection means includes a receiving side input signal level detecting means for detecting a receiving side input signal level, a transmitting side input signal level detecting means for detecting a transmitting side input/output signal level,
Transmission 1111 output signal level detection means, and a first unit that calculates the ratio of the transmission side input/output signal (transmission side output signal level/transmission side output signal level, hereinafter referred to as the first echo cancellation effect).
sub-band transmitting side input signal level detection means for detecting the output signal level of the second thinning means; and detection of the residual signal level after the second echo canceller cancels the echo signal. The subband transmitting side output signal level detecting means detects the ratio of the subband transmitting side input/output signal level (subband transmitting side input signal level/subband transmitting side output signal level, hereinafter referred to as the second echo cancellation effect). The current estimated value of the first echo-cancelling effect (hereinafter referred to as estimated an estimated echo-cancelling effect calculation means for calculating an estimated echo-cancelling effect (referred to as an echo-cancelling effect);
a second comparing means for comparing the estimated echo canceling effect and the second echo canceling effect, and a third comparing means for comparing the receiving side input signal level with a predetermined threshold value. , if either the first or second echo cancellation effect is smaller than the estimated echo cancellation effect, it is determined that there is no near-end speaker; otherwise, it is determined that there is a near-end speaker. Based on the result of the determination means, when a near-end speaker is present, the identification of the echo path characteristics of the first echo canceller is stopped, and the identified echo path of the second echo canceller is cleared in a certain period of time. It consists of means.

Operation With this configuration, when the far-end speaker is busy and the near-end speaker is absent, the first echo canceller identifies the characteristics of the echo path and cancels the echo signal. When the far-end speaker is talking and the near-end speaker also speaks, or when the echo path changes, the first echo cancellation effect calculated by the first echo cancellation effect calculation means increases and The second echo cancellation effect calculated by the second echo cancellation effect calculation means detection means also increases, and when the first and second echo cancellation effects become larger than the estimated echo cancellation effects, the first and second comparison means The determining means is informed that the first and second echo canceling effects have become larger than the estimated echo canceling effect, and the determining means determines that a near-end speaker is present. The control means immediately stops identifying the echo path of the first echo canceller.

If a near-end speaker is present, the second echo canceller cannot identify the echo path due to the near-end speaker's voice, and
Therefore, the 'r4J constant means determines that there is a near-end speaker, and the control means determines that the Bl
, the identification of the echo path of the first echo canceller is stopped (~), and the characteristics of the identified echo path are not disturbed.

Further, the second echo canceller is controlled by the control means to clear the identified echo path characteristics in a certain period of time in order to quickly converge when the near-end speaker is no longer present. However, if this is a change in the echo path, the second echo canceller will gradually identify the echo path, and the second echo cancellation effect will decrease. When it becomes equal to the estimated echo canceling effect, the second comparison means detects this, the constant means determines that there is a variation in the echo path, and the control means starts identifying the echo path of the first echo canceller. In these two ways, fluctuations in the echo path and simultaneous bidirectional communication can be determined within the convergence time of the second echo canceller.

Further, a second echo canceler that can cancel echoes of the same length as the first echo canceler by using first and second filters and first and second thinning means,
The thinning ratio can be the same as the thinning ratio for the first echo canceller.

EXAMPLE An example of the present invention will be described below with reference to the drawings. FIG. 1 shows the configuration of an echo canceling device in one embodiment of the present invention. In FIG. 1, 1 is a receiving side input terminal, 2 is a receiving side output terminal, 3 is a transmitting side input terminal, 4 is a transmitting side output terminal, and 6 is a first echo canceller provided between the above terminals. The echo returning from the receiving side output terminal 2 to the transmitting side input terminal 3 is erased. The above configuration is the same as the configuration shown in FIG.

7 is reception fl that comes in from reception 11i11 input terminal 1
The first filter extracts a signal in the 1 kHz band from 6001+ of ll+input signal Rin 1, 8 is the input signal 6 of 5in1 on the transmitting side that comes in from the transmitter's input signal 5in1
9 is a first thinning means for extracting the signal in the band from 00Hz to 1kHz; 9 is a first thinning means for extracting the output signal of the first filter 7 at a rate of once every four samples; A second decimation means 6 takes out the output signal at a rate of once every four samples, and 6 uses the output signal Rin2 of the first decimation means as a receiving side input signal, and transmits the output signal S in2 of the second decimation means. 11 is the receiving side input signal R of the second echo canceller, which has the same delay time for echo cancellation as the first echo canceller.
in 1, transmitting side input/output signal Sin 1.5out 1
, the output signal 5in2 of the second thinning means, and the cancellation residual signal 5out2 of the second echo canceller are detected to detect the presence or absence of a near-end speaker and fluctuations in the echo path, and control the first and second echo cancellers. This is a two-way simultaneous call detection means.

