JPH10229354A - Echo controller - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize a simultaneous two-way communication with simple configuration in an echo controller to prevent a receiving sound signal from being transmitted to a transmitting sound signal. SOLUTION: When only one of a receiving sound signal and a transmitting sound signal exists, the output of a transmission suppressing means 1 is selected by a selection output means 4. When both of the receiving sound signal and the transmitting sound signal exist simultaneously, a double talk state is detected by a double talk detecting means 3. Consequently, the output of an echo- removing means 2 is selected by the selection output means 4. Filter characteristics of an echo pass is simulated, a pseudo-echo signal is generated, based on the receiving sound signal, the pseudo-echo signal is subtracted from the transmitting sound signal by the echo-removing means 2. By using an adaptive digital filter with small number of taps is used for the echo-removing means 2, the configuration of the echo-removing means 2 is simplified.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an echo control device for preventing a received voice signal from wrapping around a transmitted voice signal, and more particularly to an echo control device applicable to a video conference system, a telephone having a hands-free function, and the like. About.

[0002] Video conference systems have been developed for holding conferences between remote locations. In addition, a telephone having a hands-free function that allows a user to speak without holding the telephone has been developed. In such a video conference system or telephone, the sound output from the speaker is reflected on the wall or ceiling of the room and goes around the microphone, so-called echo is generated, which is an obstacle to a comfortable conference or call.

An echo suppressor and an echo canceller have been proposed to suppress or eliminate such echoes.

[0004]

2. Description of the Related Art FIG. 7 is a diagram showing a configuration of a conventional echo suppressor. An echo suppressor 101 is provided on the transmitting side. This echo suppression unit 101 is constituted by a variable attenuation circuit. The voice detector 102 detects the voice level of the received voice signal on the receiving side. If the detected value is equal to or greater than a predetermined value, the echo suppression unit 10 is adaptively adjusted according to the magnitude of the detected value.
1 is increased. As a result, even if the received audio signal goes around to the transmitting side, generation of echo is suppressed.

[0005] The comparator 103 compares the audio level of the transmission audio signal on the transmission side with the audio level of the reception audio signal on the reception side, and determines that the audio level of the transmission audio signal is higher than the audio level of the reception audio signal. , The attenuation of the echo suppression unit 101 is set to zero. This makes it possible to send the transmission audio signal to the other party.

In this echo suppressor, if the sound on the receiving side is higher than the sound on the transmitting side, the sound on the receiving side is output. On the other hand, if the sound on the transmitting side is higher than the sound on the receiving side, the sound on the transmitting side is output. Sent to the side. There is no two-way simultaneous communication (double talk) state in which the voice of the transmitting side is transmitted to the other side at the same time as the voice of the receiving side is output.

[0007] It is assumed that the echo suppressor is also provided on the partner side. FIG. 8 is a diagram showing a configuration of a conventional echo canceller. Pseudo echo generator 104
Is constituted by an adaptive digital filter, and its tap coefficient value is determined so that the filter simulates an echo path. The received voice signal passes through the pseudo echo generating unit 104 to generate a pseudo echo signal, and the pseudo echo signal is sent to the echo canceling unit 105. Echo canceling unit 105
Is to cancel the echo component contained in the transmission voice signal using the pseudo echo signal.

[0008]

However, although the echo suppressor has a simple structure, there is a problem that two-way simultaneous communication cannot be performed essentially.

[0009] In the echo suppressor, there is a problem that the speech quality is low because the beginning and the end of the speech are cut off. In other words, if you speak out loud while the other party is speaking, the echo suppressor will select only the one with the higher audio level, so the audio level will be higher than the other party at the beginning of speaking. Until that time, your voice will not be transmitted to the other party. Also, the last part of the other party's story will not be transmitted.

[0010] On the other hand, the echo canceller is effective as a means for realizing simultaneous two-way communication, and of course, the beginning and end of the talk do not occur. However, the pseudo echo generator 104
, The configuration of the adaptive digital filter becomes large-scale, and the required amount of computation becomes enormous.

