JPH0614100A - Echo canceller - Google Patents

Abstract

PURPOSE:To stably cancel an echo even when a talking state is changed by reducing the effect of noise. CONSTITUTION:A 1st FIR filter 520 revises a tap coefficient of a register 504 based on a residual signal e1 generated from a microphone output signal and a reception signal and generates a pseudo echo signal y1*. The tap coefficient of the register 504 is copied for a period once per 2XT2 period to set the tap coefficient as a tap coefficient of a register 503 of a 2nd FIR filter 530. The 2nd FIR filter 530 generates a pseudo echo signal y2* based on the tap coefficient of the register 503 and the reception signal. A comparator 509 selects the residual signal e1 or e2 which has a smaller power and outputs the selected signal as an echo elimination acoustic signal from a switch SW1 or SW2.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an echo canceller, which is applicable to, for example, a handsfree telephone.

[0002]

2. Description of the Related Art In recent years, acoustic echo cancellers have been actively researched and developed in consideration of application to hands-free telephones.

For example, the document: Japanese Patent Laid-Open No. 3-80628 discloses a double-talk detection method applicable to an acoustic echo canceller.

According to the above-mentioned document, for example, the microphone 102 is omnidirectional, and the microphone 103 is directional different from that of the microphone 102. Then, the acoustic power of the voice uttered by the near-end speaker is S2 in the microphone 102 and S3 in the microphone 103, and a difference in power occurs between S2 and S3. Then, the reception echo signal input level of the microphone 102 is R2, the reception echo signal input level of the microphone 103 is R3, and the microphone 1
When the noise input level in 02 is N2 and the noise input level in the microphone 103 is N3, the above level relationships are as follows.

R2 + N2> R3 + N3 (1) equation, S
3 + N3> S2 + N2 (2) formula, R2> R3
Expression (3), S3> S2 (4) Expression.

Under the above conditions, the power ratio Srec in the receiving state (receiving state) is Srec = (R3 + N3)
/ (R2 + N2). On the other hand, the power ratio Ssend in the transmitting state is Ssend considering the influence of noise.
(Sending state) = (S3 + N3) / (S2 + N2). The power ratio in the double talk (transmission / reception) state is Sdouble = (R3 + S3 + N3) / (R2 +
S2 + N2). And Ssend-Sdub
le = {R2 (S3 + N3) -R3 (S2 + N2)} /
It becomes {(S2 + N2) * (R2 + S2 + N2)}. Then, according to the equations (1) to (4), Ssend-
Sdouble> 0 and Srec-Sdouble <0. Therefore, Srec <Sdouble <Ssend
Becomes

From the above relationships, Srec and Sdoubl
A threshold is provided between e and Ssend to check which state the power ratio is closest to, and to identify each state.

[0008]

However, in the above method, even when the reception speaker signal output is not output from the speaker 104, the Srec state (listening state) is caused by the spatial noises N2 and N3. Therefore, there is a problem in that the filter coefficient of the adaptive filter 101 is updated to disturb the coefficient and stable echo cancellation cannot be performed.

There is also a problem that when the receiving state is changed to the transmitting / receiving state, the tap coefficient of the adaptive filter is disturbed and stable echo cancellation cannot be performed.

The present invention has been made in view of the above problems, and an object of the present invention is to reduce the influence of noise and to stably perform echo cancellation even if the call state changes. It is to provide an echo canceller.

[0011]

In order to achieve the above object, the present invention provides a receiving sound output means for outputting a receiving signal as sound, a sound capturing means for capturing sound and outputting an acoustic signal, and An adaptive filter that estimates a received sound echo component generated by an echo due to a received sound output, a calculator that removes the received sound echo component from the acoustic signal, a call state is determined, and the adaptive filter of the adaptive filter is determined according to this state. In an echo canceller that includes a call state determination unit that performs update control of filter coefficients and outputs an echo-removed acoustic signal in which the received acoustic echo component is removed from the acoustic signal, the following characteristic adaptive filters and It was realized by means and configuration.

