JP2003264483A - Device and method for suppressing echo, telephone set, and video telephone system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an echo suppressing device for restricting echo in a short time even in surrounding noises. <P>SOLUTION: The echo suppressing device 100 includes: a first frequency divider 103 for calculating a receiving sound signal on a frequency axis from a receiving sound signal on a time axis; a second frequency divider 104 for calculating a transmitting sound signal on the frequency axis from a transmitting signal on the time axis; an echo suppressing part 106 for suppressing the echo components of the transmitting sound signal on the frequency axis; and a frequency synthesizer 111 for synthesizing the transmitting sound signal on the time axis from the transmitting sound signal on the frequency axis. The noise components and the echo components are suppressed on the frequency axis. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an echo suppressing apparatus and method for eliminating echo interference in bidirectional voice communication, and further to a telephone and a videophone system using the echo suppressing apparatus.

[0002]

2. Description of the Related Art In a telephone or a videophone system that shares video and voice through a communication line, it is possible to talk with a receiving speaker and a transmitting microphone without using a handset for voice. When the far-end party and the near-end party are communicating voice signals with each other through the communication line, the voice signal received from the far-end party through the communication line (hereinafter, simply referred to as a receiving signal). The sound of the far-end person, which is amplified from the receiving speaker on the near-end person side, is reflected to the objects around the near-end person including the near-end person, and the reflected sound of the far-end person is input to the transmission microphone. . There has been an echo problem that the voice of the far-end party input to the transmission microphone becomes an echo and returns from the receiving speaker on the far-end party side to the ear of the far-end party with a delay.

In order to suppress such echo damage,
As a conventional acoustic echo canceller, for example, a document “Acoustic System and Digital Processing” (p209-p214,
Oga, Yamazaki, Kaneda, edited by The Institute of Electronics, Information and Communication Engineers,
Those described in (March 1995) are known.

A conventional acoustic echo canceller 1301 described in the above document will be described with reference to the block diagram of FIG.

A reception signal transmitted from a telephone at a remote place connected to a communication line is converted into a voice of the far end person, which is amplified by a reception speaker 1305, and is output. The voice of the far-end person expanded by the receiving speaker 1305 propagates through the space on the near-end side and reaches the transmitting microphone 1306,
The transmission microphone 1306 converts the voice of the near-end person and the noise of the near-end side into an echo-voice signal of the near-end person, the noise signal of the near-end side, and the echo signal, as an echo, and transmits the voice. Input to the microphone 1306.

Generally, an acoustic echo canceller 1301
Removes the echo signal by calculating the same signal as the echo signal included in the input transmission signal and subtracting the same signal as the echo signal from the input transmission signal. Here, the same signal as the calculated echo signal is referred to as a pseudo echo signal to distinguish it from the echo signal.

The echo signal input to the transmission microphone 1306 is input to the transmission microphone 1306 because the sound of the far end speaker, which is amplified by the reception speaker 1305, propagates through the space on the near end side side. Earpiece speaker 1305
From the transmitting microphone 1306 directly to the transmitting microphone 1306 and the transmitting microphone 13 reflected by the object on the near end side.
Indirect components reaching up to 06 are included. Therefore, the acoustic echo canceller 1301 cannot estimate the phase and amplitude of the indirect component included in the echo signal, and therefore estimates and estimates the path (hereinafter, simply referred to as an echo path) propagated until the echo signal is input. A signal for removing an echo signal, that is, a pseudo echo signal is calculated based on the echo path and the received signal that have been received.

A conventional echo canceller 1301 shown in FIG. 18 converts a received signal received from a communication line into a voice of a far end person and outputs the voice, and a voice of a near end person as a transmission signal. The receiving microphone 1306 that is converted and input, and the sound of the far-end person, which is expanded from the receiving speaker 1305, propagates through the space on the near-end side and reaches the transmitting microphone 1306, and the receiving microphone 130 is output as an echo.
6, an adaptive filter 1302 for estimating the echo path of the voice of the far-end person, which has been amplified until it is input to 6, and calculating a pseudo echo signal based on the estimated echo path and the received signal, and this adaptive filter. A coefficient updater 1303 for updating the filter coefficient of 1302 and a subtracter 1 for subtracting the pseudo echo signal from the transmission signal input from the transmission microphone 1306.
And 304.