FIG. 2 is a block diagram showing details of the two-way simultaneous call detection means. In FIG. 2, 21 is the receiving side input signal Ri
22 is a transmitting side input signal level detecting means for detecting the level of the transmitting side input signal S inl, and 23 is a transmitting side input signal level detecting means for detecting the level of the transmitting side output signal 5out1. Output signal level detection means, 24 is the output signal Sin of the second thinning means
Sub-band transmitting side input signal level detection means for detecting the level of 2; 25 is the cancellation residual signal 5o of the second echo canceller;
Sub-band transmitting side output signal level detection means for detecting the ut20 level, 26 is the receiving side input signal level Lr in 1
and a predetermined threshold Th, 27 is a ratio Acnc1 (hereinafter referred to as a transmission side output signal level LIiout 1) of the transmission side input signal level 1,5in1 and the transmission side output signal level Lsout 1 /transmission side input signal level 1,5in1 ratio Acnc1 is called the first echo cancellation effect); 28 is a calculation means for calculating the current wl from the past first echo cancellation effect;
Estimated echo cancellation effect calculation means 29 estimates the echo cancellation effect Acnc IA of 1, and 29 is the subband transmitting side input signal level Is+in2 and the subband transmitting side output signal level L s o
ut 2Acnc2 (hereinafter referred to as subband transmitting side output signal level Lsout2/subband transmitting side input signal level Lgin
2 is called a second echo cancellation effect); 30 is a first echo cancellation effect Acnc;
1.1 and the estimated echo cancellation effect Acnc IA, 31 is a third comparison means that compares the estimated echo cancellation effect ACnC1A and the second echo cancellation effect Acnc2, and 32 is the 1st comparison means for comparing the estimated echo cancellation effect Acnc IA. Second. Third comparison means 30, 31.
determination means for detecting the presence or absence of a near-end speaker and fluctuations in the echo path from the results Cmp1, Cmp2, and Cmp3 of step 26;
33 determines whether the identification of the echo path of the first echo canceller is continued or stopped based on the result No. 33 of the determination means 32;
This is a control means for controlling the retention and clearing of the echo path identified by the echo canceller.

The operation of the echo canceling device configured as described above will be explained below.

First, the voice signal of the far end speaker is the receiving side input signal Rin 1
The signal is supplied to the echo canceller 6 via the receiving side input terminal 1 as a signal. Further, the signal R1n1 is sent from the receiving side output terminal 2 to the near-end speaker as the receiving side output signal Rout1.

If there is no near-end speaker signal, the signal Rout1 passes through the echo path and enters the transmitter input signal 3 as the transmitter input signal 5in1. In the first echo canceller 6, the transmitting side output signal 5
A pseudo echo path is created from the sequence of the receiving side input signal R1n1 and the transmitting side output signal 5out1 so that out1 becomes zero.

Although the echo signal is canceled in this way, if a near-end speaker is present, the echo path cannot be estimated successfully.

Therefore, it is necessary to detect the presence of a near-end speaker signal and stop estimating the echo path.

This will be explained below with reference to FIG. The two-way simultaneous call detection means 11 detects the presence or absence of a near-end speaker as follows. When there is no near-end speaker signal and the echo path is constant, when the far-end speaker signal is sent as the receiving side input signal Rin 1, the receiving side input signal level detection means 21 detects the far-end speaker signal level Lr.
1n1 is detected and the far end speaker signal level becomes larger than a predetermined threshold Th in the third comparison means, the third comparison means outputs Cmp3=1 to the determination means 32 (pass 1).

