JP2000013284A - Echo canceller - Google Patents

Abstract

PROBLEM TO BE SOLVED: To effectively cancel a reception signal echo even when the characteristic of the echo invaded in an input signal is variously changed. SOLUTION: The echo canceller 1 is provided with an adder means 25 that adds a reception signal S1 at a prescribed rate to an input signal S3 to produce a reception signal added input signal S5, an adaptive filter 2 whose characteristic is switched, based on a transmission signal S4 to correct a characteristic of the reception signal S1 to produce a pseudo echo signal S2 from the received signal S1, an echo canceller means 3 that subtracts the pseudo echo signal S2 from the reception signal added input signal S5 to produce a transmission signal S4 and suppresses an echo signal component included in the transmission signal S4, and a control means that controls the adaptive filter 2 to stop the characteristic switching of the filter 2, based on the transmission signal S4 and the reception signal added input signal S5. Thus, even when the characteristic of the echo signal is variously changed, the echo canceller can effectively cancel the echo signal.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an echo canceller applied to, for example, a video conference system.

[0002]

2. Description of the Related Art Conventionally, in a two-way communication system such as a video conference system, an echo canceller device is used to maintain the communication quality in the two-way communication system.

First, an example of a conventional video conference system will be described with reference to a block diagram of FIG.

[0004] Reference numeral 4 denotes the entirety of the video conference system. The video conference system 4 is connected between the video conference system 4 and a video conference system 23 of the other party (referred to as a communication target in the following description). Send and receive audio and video signals, and also send audio and video
The system is constructed with a system that can transmit and receive drawing information and the like necessary during a conference by multiplexing signals.

[0005] A system control circuit 7 is provided in the video conference system 4. The system control circuit 7 also receives and transmits signals to and from the remote controller 16, and operates signals from the operation device 15 or the remote controller 16 itself. In response to each of the operation signals, the operation of the entire video conference system 4 is controlled to perform the operations described in the following items 1 to 7.

[0006] 1. The audio signal collected by the microphone device 10 is supplied to the echo canceller 11, which converts the sound signal into a digital signal in the echo canceller 11, and performs echo cancellation processing on the echo signal in which the vocal sound of the speaker device 13 enters the microphone device 10, and thereafter performs audio processing. Audio codec 12
Performs signal compression processing to generate a compressed audio signal.

[0007] 2. The conference attendee is photographed by the television camera device 5, and an image signal is generated and supplied to the video codec 6. The video codec 6 converts the image signal into a digital signal, compresses the signal, and then performs multiplexer / demultiplexing.
The compressed audio signal supplied from the audio codec 12 in the multiplexer / demultiplexer 8 and the control signal supplied from the system control circuit 7 are supplied to the demultiplexer 8 as a multiplexed data signal by the multiplexer / demultiplexer 8. To supply.

[0008] 3. Video conference system 4 and call target 23
And a line interface 14 provided for transmitting and receiving the multiplexed data signal to and from the communication target 23 through the communication line K1. 8 to transmit and receive a multiplexed data signal.

4. The multiplexed data signal transmitted from the communication target 23 to the line interface 14 of the video conference system 4 through the communication line K1 is supplied to the multiplexer / demultiplexer 8, and the multiplexer / demultiplexer 8 converts the multiplexed data signal into a compressed digital image signal and a compressed digital image signal. The compressed digital image signal is separated into an audio signal and a control signal, and the compressed digital image signal is supplied to a video codec 6 where data is expanded and converted into an analog signal to obtain a received image signal and output to the monitor device 9. Conference system 4
To monitor the image sent to

[0010] 5. The compressed audio signal is supplied to the audio codec 12, the signal is expanded, converted into a digital audio signal, supplied to the echo canceller 11, converted into an analog signal, and output to the speaker device 13, so that the audio signal on the call target side can be heard. .

6. The line drawing data written in the tablet 17 is fetched and displayed on the monitor device 9 through the addition circuit 18, the line drawing data is converted into digital data and output to the multiplexer / demultiplexer 8. And multiplexes the data with the signal of the second line and outputs it to the rear line interface 14.