If the operation of estimating the echo signal is performed well, the pseudo echo signal and the echo signal are considered to be substantially equal to each other, and the echo component contained in the transmission signal can be removed. Also, the coefficient updater 1303
Is based on the difference signal between the pseudo echo signal and the echo signal,
It is determined whether or not the operation of estimating the echo is good, and when it is determined that the operation is not good, the filter coefficient of the adaptive filter 1302 is updated. The adaptive filter 1302 repeats the calculation of the pseudo echo signal based on the updated filter coefficient, and follows the echo signal.

Next, the conventional acoustic echo canceller 130 is used.
The operation of No. 1 will be described.

When the reception signal is input, the adaptive filter 1
302 operates to calculate a pseudo echo signal. Next, the subtractor 1304 subtracts the pseudo echo signal calculated by the adaptive filter 1302 from the transmission signal input by the transmission microphone 1306. Next, it is determined whether or not the filter coefficient of the adaptive filter 1302 is an appropriate value based on the transmission signal and the reception signal from which the pseudo echo signal is subtracted, and if the filter coefficient is not set to an appropriate value. When determined, the coefficient updater 13
The filter coefficient is updated by 03.

Next, the coefficient updating method of the coefficient updating unit 1303 will be described.

The received signal is x (i), the transmitting microphone 1
The input signal to 306 is s (i), the adaptive filter 1302
, H (i), the output signal of the adaptive filter 1302 is y (i), and the output signal of the subtractor 1304 is e (i), the output signal e (i) of the subtractor 1304 is e (i) = S (i) -y (i) ... (1) It is represented by (1). In addition, Least Mean Sq
The coefficient updating equation is expressed by equation (2) by the uree method (hereinafter, simply referred to as LMS method).

H _{n + 1} (i) = h _n (i) + αe (i) x (i) (2)

The subscript n in equation (2) indicates the current sample value arranged in time series, and the subscript n + 1 indicates the sample value next to the current sample value n arranged in time series. Also, α
Is a step size parameter that determines the degree of coefficient update. Expression (2) updates h (i) so that e (i), which is the output signal of the subtractor 1304, becomes smaller. After being sufficiently updated, e (i) becomes sufficiently small, and the co component included in the transmission signal will be removed. At this time, h (i) has a value close to the transfer function between the speaker 1305 and the microphone 1306.

As described above, the conventional acoustic echo canceller 1301 repeats updating the coefficient and calculating the pseudo echo signal so that the pseudo echo signal becomes equal to the echo component included in the transmission signal, and outputs the transmission signal. It was constructed so that the contained echo component is eliminated.

Next, a conventional noise suppressing device will be described with reference to the block diagram of FIG.

As a conventional noise suppression device, for example, the document "Suppression of Acoustic Noise in Speech Using Sp
ectral Subtraction ”(IEEE Transactions on Acousti
cs, Speech, and Signal Processing, vol. ASSP-27, N
O.2, April, 1979) is known.

The conventional noise suppressor 14 shown in FIG.
01 is a frequency divider 1402 that divides the audio signal on the time axis into predetermined frames and calculates the frequency component included in the audio signal for each divided frame; and the amplitude value of the frequency component of the audio signal. Amplitude calculator 1403 for calculating
And a noise amplitude estimator 1404 for estimating the amplitude value of the frequency component of the noise signal included in the audio signal based on the amplitude value of the frequency component of the audio signal, and the noise signal estimated as the amplitude value of the frequency component of the audio signal Coefficient calculator 1405 that calculates a noise suppression coefficient based on the amplitude value of the frequency component of the noise, and multiplier 140 that multiplies the frequency component of the audio signal by the noise suppression coefficient.
6 and a frequency synthesizer 140 for synthesizing the audio signal on the time axis from the frequency component of the audio signal multiplied by the noise suppression coefficient.
Equipped with 7.

