JPH0865211A - Echo canceller - Google Patents

Abstract

(57) [Abstract] [Purpose] By performing echo cancellation on the low-frequency component of the transmission input signal, the processing amount is reduced, and the size and cost of the device are reduced. A transmission input signal is divided into a high frequency component and a low frequency component by a band divider 12. Audio signals have an energy center of gravity in the low frequency range, and high frequencies have large air attenuation and little wraparound to the microphone. Furthermore, the lower the frequency is, the lower the autocorrelation of the signal is. Therefore, when the echo cancellation is applied to the low frequency component, the effect is high. For this reason,
The echo canceling unit 14 performs echo canceling processing only on the low frequency component, and guides the high frequency component as it is to the band integrator 15 without any processing.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an echo canceller, and more particularly to an echo canceller for canceling an acoustic echo of a device such as a video conference terminal that performs a two-way simultaneous call using a speaker and a microphone.

[0002]

2. Description of the Related Art FIG. 3 is a diagram for explaining the operation of a general echo canceller. A received signal is output from a speaker 10, a part of which goes around to a microphone 11 via an echo path. Add to the transmit input signal. This is an echo, and the echo canceller 9 attenuates this echo. The echo canceller 9 estimates the impulse response of the echo path by the learning identification method or the like, and adds the pseudo echo generated by convolving the impulse response to the reception signal in the opposite phase to the transmission input signal to attenuate the echo.

Now, in order to realize a wideband echo canceller for ensuring good speech quality, a very high speed operation is required due to the high sampling frequency.
Large-scale hardware is required, which prevents miniaturization and low price. By dividing the target signal into bands and performing echo canceller processing for each band, the effective sampling frequency can be halved, so this method is used to reduce the amount of processing. It is a split-type echo canceller.

[0004]

In the conventional band-division type echo canceller, the filter operation is halved, but the other processing remains unchanged. Considering the increase in the amount of additional processing such as band-division, the processing is so large. The amount cannot be reduced. The present invention has been made in view of the actual situation as described above, and has been made for the purpose of further reducing the processing amount while minimizing performance deterioration.

[0005]

In order to solve the above problems, the present invention provides (1) a band divider for dividing a transmission input signal into a low frequency component and a high frequency component, and the low frequency component and the high frequency component. It has a band integrator for synthesizing the components and an echo canceling unit for canceling the echo, which is a part of the reception output signal wrapping around to the transmission input signal, by the pseudo echo of the opposite phase, and echo cancels only the low frequency component of the transmission input signal And (2) a transmission determination unit that determines whether the transmission input signal is a signal to be transmitted, and when the transmission determination unit determines that transmission is unnecessary, Attenuating the high frequency component of the transmission input signal, or (3) the transmission determination unit determines that transmission is unnecessary when the power level of the transmission input signal is below a certain value, or (4) The transmission determination unit receives If the power level of the input signal is equal to or higher than a certain value, it is determined that the transmission is not necessary, or (5) if the correlation level between the received output signal and the transmission input signal is equal to or more than a certain value, the transmission determination unit determines that the transmission is unnecessary. Or
(6) When calculating the correlation level between the reception output signal and the transmission input signal, a delay device for delaying the reception signal by a propagation delay of the echo path is provided, or (7) the transmission input signal is previously processed. It is characterized in that a time delay required for the determination in the transmission determination unit is corrected by applying a delay, or (8) the determination result in the transmission determination unit is also used for detecting a double talk state.

[0006]

[Function] A band divider that divides a transmission input signal into low-frequency components and high-frequency components, a band combiner that combines low-frequency components and high-frequency components, and an echo in which a part of the reception output signal wraps around to the transmission input signal. Is canceled by a pseudo echo having an opposite phase, and only the low frequency component of the transmission input signal is passed through the echo canceling unit. Further, the transmission determining unit has a transmission determining unit that determines whether or not the transmission input signal is a signal to be transmitted, and when the transmission determining unit determines that transmission is unnecessary, the high frequency component of the transmission input signal is attenuated.

[0007]

1 is a diagram showing an electric circuit according to a first embodiment of the present invention, in which a transmission input signal is a Quadrature M
A band divider 12 such as an irror filter divides the high-frequency component and the low-frequency component. An audio signal, which is a main target signal of such a device, has an energy center of gravity in the low frequency range.
In addition, high frequencies have large air attenuation and little wraparound to the microphone. Further, the lower the frequency is, the lower the autocorrelation of the signal is, so that the effect of the echo canceller is high. For this reason, the deterioration of the overall performance is not so large even if the echo canceller processing is performed only on the low frequency components. Therefore, the low-frequency component is passed through the echo canceller 14, where the low-frequency component of the received signal is canceled by the pseudo echo of the opposite phase passed through the band divider 13, and the high-frequency component is directly processed into the band without any processing. Lead to vessel 15.

FIG. 2 is a diagram showing an electric circuit according to a second embodiment of the present invention, in which echo cancellation processing is performed on the low frequency component as in the case of the first embodiment, but for the high frequency component the transmission judging section is used. An attenuator controlled by 20 is added. In addition,
In the figure, the attenuator is represented by a switch 21 for the sake of clarity. Here, the first condition that the transmission determination unit 20 determines to perform the attenuation is that the power level of the transmission input signal is smaller than a predetermined value. That is, the gain is narrowed down except when it is considered that the voice of the near-end speaker is input. The power level calculation unit 22 calculates the power level of the transmission input signal, and the level is below a predetermined value. Attenuate if. The second condition is that the case where the received signal has a certain level of power is regarded as a case where an echo is likely to occur, and attenuation is applied when this is detected. The power level calculation unit 23
The power level of the received signal is calculated by and the attenuation is applied when the level is equal to or higher than a predetermined value. The third condition is that the received signal and the transmitted input signal have a high correlation, and in this case, the transmitted input signal is likely to be an echo. The correlation level calculation unit 24 calculates the correlation level. At this time, since there is a spatial propagation delay between the reception signal and the transmission input signal, a delay device 25 is added to the reception signal as shown in the figure.
It is better to calculate the correlation after passing through. The transmission determination is performed using at least one of the above conditions.