That is, the echo canceling section including the adaptive filter and the arithmetic unit is multiplexed, and each echo canceling section outputs each echo-removing acoustic signal from the received signal and the acoustic signal to cancel each echo. The filter coefficient of the adaptive filter of the section is updated at least in the listening state (for example, the transmitting / receiving state may be included), and any one of the echo removal is performed according to the calling state obtained by the calling state determination. It is characterized in that the echo removal acoustic signal from the unit is selected and output.

The selection of the echo-removed acoustic signal is
It is preferable to perform the power comparison of the echo-removed acoustic signals from the echo-removal units.

[0014]

According to the present invention, since the echo canceling section is constructed in a multiplexed manner, and the adaptive filters of each echo canceling section are respectively updated at least by the echo canceling acoustic signals at different times in the receiving state, Outputs each echo removal acoustic signal output based on the echo component estimated by each echo removal unit in the receiving state and the transmitting / receiving state, performs power comparison, and outputs, for example, an echo removal acoustic signal with low power. can do.

For example, the first adaptive filter of the first echo canceling section adapts the filter coefficient by the echo canceling signal of the first echo canceling section when the receiving signal is present (in the receiving state or the transmitting / receiving state). Then, the second adaptive filter of the second echo removing unit, when the filter coefficient set in the first adaptive filter in the receiving state is fetched and set for its own filter coefficient in a certain cycle, for example, from the receiving state Even if the filter coefficient of the first adaptive filter is disturbed by the transmitted signal after changing to the transmitting / receiving state, the receiving acoustic echo with the filter coefficient set in the past stable receiving state by the second adaptive filter. The echo-removed sound signal obtained by estimating the component and the echo-removed sound signal of the first echo-removing section are compared in power, and the echo-removed sound with the smaller power is compared. It is also possible to output a signal.

Therefore, echo can be stably removed in the receiving state and the transmitting / receiving state.

[0017]

DESCRIPTION OF THE PREFERRED EMBODIMENTS A preferred embodiment of the echo canceller of the present invention will be described with reference to the drawings. FIG. 1 is a functional block diagram of an adaptive filter unit and adder applicable to an echo canceller. Before making this description, an overall outline of the echo canceller will be described with reference to FIG.

FIG. 3 is a functional block diagram of the echo canceller of this embodiment.

In FIG. 3, the echo canceller comprises an omnidirectional microphone 1, a speaker 2, and an echo canceller main body 7. And
The output of the microphone 1 is supplied to the echo canceller main body 7. The echo canceller main body 7 outputs the residual signal e as an eco canceller output. Further, the received signal is supplied to the echo canceller main body 7 and the speaker 2.

Further, the echo canceller main body 7 has an adder 3, an adaptive filter section 4, a microphone output signal power detector 5, a tap coefficient update control section 6, a residual signal power detector 8, and a power ratio. It is composed of a calculator 9, a received signal power detector 10, and a call state determiner 11.

Next, the function of each unit will be described. Adder 3
Outputs the residual signal e by taking the difference between the transmission signal supplied from the microphone 1 and the pseudo echo signal y ^* supplied from the adaptive filter unit 4, and outputs the residual signal e together with the echo coefficient canceller output control unit 6 and the tap coefficient update control unit 6. It is also supplied to the differential power detector 8.

The adaptive filter unit 4 takes in the received signal and the transmitted signal from the microphone 1, and generates and adds the pseudo echo signal y ^* based on the control of the tap coefficient control signal from the tap coefficient update control unit 6. To the differential input of the instrument 3. A specific configuration example of the adaptive filter unit 4 will be described in detail in the description of FIG. 3 below.