Next, the operation of the conventional noise suppressing device shown in FIG. 19 will be described.

As shown in FIG. 19, the sound of the near-end person on which noise generated from the vicinity of the near-end person is superimposed is converted into a transmission signal by the transmission microphone of the near-end party and input. The transmission signal in which the voice signal and the noise signal of the near-end person are superimposed is divided by the frequency divider 1402 into frequency components of the transmission signal. As a frequency division processing method of the frequency divider 1402, a Fourier transform processing for converting a transmission signal on the time axis into frequency components or a subband division processing having a plurality of bandpass filters is used.

The transmission signal divided by the frequency divider 1402 is processed by the amplitude calculator 1403 for each frequency unit of the transmission signal divided by the frequency divider 1402 or for each frequency band in which a plurality of divided frequencies are collected. ,
The amplitude value of the frequency component of the transmission signal is calculated.

Based on the amplitude value of the transmission signal calculated by the amplitude calculator 1403, the noise amplitude estimator 1404 calculates the frequency of the noise signal for each frequency unit or for each frequency band in which a plurality of divided frequencies are collected. An amplitude estimate of the component is estimated.

The amplitude value of the frequency component of the transmission signal calculated by the amplitude calculator 1403 and the noise amplitude estimator 1404
The coefficient calculator 1405 calculates a noise suppression coefficient for suppressing the noise signal based on the amplitude value of the frequency component of the noise signal estimated by. The noise suppression coefficient is, for example, 1 in the frequency band including only the voice signal of the near end person, and 0 or a sufficiently small value in the frequency band including only the noise signal.

The transmission signal in which the noise signal has been suppressed by the multiplier 1406 is combined by the frequency synthesizer 1407 into the transmission signal on the time axis from the frequency components and output. The processing of the frequency synthesizer 1407 is the reverse process of the frequency divider 1402, and inverse Fourier transform or subband synthesis processing is used.

As described above, the conventional noise suppressing device is configured to suppress the noise by multiplying the received signal by the noise suppressing coefficient for suppressing the noise signal.

[0027]

However, in the conventional acoustic echo canceller, the update of the filter coefficient of the adaptive filter is sufficiently performed when the echo path of the sound of the far-end person, which is amplified between the transmitting speaker and the receiving microphone, fluctuates. , The echo remains without being canceled, or when the received signal received by the receiving microphone contains a noise component, the residual echo increases due to a delay in updating the filter coefficient of the adaptive filter. was there.

On the other hand, the conventional noise suppressing device suppresses the noise component of the reception signal received by the reception microphone, but has a problem that it cannot suppress the echo component of the reception signal.

The present invention has been made in order to solve the conventional problems, and an echo suppressing effect can be obtained even when the echo path between the receiving speaker and the transmitting microphone fluctuates or when there is noise in the surroundings. The present invention is to provide an echo suppressor.

[0030]

An echo suppressor according to the present invention comprises a transmitting speaker for converting a received signal on a time axis received from a communication line into a voice of the far end and outputting the voice of the near end. And a first frequency divider that divides the received signal on the time axis and calculates the received signal on the frequency axis, and the transmitting speaker. A transfer function multiplier that calculates the pseudo echo signal on the frequency axis by multiplying the transfer function between the reception microphones by the reception signal on the frequency axis, and divides the transmission signal on the time axis by the frequency axis. A second frequency divider for calculating the above transmission signal, and transmission on the frequency axis received by the receiving microphone based on the pseudo echo signal on the frequency axis and the transmission signal on the frequency axis Eco that suppresses the echo component of the signal And a frequency synthesizer for synthesizing a transmission signal on the time axis with the echo component suppressed from a transmission signal on the frequency axis with the echo component suppressed. . With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