Further, in the configuration as shown in FIG. 2, it takes some time to calculate the power and the correlation level in the transmission determining section, so that it is difficult to strictly detect the start point of the voice of the speaker. Although a problem such as a break in a word occurs, in the present invention, since the attenuation is applied only to the high frequency component, it does not seriously hinder the communication. Further, the input transmission signal after the branch for the transmission judgment is passed through the delay device 26,
It is also possible to prevent the beginning of a word from being cut off by delaying the time required for transmission determination. Further, since the function of the transmission determination unit as described above is very close to the function of the double talk determination that determines whether or not the filter coefficient of the echo canceller can be updated, the output of the transmission determination unit can be used even in the double talk detection unit. It is also useful to use.

[0010]

As is apparent from the above description, according to the present invention, the following effects can be obtained. [Effects corresponding to claim 1] By performing the echo canceling process only on the low frequency component of the transmission input signal, the processing amount can be greatly reduced, and the miniaturization of the device and the cost reduction can be expected. [Effect corresponding to claim 2] When transmission is not necessary, attenuation is performed on the high frequency component of the transmission input signal, so that the echo remaining in the high frequency can be attenuated and the speech quality is improved. [Effect corresponding to claim 3] When the power level of the transmission input signal is equal to or lower than a certain value, it is determined that transmission is not necessary, so that all noise including echo other than that determined to be the utterance of the speaker is attenuated. You can do it. [Effect corresponding to claim 4] By determining that transmission is not necessary when the received output signal power level is equal to or higher than a certain value, the high frequency range is attenuated only when an echo may occur, and the utterance of the speaker is generated. The possibility of making a wrong decision is reduced, and the certainty of the call is increased. [Effect corresponding to claim 5] When the correlation level between the reception signal and the transmission input signal is equal to or higher than a certain value, it is judged that the transmission is not necessary, so that echo detection can be performed more reliably, so that the certainty of the call is increased. [Effect corresponding to claim 6] When the correlation level between the reception output signal and the transmission input signal is calculated, the correlation level can be calculated accurately by having a delay device for delaying the reception signal by a delay equivalent to the propagation delay of the echo path. Because it can be done, the certainty of the call increases. [Effects corresponding to claim 7] By correcting the time delay in the determination in the transmission determination unit by delaying the transmission input signal in advance, it is possible to prevent the beginning of the utterance from being uttered and increase the certainty of the call. . [Effect corresponding to claim 8] By using the determination result in the transmission determination unit also for detecting the double talk state,
Since the configuration can be simplified, the number of parts can be reduced, the number of program steps can be reduced, and the size and cost of the device can be reduced.

[Brief description of drawings]

FIG. 1 is an electric circuit diagram for explaining a first embodiment of the present invention.

FIG. 2 is an electric circuit diagram for explaining a second embodiment of the present invention.

FIG. 3 is a diagram for explaining the operation principle of an echo canceller.

[Explanation of symbols]

10 ... Speaker, 11 ... Microphone, 12, 13 ...
Band divider, 14 ... Echo canceling unit, 15 ... Band integrating unit, 20 ... Transmission determining unit, 21 ... Switch (attenuator),
22 ... Transmission input signal power level calculation unit, 23 ... Reception signal power level calculation unit, 24 ... Correlation level calculation unit,
25, 26 ... Delay device.

Claims (8)

[Claims] 1. A band divider for dividing a transmission input signal into a low-frequency component and a high-frequency component, a band combiner for synthesizing a low-frequency component and a high-frequency component, and a part of a reception output signal as a transmission input signal. An echo canceller, which has an echo canceling unit that cancels a wraparound echo by a pseudo echo having an opposite phase, and passes only the low-frequency component of a transmission input signal to the echo canceling unit. 2. A transmission determination unit that determines whether or not the transmission input signal is a signal to be transmitted, and attenuates the high frequency component of the transmission input signal when the transmission determination unit determines that transmission is unnecessary. The echo canceller according to claim 1, wherein 3. The echo canceller according to claim 2, wherein the transmission determination unit determines that transmission is unnecessary when the power level of the transmission input signal is below a certain value. 4. The echo canceller according to claim 2, wherein the transmission determination unit determines that transmission is unnecessary when the power level of the received output signal is equal to or higher than a certain value. 5. The echo canceller according to claim 2, wherein the transmission determining unit determines that transmission is unnecessary when the correlation level between the reception output signal and the transmission input signal is a certain value or more. 6. The echo canceller according to claim 5, further comprising a delay device for delaying the reception signal by a propagation delay of an echo path when calculating the correlation level between the reception output signal and the transmission input signal. . 7. The echo canceller according to claim 2, wherein the transmission input signal is delayed in advance to correct the time delay required for the determination by the transmission determination unit. 8. The echo canceller according to claim 2, wherein the determination result of the transmission determination unit is also used for detecting a double-talk state.