That is, in general, the length of the impulse response of an echo handled in an acoustic system is 200 msec or more, and the frequency band of an audio signal handled is 7 kHz. For example, if the sampling frequency of the audio signal is 16 kHz and the length of the impulse response of the echo is 256 msec,
The number of taps required for the adaptive digital filter is 4096
(= 16 × 10 ³ × 256 × 10 ^-3 ). If a learning identification method is used as an algorithm for adaptively controlling the echo canceller, the required amount of computation in the echo canceller is about twice the number of taps per sample, so that about 12
8 MOPS (MegaOperation Per Second) [= 16 × 1
0 ³ × (4092 × 2)]. In order to perform such calculations, a DSP (Di-
(Gital Signal Processor) requires nearly 10 chips. Therefore, the product configuration has become larger,
This would have a significant effect on product costs.

SUMMARY OF THE INVENTION The present invention has been made in view of such a point, and it is an object of the present invention to provide an echo control device which realizes two-way simultaneous communication with a simple configuration.

[0013]

According to the present invention, in order to achieve the above object, as shown in FIG. 1, a transmission suppressing means 1 for suppressing a transmission audio signal by detecting a reception audio signal, and a filter characteristic of an echo path are provided. An echo canceling means 2 for generating a pseudo echo signal based on the received voice signal and subtracting the pseudo echo signal from the transmitted voice signal, and a double talk system in which the received voice signal and the transmitted voice signal exist simultaneously. The double talk detecting means 3 for detecting the state, and when the double talk detecting means 3 does not detect the double talk state, the output of the transmission suppressing means 1 is selected and output, while the double talk detecting means 3 detects the double talk state. And a selection output unit 4 for selecting and outputting the output of the echo removal unit 2 when the operation is performed.

In the above configuration, when only one of the received voice signal and the transmitted voice signal exists, the output of the transmission suppressing means 1 is selected by the selection output means 4. That is, if a received voice signal is not detected and a transmitted voice signal is present, the transmitted voice signal is transmitted to the other party, while a received voice signal is detected and if no transmitted voice signal is present, The received voice signal is output to the own side.

Next, when both the received voice signal and the transmitted voice signal are present at the same time, the double talk detecting means 3
Detects the double talk state. Thus, the output of the echo removing means 2 is selected by the selection output means 4.
The echo removing unit 2 creates a pseudo echo path filter characteristic, creates a pseudo echo signal based on the received audio signal, and subtracts the pseudo echo signal from the transmission audio signal.

Thus, two-way simultaneous communication becomes possible.
The echo removing means 2 is composed of an adaptive digital filter, and the adaptive control of the tap coefficient value is operated only when the detection level of the received voice signal is higher than a predetermined value and the double talk is not detected. ing.
By the way, the time rate at which double talk is detected is
According to actual measurements by the applicant (five examples of conversations between men), the average is as low as 8.78%, and the S / N of the received voice signal and the transmitted voice signal at the time of double talk should not hinder even a relatively small value. Conceivable. Focusing on these points, it is possible to reduce the number of taps of the adaptive digital filter constituting the echo removing means 2. If an adaptive digital filter having a small number of taps is used for the echo removing means 2, the configuration of the echo removing means 2 is simplified, and the product cost can be reduced.

[0017]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing an embodiment of an echo control apparatus according to the present invention.

First, the principle configuration of the first embodiment of the present invention will be described with reference to FIG. In the first embodiment,
A transmission suppressing unit 1 for suppressing a transmission voice signal by detecting a reception voice signal, a filter characteristic of an echo path is pseudo-created, a pseudo echo signal is generated based on the reception voice signal, and the pseudo echo signal is generated from the transmission voice signal. Echo removing means 2 for subtracting a signal, double-talk detecting means 3 for detecting a double-talk state in which a received voice signal and a transmitting voice signal are simultaneously present, and transmission suppressing means when double-talk detecting means 3 does not detect a double-talk state 1 is selected and output. On the other hand, when the double talk detecting means 3 detects the double talk state, the output of the echo removing means 2 is selected and output.