In addition, the echo suppressor of the present invention converts the received signal on the time axis received from the communication line into the voice of the far-end person and outputs it, and the voice of the near-end person on the time axis. A receiving microphone that converts the input signal into
A first frequency divider that divides the received signal on the time axis and calculates the received signal on the frequency axis, and divides the transmitted signal on the time axis to calculate the transmitted signal on the frequency axis. A second frequency divider, a transfer function between the transmitting speaker and the receiving microphone is estimated to calculate a pseudo echo signal on the frequency axis, and the pseudo echo signal on the frequency axis is calculated from the transmission signal on the frequency axis. An echo canceller unit that subtracts a signal to remove an echo component of the transmission signal on the frequency axis,
Based on the transmission signal on the frequency axis and the pseudo echo signal on the frequency axis from which the pseudo echo signal on the frequency axis is subtracted, the transmission on the frequency axis remaining without being removed by the echo canceller unit. An echo suppressor that suppresses the echo component of the speech signal, and a frequency synthesizer that synthesizes a transmission signal on the time axis with the echo component suppressed from a transmission signal on the frequency axis with the echo component suppressed And a container. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, the echo suppressor according to the present invention converts the received signal on the time axis received from the communication line into the voice of the far-end person and outputs it, and the voice of the near-end person on the time axis. Receiving microphone which is converted into a transmitting signal and input, a first frequency divider which divides the receiving signal on the time axis and calculates the receiving signal on the frequency axis, and the transmitting signal on the time axis A second frequency divider that divides the frequency component into a transmission signal on the frequency axis, a transmission noise suppression unit that suppresses a noise component of the transmission signal on the frequency axis, the transmission speaker, and the reception. The pseudo echo signal on the frequency axis is calculated by estimating the transfer function between the microphones, and the pseudo echo signal on the frequency axis is subtracted from the transmission signal on the frequency axis in which the noise component is suppressed, and the frequency is calculated. Echo component of on-axis transmission signal is removed Based on the echo canceller unit, the pseudo echo signal on the frequency axis, and the transmission signal on the frequency axis obtained by subtracting the pseudo echo signal on the frequency axis from the echo canceller unit. An echo suppressor for suppressing the echo component of the transmission signal on the frequency axis, and a transmission on the time axis with the echo component suppressed from the transmission signal on the frequency axis with the echo component suppressed And a frequency synthesizer for synthesizing signals. With this configuration, after suppressing the noise component included in the reception signal on the frequency axis, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the frequency axis Since the echo component contained in the received signal is suppressed above,
Even in the presence of ambient noise, the residual echo does not increase, and sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, the echo suppressor of the present invention converts the received signal on the time axis received from the communication line into the voice of the far-end person and outputs it, and the voice of the near-end person on the time axis. Receiving microphone which is input after being converted into a transmitting signal, a first frequency divider for dividing the receiving signal on the time axis and calculating the receiving signal on the frequency axis, and the transmitting signal on the time axis A second frequency divider that divides the transmission signal on the frequency axis, calculates a transmission signal on the frequency axis, a reception noise suppression unit that suppresses a noise component of the reception signal on the frequency axis, and a transmission signal on the frequency axis. A transmission noise suppression unit that suppresses noise components,
The pseudo echo signal on the frequency axis is calculated by estimating the transfer function between the transmitting speaker and the receiving microphone,
An echo canceller unit that subtracts a pseudo echo signal on the frequency axis from a transmission signal on the frequency axis in which the noise component is suppressed to remove an echo component of the transmission signal on the frequency axis, and the echo canceller. An echo suppression unit that suppresses the echo component of the transmission signal on the frequency axis that remains without being removed by a unit, and the echo component is suppressed from the transmission signal on the frequency axis that has the echo component suppressed. And a frequency synthesizer for synthesizing the transmission signal on the time axis. With this configuration, after suppressing the noise component included in the reception signal on the frequency axis, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the frequency axis is calculated. Since the echo component contained in the received signal is suppressed above, the residual echo does not increase even in the presence of ambient noise, and the processing calculation amount is smaller than the processing calculation amount on the time axis. Thus, sufficient echo suppression performance can be obtained.