Next, this far-end speaker signal returns via the echo path as the transmitting side input signal 5in1. Transmission side input signal Sin 1
In other words, the echo signal is detected by the transmission side input signal level detection means 22, the cancellation residual signal of the echo canceller is detected by the transmission side output signal level detection means 23, and the signal level is detected by the first echo cancellation effect calculation means 27. The echo cancellation effect Acnc1 by the No. 1 echo canceller is calculated. The estimated echo cancellation effect calculating means 28 compares the value obtained by adding a predetermined value to the estimated echo cancellation effect of one sample before and the first echo cancellation effect Acnc1, and adopts the smaller one as the estimated echo cancellation effect AcncIA. do. The first ratio/1 ratio means compares the first echo cancellation effect Acnc1 and the estimated echo cancellation effect AcncIA, and calculates the first echo cancellation effect Acnc1.
When is equal to the estimated echo cancellation effect AcncIA, Cmp1=
1 is output to the determining means 32. In the case of only the far-end speaker signal, the first echo canceller identifies the echo path so that the transmitting side output signal 5out1 becomes small, so the first echo cancellation effect Acnc1 becomes equal to the estimated echo cancellation effect AcncIA, and Cmp =1 is output from the first comparing means. The determination means f4j determines that there is no near-end speaker when Cmp1=1 and Cmp3=1 (bus 2).

When the near-end speaker speaks, the near-end speaker signal is added to the transmitting side input/output signal 5in1.5out1, so the first echo cancellation effect Acncl calculated by the first echo cancellation effect calculation means 27 The first comparison method is Cmpl
-0, and the second echo canceller continues to identify the echo path, but the cancellation residual output signal 5out 2 does not converge due to the near-end speaker signal, and the estimated echo cancellation effect Acnc
The second echo cancellation effect Acnc2 becomes larger than IA, and the output Qnp2 of the second comparing means becomes 0, and the determining means determines that there is a near-end speaker signal (bus 3).

In this case, in the control means 33, Cmpl=Cmp2−○
During the period, identification of the echo path of the first echo canceller is stopped, and the echo path identified by the second echo canceler is cleared for a certain period. If the echo path changes, the 1.
The echo path identified by the second echo canceller and the actual echo path! Because there is a difference in characteristics, 1. The cancellation residual output 5out2 of the second echo canceller increases, and the first echo canceller increases. The second echo cancellation effect becomes larger than the estimated echo cancellation effect, and the second echo cancellation effect becomes larger than the estimated echo cancellation effect. Output Cmp 1 = Cmp of the second comparison means 30.31
2-0.The quantification means 32 determines that there is a near-end speaker signal, but since the second echo canceller continues to identify the echo path, the second echo canceller will eventually converge. V second comparison means 31 in which the second echo cancellation effect Acnc2 is equal to the estimated echo cancellation effect ACnC1A
outputs Cmp2=1, the determination means determines that there is no near-end speaker signal (bus 4), and the control means 33 controls the first echo canceller to identify the echo path of the first echo canceller. do. In this way, fluctuations in the echo path and the utterance of the near-end speaker can be determined within the convergence time of the second echo canceller. The second echo canceller is only necessary to distinguish between fluctuations in the echo path and utterances of the near-end speaker, and is not required to cancel echo signals in the entire signal transmission band.

In this embodiment, the band for canceling echoes is set to 2 kHz or less, as shown in FIG. 4. As shown in FIG. 6, the audio signal has power in a relatively wide range, mainly from 30oH to 1kHz, so it is sufficient for the second echo canceller to cancel the echo signal in this band. Therefore, Figure 4 ('
As shown in B), the first. The passband of the second filters 8, 9 is 5oot (from z to 1kHz.) Next, the signals extracted by the first and second filters 8, 9 are divided into the first and second filters as shown in FIG. Since the band ratio of the input and output signals of the filters 8 and 9 is 4:1, the first and second thinning means perform the operation of extracting the signal only once every four samples. The output signal of the second thinning means is a signal with a band of 901 to 600 as shown in FIG.
First, because it will double. Even if the second echo canceller is capable of canceling echo signals with the same delay time, the hardware of the second echo canceller is
It only costs 1/1.

As described above, according to this embodiment, when there is no near-end speaker, the first echo canceller simulates the echo path and cancels the echo signal. When the near-end speaker or the echo path changes, the first echo cancellation effect of the first echo canceller calculated by the first echo cancellation effect calculation means decreases, and the first echo cancellation effect calculated by the estimated echo cancellation effect calculation means decreases. When the first comparison means detects this, the determination means immediately stops identifying the echo path of the first echo canceller, and the determination means further calculates the second echo cancellation effect. The second echo canceling effect of the second echo canceller calculated by the means is compared with the estimated echo canceling effect, and when the second echo canceling effect is equal to the estimated echo canceling effect, it is determined that the echo path has fluctuated, and the control means starts estimating the echo path of the first echo canceller. In cases other than the above, it is assumed that there is a two-way simultaneous conversation, and identification of the echo path of the first echo canceller remains stopped.

At this time, the second echo canceller clears the identified echo path at regular intervals by the control means.