7. The data received from the communication target 23 through the line interface 14 is supplied to the multiplexer / demultiplexer 8, the digital line drawing data is separated by the multiplexer / demultiplexer 8 and converted into the original analog line drawing data, and the analog line drawing data is added to the addition circuit. 18 to be added to the received image signal output from the video codec 6 and monitored by the monitor device 9.

Next, an example of the configuration of the echo canceller 11 will be described with reference to a circuit block diagram shown in FIG.

In FIG. 4, reference numeral 11 denotes the entire echo canceller.
Japanese name of aptive Digital Filter, in the following description, this adaptive filter is referred to as ADF) 2, subtractor 3, two-way simultaneous call detector 21, input terminal T1 of received sound signal S1 from call target 23, received sound It comprises an output terminal T2 for the signal S1, an input terminal T3 for the transmission voice signal S3 to be transmitted to the call target 23, and an output terminal T4 for the transmission voice signal S4. Note that ADF2 uses FIR (Finite Impulse Response).
) Consist of a filter.

Further, the two-way simultaneous call detector 21 is in a state where both callers speak together with the call target 23 in the video conference system 4, that is, in a state where the caller's voice signal is input to the terminal T1. , Terminal T3 is a detector provided to detect a state in which a speech input of a caller is input from the microphone device 10 (hereinafter, referred to as a double talk state).

The thus configured echo canceller 11
, The received voice signal S1 input from the input terminal T1
Is output from the output terminal T2. Then, the received voice signal S1 is supplied to the speaker device 13 and is output from the speaker device 13 as a sound to the caller on the TV conference system 4 side.

In a state where the output is present, the sound wraps around the microphone device 10 and transmits the transmission voice signal S3.
The transmission audio signal S4 including the echo signal generated therein is supplied to the + end side of the subtractor 3, and the output signal of the transmission audio signal S4 is monitored and the coefficient of the ADF 2 is switched while monitoring the output signal as shown by an arrow Y1. The received voice signal S1 is input to the ADF 2 in a state of being turned on, a pseudo echo signal S2 is generated by the ADF 2 based on the received voice signal S1, and the generated pseudo echo signal S2 is supplied to the minus end of the subtractor 3, and the subtractor In 3, the pseudo echo signal S2 is subtracted from the transmission voice signal S3,
The transmission audio signal S4 maintaining the speech quality by canceling the echo signal component is generated and output from the output terminal T4.

The echo signal component is a sound based on the received voice signal S1 supplied from the output terminal T2 to the speaker device 13 when the caller is not speaking toward the microphone device 10.
(Referred to as an echo signal component in the following description).

Further, in the state where the characteristics of the ADF 2 are variably controlled, the transmission voice signal S3 and the transmission voice signal S4 are input to the two-way simultaneous call detector 21, and the two-way simultaneous call detector 21 At the transmission voice signal S
3 and the transmission audio signal S4 are compared, and if the signal level of the transmission audio signal S4 exceeds a predetermined threshold value with respect to the signal level of the transmission audio signal S3, it is determined that a double talk state exists. The ADF 2 is controlled from the two-way simultaneous call detector 21 so that the characteristics of the ADF 2 do not change from the characteristics at the time of making this determination.

Under this control, the transmission voice signal S3 including the echo signal is supplied to the subtractor 3, and the pseudo echo signal S2 is subtracted from the signal S3 by the subtracter 3, based on the transmission voice signal S4 obtained. The characteristic of the ADF 2 is prohibited from being variably controlled, and in the double talk state, the S / N of the echo signal deteriorates due to the superposition of the voice signal input by the caller from the microphone device 10 on the echo signal. Therefore, the variable control Y1 of the characteristic of the ADF2 becomes inaccurate and the echo canceling operation becomes insufficient.
This prevents the remaining amount of the echo signal component in the transmission audio signal S4 from increasing and the communication quality of the transmission audio signal S4 from deteriorating.