Further, the echo suppressor according to the present invention converts the received signal on the time axis received from the communication line into the voice of the far-end person and outputs it, and the voice of the near-end person on the time axis. Receiving microphone which is converted into a transmitting signal and input, a first frequency divider which divides the receiving signal on the time axis and calculates the receiving signal on the frequency axis, and the transmitting signal on the time axis A second frequency divider that calculates a transmission signal on the frequency axis, a reception noise suppression unit that suppresses a noise component of the reception signal on the frequency axis, and a transmission signal on the frequency axis. A transmission noise suppression unit that suppresses noise components,
A double-talk determiner for determining whether or not the far-end voice component of the reception signal on the frequency axis and the voice component of the near-end voice signal of the transmission signal on the frequency axis are in a double-talk state or not. The pseudo echo signal is calculated by estimating the transfer function between the transmitting speaker and the receiving microphone, and the pseudo echo signal on the frequency axis is subtracted from the transmission signal on the frequency axis in which the noise component is suppressed. An echo canceller unit that removes the echo component of the transmission signal on the frequency axis, and an echo suppression unit that suppresses the echo component of the transmission signal on the frequency axis that remains without being removed by the echo canceller unit. A frequency synthesizer for synthesizing a transmission signal on the time axis in which the echo component is suppressed from a transmission signal on the frequency axis in which the echo component is suppressed. With this configuration, after suppressing the noise component included in the reception signal on the frequency axis, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the frequency axis is calculated. Since the echo component contained in the received signal is suppressed above, the residual echo does not increase even in the presence of ambient noise, and the processing calculation amount is smaller than the processing calculation amount on the time axis. Thus, sufficient echo suppression performance can be obtained.