In the above configuration, when only one of the received voice signal and the transmitted voice signal is present, the output of the transmission suppressing means 1 is selected by the selection output means 4. That is, if a received voice signal is not detected and a transmitted voice signal is present, the transmitted voice signal is transmitted to the other party, while a received voice signal is detected and if no transmitted voice signal is present, The received voice signal is output to the own side.

Next, if both the received voice signal and the transmitted voice signal are present simultaneously, the double talk detecting means 3
Detects the double talk state. Thus, the output of the echo removing means 2 is selected by the selection output means 4.
The echo removing unit 2 creates a pseudo echo path filter characteristic, creates a pseudo echo signal based on the received audio signal, and subtracts the pseudo echo signal from the transmission audio signal.

Thus, two-way simultaneous communication is possible.
The echo removing means 2 is composed of an adaptive digital filter, and the adaptive control of the tap coefficient value is operated only when the detection level of the received voice signal is higher than a predetermined value and the double talk is not detected. ing.
By the way, the time rate at which double talk is detected is
According to actual measurements by the applicant (five examples of conversations between men), the average is as low as 8.78%, and the S / N of the received voice signal and the transmitted voice signal at the time of double talk should not hinder even a relatively small value. Conceivable. Focusing on these points, it is possible to reduce the number of taps of the adaptive digital filter constituting the echo removing means 2. If an adaptive digital filter having a small number of taps is used for the echo removing means 2, the configuration of the echo removing means 2 is simplified, and the product cost can be reduced.

Next, the first embodiment will be described in detail. FIG. 2 is an overall configuration diagram of the first embodiment. FIG.
1 corresponds to the echo suppression unit 11 and the voice detector 12, and the echo removal unit 2 in FIG.
Correspond to the subtractor 13, the pseudo echo generation unit 14 and the voice detector 12, the double talk detection means 3 in FIG. 1 corresponds to the double talk detection unit 15, and the selection output means 4 in FIG. Corresponding.

The voice detector 12 detects the voice level of the received voice signal and sends the detected value to the echo suppressor 11 and the pseudo echo generator 14. The echo suppression unit 11 is configured by a variable attenuation circuit, and operates when the detected value of the sound level sent from the sound detector 12 is equal to or more than a predetermined value. In the operation, the attenuation increases as the detected value of the audio level increases. As a result, the transmission voice signal is suppressed, and the occurrence of echo is prevented.

The pseudo echo generator 14 is composed of an adaptive digital filter, and its detailed configuration will be described later with reference to FIG. The pseudo echo generation unit 1
4 is a voice detector 12 as described later with reference to FIG.
The adaptive operation is not performed if the detected value of the audio level sent from the device is equal to or less than a predetermined value, and the adaptive operation is performed if the detected value exceeds the predetermined value. However, this component is not shown in FIG.

The double talk detecting section 15 detects each audio level of the received audio signal and the transmitted audio signal. And
If both audio levels exceed a predetermined value, the switch 16 is caused to select the output of the subtractor 13, otherwise, the switch 16 is caused to select the output of the echo suppressor 11. Note that the double talk detection unit 15 may use a method in which the number of zero crossings of the signal is used in addition to the sound level instead of detecting the sound levels of the received sound signal and the transmitted sound signal.

FIG. 3 is a diagram showing the internal configuration of the pseudo echo generator 14 and the subtractor 13. Pseudo echo generation unit 14
Adaptive digital filter includes delay units 21a-21d
, Multipliers 22a to 22d, and adders 23a to 23c. The pseudo echo generation unit 14 further includes a tap coefficient value control unit 24. The tap coefficient value control unit 24
The value of each tap coefficient is determined based on the output of the subtractor 13,
These values are sent to the respective multipliers 22a to 22d.
The value of each tap coefficient is adaptively determined in a direction in which the received voice signal component included in the transmitted voice signal output from the subtractor 13 decreases.