Microphone output signal power detector 5
Calculates the average power of the microphone output signals provided. In this averaging, a short cycle t (for example, 1
Average value Pt1 _ave of 0 msec) and the long period T
Average value P _T1ave (for example, about 100 msec)
And calculate. In the averaging based on these two types of time, the accuracy of signal detection is improved by setting the cycle so that there is a difference in the change of the average value with respect to the time variation of the microphone output signal. Then, the average value Pt1 _ave and the average value P
The ratio Pp1 to _T1ave is calculated. This power ratio Pp1
Is greater than a predetermined threshold Pth1. If it is determined that the power ratio Pp1 is larger than the predetermined threshold Pth1, it is considered that there is a microphone output signal,
Information on the presence of the microphone output signal
Supply to. Also, the averaged microphone output signal power value P1 is supplied to the power ratio calculation unit 9.

The residual signal power detector 8 has basically the same function as the microphone output signal power detector 5. That is, the average value of the powers of the residual signals supplied is calculated. In this averaging, a short cycle t (for example, 10 m
An average value Pt2 _{ave of} about sec) and an average value P _{T2ave of} the long cycle T (for example, about 100 msec) are calculated. In the averaging based on these two types of time, the accuracy of signal detection is improved by setting the cycle so that there is a difference in the change of the average value with respect to the time variation of the residual signal. Then, the power ratio Pp2 between the average value Pt2 _ave and the average value P _T2ave is obtained. This power ratio Pp2 is the predetermined threshold Pth.
It is determined whether it is greater than 2. Here, the power ratio Pp
If it is determined that 2 is larger than the predetermined threshold value Pth2, it is considered that there is a residual signal, and the residual signal presence information is supplied to the call state determining unit 11. Then, the averaged residual signal power value P2 is supplied to the power ratio calculation unit 9.

The reception signal power detector 10 has basically the same functions as the microphone output signal power detector 5 and the residual signal power detector 8. That is, the average value of the powers of the residual signals supplied is calculated. This averaging, short-period t (e.g., about 10 msec) for calculating the average value _{Pt3 ave} of the long period T (e.g., about 100 msec) and a mean value _P T _{3 ave} of. In these two types of averaging, the period is set so that there is a difference in the change of the average value with respect to the time fluctuation of the received signal, and the accuracy of signal detection is improved. Then, the average value _{Pt3 ave} , the average value P _{T3a ve,} and the ratio Pp3 are obtained. It is determined whether or not this power ratio Pp3 is larger than a predetermined threshold value Pth3. Here, if it is determined that the power ratio Pp3 is larger than the predetermined threshold value Pth3, it is considered that there is a reception signal, and the reception signal presence information is supplied to the call state determination unit 11.

The power ratio calculation unit 9 acquires the transmission signal power value P1 from the transmission signal power detector 5 and the residual signal power value P2 from the residual power detector 8 to obtain the echo cancellation amount Ys. Specifically, this echo cancellation amount Y
s is, for example, Ys = 10 × log ₁₀ (P1 / P2)
Then, the echo cancellation amount Ys is supplied to the call state determiner 11. For example, when there is only a received signal, the echo cancellation amount Ys changes so as to take a value near a certain steady value V1. For example, the curve shown in FIG. 4 can represent Ys.

The call state judging device 11 fetches the microphone signal presence / absence information from the microphone signal power detector 5, the residual signal presence / absence information from the residual signal power detector 8, and the reception signal power detector 10 Information on whether or not a received signal is present is fetched from the device, a coefficient update / stop control signal for controlling whether or not to perform filter tap coefficient update control is generated based on these information, and is supplied to the tap coefficient update control unit 6. . In addition, the echo cancellation amount Ys supplied from the power ratio calculation unit 9 is fetched, and a constant value V
Until the value becomes 1, the coefficient updatable control signal is supplied to the tap coefficient update controller 6. FIG. 4 shows a characteristic diagram of an example of the echo cancellation amount Ys.