Further, the echo suppressor according to the present invention converts the received signal on the time axis received from the communication line into the voice of the far-end person and outputs it, and the voice of the near-end person on the time axis. Receiving microphone which is converted into a transmitting signal and input, a first frequency divider which divides the receiving signal on the time axis and calculates the receiving signal on the frequency axis, and the transmitting signal on the time axis A second frequency divider that calculates a transmission signal on the frequency axis, a reception noise suppression unit that suppresses a noise component of the reception signal on the frequency axis, and a transmission signal on the frequency axis. A transmission noise suppression unit that suppresses noise components,
A double-talk determiner for determining whether or not the far-end voice component of the reception signal on the frequency axis and the voice component of the near-end voice signal of the transmission signal on the frequency axis are in a double-talk state or not. A transfer function multiplier for calculating a pseudo echo signal on the frequency axis by multiplying a transfer function between the transmission speaker and the reception microphone by the reception signal on the frequency axis, and based on the determination of the double talk determination device. A transfer function updater that updates the transfer function, an echo suppressor that suppresses the echo component of the transmission signal on the frequency axis that remains without being removed by the echo canceller, and an echo component that is suppressed. And a frequency synthesizer for synthesizing the transmission signal on the time axis in which the echo component is suppressed from the transmission signal on the frequency axis. With this configuration, after suppressing the noise component included in the received signal on the frequency axis,
The echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component contained in the reception signal is suppressed on the frequency axis. Even in the presence of noise, the residual echo does not increase, and sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, in the echo suppressor of the present invention, the echo suppressor calculates the amplitude value of the pseudo echo signal on the frequency axis, and the amplitude of the transmission signal on the frequency axis. A second amplitude calculator for calculating a value, and a coefficient calculator for calculating an echo suppression coefficient based on a comparison judgment of the amplitude value of the pseudo echo signal on the frequency axis and the amplitude value of the transmission signal on the frequency axis And a multiplier that multiplies the transmission signal on the frequency axis by the echo suppression coefficient to suppress the echo component of the transmission signal on the frequency axis. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, in the echo suppressor of the present invention, the echo suppressor subtracts the first amplitude calculator for calculating the amplitude value of the pseudo echo signal on the frequency axis from the pseudo echo signal on the frequency axis. A second amplitude calculator for calculating an amplitude value of the transmission signal on the frequency axis, and an amplitude value of the pseudo echo signal on the frequency axis and the pseudo echo signal on the frequency axis on the frequency axis A coefficient calculator that calculates an echo suppression coefficient based on a comparison determination with the amplitude value of the transmission signal, and the amplitude value of the transmission signal on the frequency axis from which the pseudo echo signal on the frequency axis is subtracted And a multiplier that multiplies the echo suppression coefficient to suppress the echo component of the transmission signal on the frequency axis that remains without being removed. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, in the echo suppressor of the present invention, the echo canceller section updates the transfer function between the transmitting speaker and the receiving microphone, and the transfer function for the receiving signal on the frequency axis. A first multiplier for calculating a pseudo echo signal on the frequency axis by multiplying by, and a pseudo echo signal on the frequency axis by subtracting the pseudo echo signal on the frequency axis from the transmission signal on the frequency axis A subtractor for removing the echo component, and a second multiplier for multiplying the reception signal on the frequency axis by the transmission signal on the frequency axis from which the echo component is removed and outputting the product to the transfer function updater. Is provided. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, in the echo suppressor of the present invention, the echo canceller unit updates the transfer function between the transmitting speaker and the receiving microphone, and the transfer function for the receiving signal on the frequency axis. A first multiplier that multiplies to calculate a pseudo echo signal on the frequency axis; and a frequency multiplier that subtracts the pseudo echo signal on the frequency axis from a transmission signal on the frequency axis in which noise components are suppressed. A subtracter for removing the echo component of the above transmission signal, a reception signal on the frequency axis and a transmission signal on the frequency axis from which the echo component has been removed are multiplied and output to the transfer function updater. And a second multiplier that does. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Furthermore, in the echo suppressor of the present invention, the echo canceller section updates the transfer function between the transmitting speaker and the receiving microphone, and a transfer function updater on the frequency axis in which noise components are suppressed. A first multiplier for calculating a pseudo echo signal on the frequency axis by multiplying the reception signal by the transfer function; and a frequency for subtracting the pseudo echo signal on the frequency axis from the transmission signal on the frequency axis. A subtracter for removing the echo component of the on-axis transmission signal, the transfer function updater by multiplying the reception signal on the frequency axis by the transmission signal on the frequency axis from which the echo component is removed And a second multiplier for outputting to.