Here, the adaptive digital filter of the pseudo echo generator 14 is used only during the two-way simultaneous communication period, so that the degree of echo removal of the adaptive digital filter may be low. More specifically, it is sufficient to set the length of the impulse response of the echo to a value of 1/10 (about 25 msec) or less and the number of taps to a value of 1/10 (about 400) or less. Practical results are obtained. For example, when the length of the impulse response is suppressed to 20 msec and the sampling frequency of the audio signal is 16 kHz, the number of taps is 320 (= 16 × 10 ³ × 20 × 10
^-3 ). Therefore, the required calculation amount is about 12 MOPS
[= 16 × 10 ³ × (320 × 2)], and the adaptive digital filter can be realized by a one-chip DSP.

FIG. 4 is a flowchart showing the operation of the echo control device having such a configuration. Hereinafter, description will be given along the step numbers in the figure. [S1] The double-talk detector 15 monitors the audio levels of the received audio signal and the transmitted audio signal, and if it detects two-way simultaneous communication, proceeds to step S7. If not, the process proceeds to step S2.

[S2] The switch unit 16 is connected to the upper side in FIG. That is, the output of the echo suppression unit 11 is selected by the switch unit 16. [S3] The pseudo echo generator 14 determines whether or not the detected value of the audio level sent from the audio detector 12 is equal to or less than a predetermined value. As a result, if the detected value of the audio level is equal to or less than the predetermined value, the process proceeds to step S6, and if it exceeds the predetermined value, the process proceeds to step S4.

[S4] The echo suppression unit 11 includes the detector 12
If the detected value of the audio level sent from the device is equal to or more than a predetermined value, the attenuation is changed according to the detected value of the audio level.
That is, the larger the detected value of the audio level, the larger the attenuation. When the detected value of the audio level sent from the detector 12 is smaller than a predetermined value, the echo suppression unit 11 stops operating.

[S5] The pseudo echo generator 14 performs adaptive correction of the tap coefficient value of the adaptive digital filter. That is, the tap coefficient value of the adaptive digital filter is adaptively corrected so that the output of the subtractor 13 becomes small.

[S6] The pseudo echo generator 14 stops the adaptive correction of the tap coefficient value of the adaptive digital filter. This is to prevent the pseudo echo generation unit 14 from operating with a small level of audio signal due to insufficient calculation accuracy and malfunctioning.

[S7] The pseudo echo generator 14 stops the adaptive correction of the tap coefficient value of the adaptive digital filter. That is, the tap coefficient value of the adaptive digital filter is adaptively modified so that the output of the subtractor 13 becomes smaller. As a result, the transmission signal itself is reduced. To prevent this, the adaptive correction is stopped.

[S8] The switch section 16 is connected to the lower side in FIG. That is, the output of the subtractor 13 is switched to the switch 1
Let 6 be selected. As described above, when it is determined that an audio signal is present on both the transmitting side and the receiving side, the echo canceller functions, and otherwise, the echo suppressor functions.

Next, a second embodiment will be described. FIG.
FIG. 3 is an overall configuration diagram of a second embodiment. In the second embodiment, the basic configuration is the same as the configuration of the first embodiment. Therefore, the same components are denoted by the same reference numerals, and the description thereof will be omitted. Only different portions will be described.

In the second embodiment, a smooth switching section 31 is used instead of the switch section 16 in the first embodiment.
Is provided. The smooth switching unit 31 performs the selection output smoothly over time. That is, immediately after the transition from the double talk state to the non-double talk state, the selected output is output from the subtraction unit 13 [this is x1 (t)
From the output of the echo suppression unit 11 [this is x2
(T)], and immediately after the transition from the non-double talk state to the double talk state, the selected output is subtracted from the output of the echo suppression unit 11 by the subtraction unit 1.
3 is gradually and smoothly changed to the output. Specifically, the smooth switching unit 31 obtains a value y (t) by performing an operation of the following equation (1), and outputs the value y (t).