The second echo canceller performs echo cancellation in the first echo canceller band.
．． The second filter limits the main band of the audio signal; By converting this band into a low-frequency signal by the second thinning means, the scale of the noise is small, and the fluctuation of the echo path and simultaneous two-way communication can be controlled at high speed (within the convergence time of the second echo canceller). ) can be determined.

In addition, in this embodiment, the band of the first echo canceller is 2.
kHz or less, but if the band is 4 kHz, if the band of the second echo canceller is from 500 to 1 kHz, the thinning rate by the thinning means may be set to once every eight times. In this way, if the band of the second echo canceller is set to an integer fraction of the band of the first echo canceller, the bands of the first and second echo cancellers may be changed as necessary.

Further, although the second echo canceller has the same delay amount as the first echo canceler, it may have a smaller delay amount than the first echo canceller. In this case, in the determination means, when comparing the first echo canceling effect and the second echo canceling effect, the deterioration of the echo canceling effect corresponding to the smaller delay amount of the second echo canceller is calculated as the second echo canceling effect. Just add it to the echo cancellation effect. By reducing the delay amount of the second echo canceller, the scale of the device can be further reduced, and the convergence speed of the second echo canceller becomes faster, making it easier to distinguish between simultaneous two-way communication and fluctuations in the echo path. Become faster.

Effects of the Invention As described above, in the present invention, when there is no near-end speaker, the first echo canceller simulates the echo path and cancels the echo signal. When the near-end speaker or the echo path changes, the first echo cancellation effect of the first echo canceller calculated by the first echo cancellation effect calculation means decreases, and the first echo cancellation effect calculated by the estimated echo cancellation effect calculation means decreases. When the first comparison means detects this, the determination means immediately stops identifying the echo path of the first echo canceller, and the determination means further calculates the second echo cancellation effect. The second echo canceling effect of the second echo canceller calculated by the means is compared with the estimated echo canceling effect, and when the second echo canceling effect is equal to the estimated echo canceling effect, it is determined that the echo path has fluctuated, and the control means starts estimating the echo path of the first echo canceller.

In cases other than the above, it is assumed that there is a two-way simultaneous conversation, and identification of the echo path of the first echo canceller remains stopped. At this time, the second echo canceller clears the identified echo path at regular intervals by the control means. The second echo canceller performs echo cancellation in the first band. The second filter limits the main band of the audio signal; By converting this band into a low-frequency signal by the second thinning means, the scale of the hardware can be reduced, and echo path fluctuations and simultaneous two-way communication can be controlled at high speed (within the convergence time of the second echo canceller). This allows the first echo canceller to always accurately simulate the echo path and perform excellent echo cancellation. In particular, even when an extremely large piece of hardware is required, such as an acoustic echo canceller, the scale of the second echo canceller can be made small, and it can quickly follow fluctuations in the echo path, allowing a large number of people to use the echo canceller. Excellent echo cancellation can be performed even in teleconference systems, and its practical effects are significant.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of an echo canceling device in an embodiment of the present invention, FIG. 2 is a block diagram showing the configuration of a two-way simultaneous call detection means in the same embodiment, and FIG. 3 is a two-way A flowchart explaining the algorithm of the simultaneous call detection means, FIG. 4 is a characteristic diagram showing the input and output bands of the first and second filters and the first and second thinning means, and FIG. 6 shows the input and output signals of the thinning means. FIG. 6 is a waveform diagram showing an example of this, FIG. 6 is a voice power distribution diagram, and FIG. 7 is a block diagram showing the configuration of a conventional echo canceling device. 1... Receiving side input terminal, 2... Receiving side output terminal, 3... Transmitting side input terminal, 4......
Transmission side output terminal, 5...first echo canceller, 6...second echo canceller, 7...
・First filter, 8...Second filter, 9
...First thinning means, 10...Second thinning means, 11...Two-way simultaneous call detection means,
21... Receiving side input signal level detection means, 22
...Transmission side input signal level detection means, 23...
...Transmission side output terminal level detection means, 24...
... Sub-band transmitting side input signal level detection means, 26...
...Subband transmitting side output signal level detection means, 26...
... Third comparison means, 27 ... First echo cancellation effect calculation means, 28 ... Estimated echo cancellation effect calculation means, 29 ... Second echo effect calculation means ,3o...
. . . first comparison means, 31 . . . second comparison means, 32 . . . determination means, 33 . . . control means. Name of agent: Patent attorney Toshio Nakao and 1 other Tadayoshi
111 3-! fi assistant 1 person 1 4 --- # Eiki If Figure 2 Figure 31. j+ 4th ['1 5th calculation 16L<l Yu-21 filter output 1τ5Lltt<
+s'\L2nA'1.51. ! jJtn! Book No. 711 No. G F-41Ir+ No. 71-1