Further, the call state judging device 11 is, for example,
When it is determined that the reception state is achieved by the information that the reception signal is present, the coefficient update control signal is supplied to the tap coefficient update control unit 6. When it is determined that there are a microphone signal, a reception signal, and a residual signal (in the double talk state, the two-way communication state, and the transmission / reception state), the coefficient update stop control is supplied to the tap coefficient update control unit 6. To do. If it is determined that there is a microphone signal and a residual signal, the coefficient update stop signal is supplied to the tap coefficient update control unit 6. When it is determined that there is neither a transmission signal nor a reception signal, the coefficient update stop signal is supplied to the tap coefficient update control unit 6. In addition, the information with / without microphone signal,
Microphone output signal power detector 5, residual signal power detector 8, received signal power detector 10 so that other combinations of the received signal presence / absence information and the residual signal presence / absence information cannot occur. It is assumed that the respective thresholds of are set. Therefore, states other than the above combinations cannot occur in this embodiment.

The tap coefficient update control unit 6 basically takes in the residual signal e output from the adder 3, and the call state judging unit 1
The tap coefficient control signal is supplied to the adaptive filter unit 4 so that the residual signal e is minimized under the control of the coefficient update / stop control signal supplied from 1. This tap coefficient control algorithm can be controlled by, for example, a learning identification method.

Therefore, from the start-up of the initial state until the echo cancellation amount Ys takes a constant value V1, the above-mentioned control algorithm performs sequential processing to update the tap coefficient. Then, the method of updating the tap coefficient after the echo cancellation amount Ys comes to take a value near the constant value V1 will be described later with reference to FIGS. 5 and 6.

FIG. 5 is an explanatory diagram showing the time relationship between the change of the echo cancellation amount and the detection of the microphone output signal in this embodiment.

In FIG. 5, (A) shows a curve of an example of the temporal change of the echo cancellation amount in the power ratio calculating section 9. (B) shows an example of a waveform diagram of a microphone output signal from the microphone 1 in a relationship corresponding to the curve of (A). (C) is a diagram showing the presence / absence of a signal in which the microphone output signal power detector 5 detects sound / silence corresponding to the output signal of the microphone 1 in (B). Here, the time from when the microphone 1 starts to output a microphone output signal due to the capture of voice or the like (B in FIG. 5B) to the time when the presence of a signal is determined and output (A in FIG. 5C). To T
1 (detection response time T1).

FIG. 1 is a functional block diagram of an example of the adaptive filter unit 4 and the adder 3 according to this embodiment.

In FIG. 1, the adder 3 is the same as that shown in FIG.
It shows the same part as the adder 3 in the inside, and specifically, the adder 3 is composed of adders 507 and 508. Parts other than the adder 3 represent internal functional blocks of the adaptive filter unit 4. The adaptive filter unit 4 includes tap coefficient multiplication registers 503 and 504,
A reference signal register 505, adders 501 and 502,
It is composed of a switch 506, an average power comparator 509, and switches SW1, SW2, SW3, and SW4.

Then, the second FI which comprises the reference signal register 505, the switch 506, the tap coefficient multiplication register 503, and the adder 501 of the adaptive filter unit 4.
The R-type filter 530 produces a pseudo echo y2 ^* . Further, a pseudo echo y1 ^* is generated by the first FIR type filter 520 composed of the reference signal register 505, the tap coefficient multiplication register 504 and the adder 502.

Therefore, the reference signal register 505 is shared by the second FIR type filter 530 and the first FIR type filter 520.

As for the time lag (T) register length of the reference signal register 505, the time lag registers 505a to 505n can hold sample data at least twice as long as the above T1 time (detection response time) (T2 ≧ 2 × T1). And In the adaptive filter unit 4, the T1 time is the time from when the call state determiner 11 determines that the microphone output signal is the double talk or when the call state determiner 11 determines the transmission state.
1 is set.

For example, assuming that the sampling frequency is fHz (for example, about 8 kHz), the reference signal register 50
5 is f × T2 (for example, 8 kHz × 0.05 sec)
= 400 or so) time lag registers 505a to 50
It is configured to include 5n. Then, the above times T1 and T2
An example of the relationship of is also shown in FIG.