Reference: JP-A-9-153848.

[0022]

However, in such an echo canceller device of a two-way communication system such as a telephone call between the video conference system 4 and the communication target 23, the location of the speaker device 13 and the microphone device 10 is set. Due to differences in sound transfer characteristics, differences in the audio output level of the speaker device 13, and differences in the input sensitivity settings of the microphone device 10, the echo characteristics such as the signal level and the echo length of the echo signal generated in the transmission voice signal S3 are wide. Varies over time.

Further, the speaker device 1 has a certain positional relationship.
3 and the microphone device 10 are set, and when the speaker device 13 and the microphone device 10 are set in another positional relationship different from the positional relationship, the speaker device 13 and the microphone device 10 are set in the transmission voice signal S3 according to the difference between these setting conditions. Even when the level of the generated echo signal is different, the signal level of the transmitted audio signal S4 may be substantially the same when compared with the signal level of the transmitted audio signal S4 regardless of the difference in the setting conditions. .

However, even if the level of the echo signal generated in the transmission voice signal S3 is different depending on the difference between the setting conditions, the threshold value of the transmission voice signal S3 may be changed. This threshold value needs to be set based on the higher ratio of the signal level of the transmission audio signal S4 to the signal level.

When the threshold value is set as described above, the double talk state may be detected more accurately if the threshold value is set to a lower level.
In this case, the detection performance of the double talk state deteriorates, and the variable control of the characteristics of the ADF 2 in the double talk state cannot be accurately prohibited in accordance with the double talk state. The echo cancellation operation becomes insufficient due to inaccuracy, and the residual amount of the echo signal component in the transmission audio signal S4 increases, resulting in a problem that the communication quality of the transmission audio signal S4 deteriorates.

In view of the foregoing, it is an object of the present invention to surely prevent the echo canceller device from deteriorating the echo suppression capability in the transmitted voice signal S4 in the double talk state.

[0027]

An echo canceller according to the present invention generates a pseudo echo signal from a received signal,
In the echo canceller device which suppresses the echo signal component in the transmission signal caused by the reception signal by subtracting the pseudo echo signal from the transmission signal generated from the input signal, the reception is performed at a predetermined ratio with respect to the input signal. Adding means for adding a signal to generate a received signal addition input signal; and an adaptive filter for adaptively switching characteristics based on the transmission signal to correct characteristics of the received signal and generating a pseudo echo signal from the received signal. An echo canceller for generating a transmission signal by subtracting the pseudo echo signal from the received signal added input signal and suppressing an echo signal component contained in the transmitted signal; And control means for stopping and controlling the characteristic switching operation of the adaptive filter based on the control signal.

According to the present invention, it is possible to reliably suppress an echo signal in an echo canceller device of a two-way communication system such as a video conference system.

[0029]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Referring to FIGS. 1 and 2, the same parts as those in FIGS. 3 and 4 are denoted by the same reference numerals, and detailed description thereof will be omitted. An example of the embodiment will be described.

In FIG. 1, reference numeral 1 denotes the entire echo canceller device, and the echo canceller device 1 has an input terminal T1,
It has an output terminal T2, an input terminal T3, and an output terminal T4, and has an ADF 2, a subtractor 3, a two-way simultaneous call detector 21,
It comprises an adder 25, a multiplier 26 and a second multiplier 27.

The input terminal T1 is an input terminal of a receiving voice signal S1 to which a voice signal from the communication target 23 is supplied through the line interface 14, the multiplexer / demultiplexer 8 and the audio codec 12, and the output terminal T2 is a receiving voice. The signal S1 is transmitted to the speaker device 1
3, an input terminal T3 is an input terminal of the transmission voice signal S3 from the microphone device 10, and an output terminal T4 is to the communication target 23 side through the audio codec 12, the multiplexer / demultiplexer 8, and the line interface 14. This is an output terminal for outputting the transmission audio signal S4.

Next, the operation of the echo canceller 1 will be described.