Claims (6)

[Claims] (1) Identify the receiving side input terminal, the receiving side output terminal, the transmitting side input terminal, the transmitting side output terminal, and the characteristics of the echo path between the receiving side output terminal and the transmitting side input terminal, and the receiving side output a first echo canceller that cancels a reverberant signal that has returned from the terminal to the transmitter input terminal; a first filter that extracts only a specific band of the receiver input signal that is input from the receiver input terminal; a second filter that extracts only a specific band of a transmission-side input signal input from a transmission-side input terminal; first and second thinning means for thinning out output signals of the first and second filters; identifying the characteristics of the echo paths of the passbands of the first and second filters, using the output signal of the first thinning means as the receiving side input signal and the output signal of the second thinning means as the transmitting side input signal; a second echo canceller that cancels echo signals in the passbands of the first and second filters; a receiving side input signal, a transmitting side input/output signal, an output signal of the second thinning means; Upon receiving the residual signal of the second echo canceller which has eliminated the echo signal from the output signal of the thinning means, it detects that a near-end speaker's signal other than the echo signal is added to the transmitting side input signal, and cancels the near-end talk. stopping identification of the echo path of the first echo canceller when a signal from a person or the like is added;
It is characterized by comprising a two-way simultaneous conversation detection means for controlling the second echo canceller to clear the characteristics of the identified echo path in a certain period of time when a signal from a near-end speaker or the like is added. echo canceller. (2) The echo canceling device according to claim 1, wherein the first and second filters have a width that is an integer fraction of the band to be canceled by the first echo canceller. (3) The first and second thinning means thin out the signal at a ratio of the echo canceling band of the first echo canceller to the passband of the first and second filters. Echo canceling device as described in section. (4) The two-way simultaneous call detection means includes a receiving side input signal level detecting means for detecting the receiving side input signal level, a transmitting side input signal level detecting means for detecting the transmitting side input/output signal level, and a transmitting side output signal level. a detection means, a first echo cancellation effect calculation means for calculating a ratio of the transmission side input/output signal (transmission side output signal level/transmission side output signal level) which is a first echo cancellation effect, and a second thinning. sub-band transmitting-side input signal level detecting means for detecting the output signal level of the means; sub-band transmitting-side output signal level detecting means for detecting the residual signal level after the second echo canceler cancels the echo signal; The second echo cancellation effect is the ratio of the subband transmitting side input/output signal level (subband transmitting side input signal level/
a second echo cancellation effect calculation means for calculating the sub-band transmission side output signal level), and a second echo cancellation effect calculation means for calculating the current first echo cancellation effect from the past value of the first echo cancellation effect which is the output of the first echo cancellation effect calculation means. an estimated echo-cancelling effect calculating means for calculating an estimated value of the echo-cancelling effect; a first comparison means for comparing the actual first echo-cancelling effect with the estimated first echo-cancelling effect; a second comparing means for comparing the echo canceling effect of the second echo canceling effect with a second echo canceling effect; a third comparing means for comparing a predetermined threshold value and a receiving side input signal level; determining means for determining the presence or absence of a near-end speaker based on the results of the second and third comparing means; and identifying the echo path of the first echo canceller when the determining means determines that the near-end speaker is present. The echo canceling device according to claim 1, further comprising a control means for controlling the second echo canceller to clear the characteristics of the identified echo path in a certain period of time. . (5) The third comparing means determines that the receiving side input signal level has a level equal to or higher than the threshold value, and the first comparing means determines that the actual first echo canceling effect is the estimated first echo canceling effect. It is characterized in that it is determined that the near-end speaker does not exist when it is determined that the second echo cancellation effect is smaller than the estimated echo cancellation effect, or when the third comparing means determines that the second echo cancellation effect is smaller than the estimated echo cancellation effect. An echo canceling device according to claim 4. (6) When the determining means determines that there is no near-end speaker, the first
identifying the echo path of the first echo canceller, otherwise stopping the identification of the echo path of the first echo canceller, and determining means determines that a near-end speaker is present, and this period continues for a certain period of time or more. Claim 4, further comprising control means for controlling the first and second echo cancellers so as to clear the characteristics of the identified echo path of the second echo canceller when Echo canceller.