The switch 506 forming a part of the second FIR type filter 530 is on / off controlled by the control signal A1 from the tap coefficient update control unit 6. The supply timing of the control signal A1 is, for example, TP of FIG. 6D in the operation timing chart of FIG.
1, TP2 (every 2 × T2 time in the receiving state).

Every time the switch 506 is turned on once in the T2 period, the tap coefficient at that time is read from the tap coefficient register 504 of the first FIR type filter 520, and the tap coefficient of the second FIR type filter 530 is read. It is transferred to the tap coefficient multiplication register 503 and set. Then, while the switch 506 is turned on, the reception signal from the reference signal register 505 is supplied to the tap coefficient multiplication register 503.
Is taken in and multiplied by the set tap coefficient,
The pseudo echo signal y2 ^* is generated by addition in the adder 501.

Then, the generated pseudo echo signal y2 ^*
Is supplied to the difference input of the adder 507, where the difference from the signal from the microphone 1 (speech signal S + echo signal y) is calculated, the residual signal e2 is output, and the average power comparator 509 , Switch SW1.

On the other hand, the first FIR type filter 520 is
While the control signal A1 is on-controlled (in the receiving state) from the tap coefficient update control unit 6 to the switch SW3, the value of the residual signal e1 generated by the supply of the echo signal from the microphone 1 while being supplied is sequentially applied. The coefficient value of the tap coefficient register 504 is updated, and the pseudo echo signal y1 ^* generated with the update of this coefficient value also changes. Therefore,
When transmission is suddenly started in the receiving state, the adder 50
The 8-output residual signal e1 is disturbed, and the disturbed residual signal e1 is
The tap coefficient of the tap coefficient multiplication register 504 is updated by this, the pseudo echo signal y1 ^* is also disturbed, the power of the residual signal e1 increases, and the average power comparator 509
Is determined to have a smaller power than the residual signal e2 due to the pseudo echo y2 ^* which is set in the past reception state of the second FIR type filter 530 and is generated by the tap coefficient of the tap coefficient multiplication register 503. Then, the residual signal e2 is output as an echo canceller output.

The average power comparator 509 has a predetermined threshold value S1 set in advance and is supplied with the residual signals e1 and e.
2 and S1 described above, the residual signal e1 or e2 having the smaller residual echo component is determined and selected. In order to make this judgment, in the case of the power relation of (e2) ² + s1 <(e1) ² , in order to select the residual signal e2,
The ON control signal is supplied to the switch SW1 to turn it on.
Further, an OFF control signal is supplied to the switch SW2 to turn it off, and the tap coefficient update control unit 6 is switched to the switch 506.
A control signal A3 for turning on is supplied. Upon receiving this control signal, the tap coefficient update control unit 6 switches the switch 506.
To turn on the control signal A2.

Further, the average power comparator 509 determines (e2) ² in order to judge and select the residual signal e1 or e2.
When it is determined that the power relationship is + s1 ≧ (e1) ² , the switch SW is used to select the residual signal e1.
An OFF control signal is supplied to the switch SW1 to turn it OFF, an ON control signal is supplied to the switch SW2 to turn it ON, and the tap coefficient update control unit 6 is switched to the switch 506.
A control signal A3 for turning off is supplied. Upon receiving this control signal, the tap coefficient update control unit 6 switches the switch 506.
Is supplied with the ON control signal A2 to turn it on, and the switch SW3 is supplied with the ON control signal A1 to turn it on. Then, by turning on the switch SW3, each coefficient of the tap coefficient multiplication register 504 is updated by the residual signal e1.

FIG. 6 is an operation timing chart of this embodiment.

In FIG. 6, (A) shows a transition from the receiving state to the double talk state to the transmitting state to the silent state as an example of the state transition. (B) represents the detection result of the presence or absence of the microphone output signal in the microphone output signal power detector 5. (C) represents the presence / absence detection result of the reception signal in the reception signal power detector 10.