The transmission voice signal S3 supplied to the input terminal T3
Is supplied to the multiplier 26, the transmission voice signal S 3 is multiplied by the coefficient a in the multiplier 26, and supplied to one input side of the adder 25. On the other hand, the received voice signal S1 is
And the received voice signal S1 is multiplied by the coefficient b and supplied to the other input side of the adder 25. The transmitted voice signal S3 multiplied by the coefficient a in the adder 25 is multiplied by the coefficient b. A corrected audio signal S5 to which S1 is added is generated.

The received voice signal S input from the input terminal T1
1 is supplied to the output terminal T2. Then, the transmission audio signal S
While monitoring the output signal of Y4 as indicated by Y1, the AD in the state where the coefficient is switched so that the output signal level of the transmission audio signal S4 is maintained at the minimum value.
The received voice signal S1 is supplied to F2 to generate a pseudo echo signal S2 based on the received voice signal S1, and the pseudo echo signal S2 is supplied to the subtractor 3, where the corrected echo signal S5 is converted from the pseudo echo signal S2 to the subtractor 3. Is subtracted, and the transmission audio signal S4 output from the subtractor 3 is supplied to the output terminal T4 and the ADF 2 for monitoring the same.

The corrected voice signal S5 and the transmission voice signal S4 are input to the two-way simultaneous call detector 21, and the corrected voice signal S5
And the signal level of the transmitted voice signal S4 is compared. If the signal level of the transmitted voice signal S4 exceeds a predetermined threshold value with respect to the signal level of the signal corrected voice signal S5, Is determined to be in a double talk state in which both parties speak. Based on this determination, the two-way simultaneous call detector 21 controls the ADF 2 to monitor the output signal of the outgoing voice signal S4 and maintain the level of the outgoing voice signal S4 at the minimum value as shown by Y1. The characteristics of the ADF 2 in the state where the coefficient switching operation is being performed are not changed from the characteristics at the time when the two-way simultaneous call detector 21 makes this determination.

Next, as described with reference to FIG. 1, a signal obtained by multiplying the transmission voice signal S3 by the coefficient a and a signal obtained by multiplying the reception voice signal S1 by the coefficient b are added. A description will be given of an effect obtained by obtaining the corrected audio signal S5 and supplying the corrected audio signal S5 to the subtractor 3 to generate the transmission audio signal S4.

As an example, the coefficient a is 1 and the coefficient b is 1 /
2. When the level of the received voice signal S1 is L, the level of the echo signal in the transmitted voice signal S3 is L·Ｌ (when the level of the echo signal in the transmitted voice signal S3 is not double-talked). 1/8 of the level L of the received voice signal S1)
And the case where the level of the echo signal in the transmitted voice signal S3 is L · 1/32 (the level of the echo signal in the transmitted voice signal S3 is 1/32 of the level L of the received voice signal S1) Microphone device 1 to meet two conditions
Description will be made assuming that 0 and the speaker device 13 are set.

In the case of this example, a signal obtained by multiplying the transmitted voice signal S3 by the coefficient a and a signal obtained by multiplying the received voice signal S1 by the coefficient b are added to generate a corrected voice signal S5. Since the level of the echo signal in the transmission voice signal S3 is L · １ ／ and L · 1/32, the adder 25 adds a signal of a sufficiently large signal level. become.

That is, the signal level of the corrected voice signal S5 can be regarded as a signal level substantially equal to half the level L of the transmission voice signal S1, so that the echo signal level in the transmission voice signal S3 is L · 1. Even if the value fluctuates from / 8 to L · 1/32, it can be considered that the level of the corrected audio signal S5 is maintained at a substantially constant level which is hardly affected by this effect.

Therefore, even if the level of the echo signal in the transmission voice signal S3 fluctuates between L · １ ／ and L · 1/32, the threshold value of the bidirectional conversation detector 21 is fixed. The value can be adopted.