[0037]

Y (t) = a · x1 (t) + (1−a) · x2 (t) (1) where the coefficient a is a function of time as shown in FIG. In the case of transition from the double talk state to the non-double talk state, the curve a1 in FIG. 6 is applied, and in the case of transition from the non-double talk state to the double talk state, the curve a2 in FIG. 6 is applied. .

Thus, the output can be smoothly switched. Although the coefficient a changes linearly in the example shown in FIG. 6, the coefficient a may not be linear as long as it changes smoothly.

[0039]

As described above, according to the present invention, the double talk detecting means and the selection output means are provided, and when the double talk detecting means detects the double talk state, the output of the echo removing means is selected by the selection output means. . The echo removing unit creates a filter characteristic of the echo path in a pseudo manner, creates a pseudo echo signal based on the received voice signal, and subtracts the pseudo echo signal from the transmission voice signal.

Thus, two-way simultaneous communication becomes possible.
Further, since the echo removing means is made to function only when double talk is detected, the number of taps of the adaptive digital filter constituting the echo removing means can be reduced. This simplifies the configuration of the echo removing means, and makes it possible to reduce the product cost.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is an overall configuration diagram of the first embodiment.

FIG. 3 is a diagram illustrating an internal configuration of a pseudo echo generation unit and a subtractor.

FIG. 4 is a flowchart showing the operation of the echo control device.

FIG. 5 is an overall configuration diagram of a second embodiment.

FIG. 6 is a graph showing a change curve of a coefficient a.

FIG. 7 is a diagram showing a configuration of a conventional echo suppressor.

FIG. 8 is a diagram showing a configuration of a conventional echo canceller.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Transmission suppression means 2 Echo removal means 3 Double talk detection means 4 Selection output means

Claims (5)

[Claims] An echo control device for preventing a received voice signal from wrapping around a transmitted voice signal, wherein a transmission suppressing means for suppressing a transmitted voice signal by detecting a received voice signal, and a filter characteristic of an echo path are created in a pseudo manner. An echo removing means for generating a pseudo echo signal based on a received voice signal and subtracting the pseudo echo signal from the transmitted voice signal; and a double talk state for detecting a double talk state in which the received voice signal and the transmitted voice signal exist simultaneously. Detecting means, when the double talk detecting means does not detect the double talk state, selects and outputs the output of the transmission suppressing means, while, when the double talk detecting means detects the double talk state, the echo removing means An echo control device, comprising: selection output means for selecting and outputting an output. 2. The apparatus according to claim 1, wherein said echo removing means has 40 taps.
2. The echo control device according to claim 1, wherein the echo control device comprises an adaptive digital filter of 0 or less. 3. The echo removing means according to claim 1, wherein said echo removing means comprises an adaptive digital filter for suppressing the length of an impulse response of the echo to 25 msec or less.
The described echo control device. 4. An apparatus according to claim 1, further comprising: an audio level detecting unit for detecting an audio level of the received audio signal, wherein said echo removing unit comprises an adaptive digital filter, and wherein the audio level detected by said audio level detecting unit is lower than a predetermined value. 2. The echo control device according to claim 1, wherein the tap coefficient value is adaptively corrected when the value is large and the double talk detecting means does not detect the double talk state. 5. The apparatus according to claim 1, wherein the selecting and outputting unit gradually switches from an output of the transmission suppressing unit to an output of the echo removing unit immediately after the double-talk detecting unit detects a double-talk state. 2. The echo control device according to claim 1, further comprising a switching unit that gradually switches from the output of the echo removing unit to the output of the transmission suppressing unit immediately after the detecting unit stops detecting the double talk state.