In FIG. 6D, the switch 506 of the second FIR type filter 530 of the adaptive filter section 4 is turned on / off.
The timing of the off control is shown. Switch 506
Is controlled to be turned on by the control signal A2 from the tap coefficient update control unit 6 in the transmission / reception state (double talk state). In the states other than the transmission / reception state (double talk state), the off control is performed.

FIG. 6 (E) shows the ON / OFF control timing of the switch SW1 of the adaptive filter section 4. When it is determined that the average power of the residual signal e2 is smaller than the average power of the residual signal e1 in the double talk state, the switch SW1 is on-controlled and outputs the residual signal e2, for example. When the average power of the residual signal e1 is smaller, the switch SW2 is turned on.

FIG. 6F shows the on / off control timing of the switch SW2 of the adaptive filter section 4.
That is, for example, in the listening state, the switch SW2
Is turned on to output the residual signal e1 as an echo canceller output. Further, even in the transmitting state and the silent state, the switch SW2 is ON-controlled to output the residual signal e1 or the silent signal.

FIG. 6G shows the on / off control timing of the switch SW3 of the adaptive filter section 4.
That is, since the SW3 is ON-controlled by the control signal A1 from the tap coefficient update control unit 6 while the reception signal is being supplied to the echo canceller main body 7, it is ON-controlled in the reception state and the double talk. Then, for example, when SW3 is turned on in the receiving state, the tap coefficient of the tap coefficient multiplication register 504 of the first FIR filter 520 is controlled to be updated by the residual signal e1. Also in the double talk state, SW3 is on-controlled, and the tap coefficient of the tap coefficient multiplication register 504 is updated by the residual signal e1.

FIG. 6H shows the ON / OFF control timing of the switch SW4 of the adaptive filter section 4.
The switch SW4 is ON-controlled by the control signal A4 from the tap coefficient update control unit 6 in the receiving state and the transmitting / receiving state. That is, the ON control is performed when there is a reception signal (or echo signal). Further, the OFF control is performed in the transmitting state and the silent state. That is, when there is no reception signal (or echo signal), the off control is performed.

FIG. 6J shows the tap coefficient copy transfer timing from the tap coefficient register 504 of the first FIR type filter 520 of the adaptive filter unit 4 to the tap coefficient multiplication register 503 of the second FIR type filter 530. It represents. Then, 1 in 2 × T2 period in the listening state
Every time (TP1, TP2), the tap coefficient of the tap coefficient register 504 of the first FIR type filter 520 is changed to the second tap coefficient.
FIR type filter 530 of tap coefficient multiplication register 5
03 to transfer and set. This is the first
When the tap coefficient of the FIR type filter 520 is disturbed (for example, when the tap coefficient is disturbed when the state is changed from the receiving state to the transmitting / receiving state), the effect of the tap coefficient register of the second FIR type filter 530 is This is for making it difficult to reach 503.

According to the above-described embodiment, since the echo cancellation is performed by using the single omnidirectional microphone 1, it is difficult to be affected by the location of the noise source,
Even if there are multiple speakers, double-talk detection accuracy is less likely to deteriorate.

Further, from the presence / absence information of the microphone output signal, the presence / absence information of the reception signal, and the presence / absence information of the residual signal, whether it is the reception state, the transmission state, the transmission / reception state, or the silent state. Since it is determined that the tap coefficient of the adaptive filter unit 4 is inappropriately updated due to erroneous detection of the call state due to transmission noise or the like when there is no received signal (eg, echo signal). You can

As shown in FIG. 3, the adaptive filter unit 4 has a first FIR type filter 520 and a second FIR type filter 520.
Type filter 530, the
The second FIR type filter 53 at a cycle of once every × T2 period.
Copy and transfer the tap coefficient of 0 and set it in the tap coefficient multiplication register 503 of the second FIR type filter 530,
Even if the tap coefficient of the first FIR type filter 520 is disturbed when the receiving state changes to the double talk state, the pseudo echo y2 ^* is set by using the tap coefficient set in the second FIR type filter 530 ^. Can be used to perform stable echo cancellation.