Next, when the coefficient a is 1, the coefficient b is 1/4, and the level of the received voice signal S1 is L1, the level of the echo signal in the transmitted voice signal S3 is L1 ・ 1/4.
0 is the condition 1 and the case where the level of the echo signal in the transmission voice signal S3 is L1 ・ 1/2 is the condition 2, and the data S4 in the conventional echo canceller shown in FIG.
/ S3 (S4 of transmission voice signal / data of transmission voice signal S3) and data S4 / S5 (example of transmission voice signal S4 / signal S5) in the embodiment of the echo canceller apparatus according to the present invention shown in FIG. 2) will be described with reference to FIG.

It should be noted that the data shown in FIG.
1 is data measured using a white noise signal.

As is apparent from FIG. 2, in the conventional echo canceller shown in FIG. 4, the level fluctuation of the transmission voice signal S3 between these conditions 1 and 2 is 13
In the case of dB, the data S4 / S3 obtained by dividing the transmission voice signal S4 by the transmission voice signal S3 has a level difference of 10 dB between the conditions 1 and 2, whereas FIG. In the example of the embodiment of the echo canceller apparatus according to the present invention shown in FIG. 1, in data S4 / S5 obtained by dividing S4 of the transmission audio signal by the correction audio signal S5, the difference between the condition 1 and the condition 2 is different. You can see that it stays at 3dB.

That is, even when the coefficient b is set to 1/4 in the embodiment of the echo canceller according to the present invention shown in FIG.
A fixed value can be adopted as this threshold value of 1.

According to this example, when the speaker device 13 and the microphone device 10 are set in a certain positional relationship,
The speaker device 1 has another positional relationship different from this positional relationship.
Even when the level of the echo signal generated in the transmission voice signal S3 changes due to the difference between these setting conditions, for example, when the microphone device 10 and the microphone device 10 are set, the threshold value is fixed and the double The state of the talk can be accurately detected. Therefore, the configuration of the two-way simultaneous call detector 21 can be simplified.

[0046]

According to the present invention, in an echo canceller device of a two-way communication system such as a video conference system, it is possible to reliably detect a double talk state even when the characteristics of an echo signal are variously changed. can do.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an example of an embodiment of an echo canceller device of the present invention.

FIG. 2 is a diagram showing measurement data of the echo canceller device shown and described in FIG. 1;

FIG. 3 is a block diagram showing an example of a configuration of a conventional video conference system.

FIG. 4 is a block diagram illustrating an example of a configuration of a conventional echo canceller device.

[Explanation of symbols]

2 ‥‥ adaptive filter (ADF), 3 ‥‥ subtractor, 5 ‥‥
Television camera device, 6 ‥‥ video codec,
7 system control circuit, 8 multiplexer / demultiplexer, 9 monitor device, 10 microphone device, 11 echo canceller, 12 audio codec, 13 speaker device, 18 adder circuit , 21 ‥‥ two-way simultaneous conversation state detector (two-way simultaneous call detector), 23 ‥‥ other party's video conference system (object of call), 25 ‥‥ adder, 26 ‥‥ multiplier, 27
{Second multiplier, S1} Received voice signal, S2} Pseudo echo signal, S3 {Transmitted voice signal, S4} Transmitted voice signal, S5} Corrected voice signal, Y1 Control of ADF2 Arrow indicating status

Claims (1)

[Claims] 1. A pseudo echo signal is generated from a received signal,
An echo canceller device configured to generate a transmission signal in which an echo signal component in the transmission signal caused by the reception signal is suppressed by subtracting the pseudo echo signal from a transmission signal generated from the input signal. Adding means for adding the reception signal at a predetermined ratio to generate a reception signal addition input signal; and a pseudo echo signal in which characteristics are switched adaptively based on the transmission signal to correct characteristics of the reception signal. An adaptive filter to generate, an echo canceller for generating a transmission signal by subtracting the pseudo echo signal from the reception signal addition input signal, and suppressing the echo signal component included in the transmission signal; Stop control of the characteristic switching operation in the adaptive filter based on an input signal and the transmission signal. An echo canceller device comprising a control unit.