Therefore, it is effective when applied to a hands-free telephone or the like.

In the above embodiment, the microphone output signal power detector 5, the residual signal power detector 8,
Short cycle t of the reception signal power detector 10 = 10 msec,
The long period T is 100 msec, but the present invention is not limited to this.

Further, in the above embodiment, the detection of the power change of the microphone output signal power detector 5, the residual signal power detector 8 and the reception signal power detector 10 is detected by the predetermined threshold value. The threshold value may be different for each power detector 5, 8, 10.

Further, in the above-mentioned embodiment, the configuration of the adaptive filter section 4 is as shown in FIG. 3, but the configuration is not limited to this. For example, although the FIR type filter is used, another IIR type filter or the like may be used. Also, the sampling frequency is 8k
However, the reference signal register length T2 is set to 0.05 sec, but the present invention is not limited to this.

Further, in the above embodiment, the learning method is used as an example in the tap coefficient update control section 6, but the invention is not limited to this.

In the above embodiment, the echo canceling section (adaptive filters 520 and 530, and the adder 50) are used.
7, 508) shows an example in which it is duplicated, but the present invention is not limited to this. In other words, it can be applied even to a configuration in which the number of layers is three or more.

Further, in the above embodiment, the microphone output signal power detector 5 and the residual signal power detector 8 in FIG. 3 detect the presence or absence of the signal based on the detection of the power, but the present invention is not limited to this. Absent. In addition to this, for example, a detection method using autocorrelation may be used.

[0063]

As described above, according to the present invention, the echo removing section is configured to be multiplexed, and each echo removing section outputs each echo removing acoustic signal from the reception signal and the acoustic signal,
The filter coefficient of the adaptive filter of each echo removing unit is
It is configured to be updated at least in the listening state, and to select and output the echo-removed acoustic signal from any of the echo removing units according to the call state obtained by the call state determination. , It is possible to reduce the influence of noise and perform stable echo cancellation even if the call state changes.

[Brief description of drawings]

FIG. 1 is a functional block diagram of an adaptive filter unit and an adder of an echo canceller according to an embodiment of the present invention.

FIG. 2 is a functional block diagram of a conventional acoustic echo canceller.

FIG. 3 is a functional block diagram of an echo canceller according to an embodiment.

FIG. 4 is a characteristic diagram of an echo cancellation amount according to an embodiment.

FIG. 5 is an operation explanatory diagram of the embodiment.

FIG. 6 is an operation timing chart of one embodiment.

[Explanation of symbols]

501, 502, 507, 508 ... Adder, 503, 5
04 ... Tap coefficient multiplication register, 505 ... Reference signal register, 509 ... Average power comparator, 520 ... First FI
R type filter, 530 ... Second FIR type filter, SW
1 to SW4 ... Switches.

Claims (2)

[Claims] 1. A reception sound output means for outputting a reception signal as sound, a sound capturing means for capturing sound and outputting a sound signal, and an adaptive filter for estimating a reception sound echo component generated by an echo due to the reception sound output. A computing unit that removes the received acoustic echo component from the acoustic signal, and a call state determination unit that determines a call state and controls updating of the filter coefficient of the adaptive filter according to the state, In an echo canceller that outputs an echo-removed acoustic signal in which the received acoustic echo component has been removed from the acoustic signal, an echo removal unit consisting of the adaptive filter and the computing unit is configured in a multiplexed manner, and each echo removal unit operates as a received signal. Output each echo removal acoustic signal from the acoustic signal, the filter coefficient of the adaptive filter of each echo removal unit,
Update at least when in the listening state, and depending on the call state obtained by the above call state judgment,
An echo canceller, which selects and outputs the echo removal acoustic signal from any one of the echo removal units. 2. The echo canceller according to claim 1, wherein the echo canceling acoustic signal is selected by comparing the powers of the echo canceling acoustic signals from the echo canceling units.