With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, in the echo suppressor of the present invention, the echo canceller unit updates the transfer function between the transmitting speaker and the receiving microphone, and a transfer function updater on the frequency axis in which noise components are suppressed. A first multiplier for calculating a pseudo echo signal on the frequency axis by multiplying the reception signal by the transfer function, and a pseudo echo signal on the frequency axis from the transmission signal on the frequency axis in which a noise component is suppressed And a subtracter that removes the echo component of the transmission signal on the frequency axis by multiplying the reception signal on the frequency axis by the transmission signal on the frequency axis from which the echo component is removed. And a second multiplier that outputs the transfer function to the transfer function updater. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, in the echo suppressor of the present invention, the receiving noise suppressing section calculates the amplitude value of the receiving signal on the frequency axis, and the amplitude value of the noise component of the receiving signal on the frequency axis. And a coefficient calculator that calculates a noise suppression coefficient based on a comparison determination of the amplitude value of the received signal on the frequency axis and the amplitude value of the noise component of the received signal on the frequency axis. And a multiplier that multiplies the received signal on the frequency axis by the noise suppression coefficient to suppress the noise component of the received signal on the frequency axis. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, in the echo suppressor of the present invention, the noise amplitude estimator determines whether or not the amplitude value of the noise component of the reception signal on the frequency axis is larger than the amplitude value of the reception signal on the frequency axis. An amplitude value comparator for comparison and determination, and when the amplitude value of the noise component of the reception signal on the frequency axis is determined to be less than or equal to the amplitude value of the reception signal on the frequency axis, the noise component of the reception signal on the frequency axis. A noise amplitude value transposition device that replaces the amplitude value of the reception signal on the frequency axis with the amplitude value of, and the amplitude value of the noise component of the reception signal on the frequency axis is greater than the amplitude value of the reception signal on the frequency axis. When it is judged to be large,
And a noise amplitude value updater for updating the amplitude value of the noise component of the received signal on the frequency axis. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, in the echo suppressor of the present invention, the transmission noise suppressing section calculates the amplitude value of the transmission signal on the frequency axis, and the amplitude value of the transmission signal on the frequency axis. A noise amplitude estimator for estimating the amplitude value of the noise signal on the frequency axis included in the transmission signal on the frequency axis, and the amplitude value of the transmission signal on the frequency axis and the frequency axis The coefficient calculator that calculates a noise suppression coefficient based on the comparison determination with the amplitude value of the noise signal, and the transmission signal on the frequency axis is multiplied by the noise suppression coefficient to obtain the transmission signal on the frequency axis. And a multiplier that suppresses noise components. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, in the echo suppressor of the present invention, the noise amplitude estimator determines whether the amplitude value of the noise component of the transmission signal on the frequency axis is larger than the amplitude value of the transmission signal on the frequency axis. Whether the amplitude value of the noise component of the transmission signal on the frequency axis is less than or equal to the amplitude value of the transmission signal on the frequency axis. A noise amplitude value interchanger for replacing the amplitude value of the noise component of the speech signal with the amplitude value of the transmission signal on the frequency axis, and the amplitude value of the noise component of the transmission signal on the frequency axis on the frequency axis When it is determined that it is larger than the amplitude value of the transmitted signal,
And a noise amplitude value updater for updating the amplitude value of the noise component of the transmission signal on the frequency axis. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, in the echo suppressor of the present invention, the double talk determiner compares the amplitude value of the received signal on the frequency axis with the estimated amplitude of the noise component of the received signal on the frequency axis, and A first amplitude value comparator for determining whether or not a far-end voice component is detected in the received signal on the frequency axis;
The amplitude value of the transmission signal on the frequency axis is compared with the amplitude estimation value of the noise component of the transmission signal on the frequency axis, and the voice component of the near end is detected in the transmission signal on the frequency axis. A second amplitude value comparator for determining whether or not the voice signal of the far end person is detected in the reception signal on the frequency axis, and a voice signal of the near end person is detected in the transmission signal on the frequency axis. If not, a first flag setter that sets the first flag and outputs it to the echo canceller section, and a voice component of the far end person is not detected in the received signal on the frequency axis, and the signal on the frequency axis is transmitted. When the voice component of the near end is detected in the speech signal,
Alternatively, the second flag is set when the voice component of the far-end person is detected in the received signal on the frequency axis and the voice component of the near-end person is detected in the transmitted signal on the frequency axis, and the echo is set. And a second flag setting device for outputting to the canceller unit. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, in the echo suppressor of the present invention, the transfer function updater transmits the amplitude value of the received signal on the frequency axis in which the noise component is suppressed and the amplitude value of the received signal on the frequency axis in which the noise component is suppressed. An amplitude value comparator that determines whether or not the amplitude value of the speech signal is larger than the amplitude value of the speech signal, and the amplitude value of the reception signal on the frequency axis where the noise component is suppressed is on the frequency axis where the noise component is suppressed. When the amplitude value of the transmission signal is larger than the amplitude value of the transmission signal, the first transfer function amplitude value updater for updating the amplitude value of the transfer function, and the amplitude value of the reception signal on the frequency axis in which the noise component is suppressed are And a second transfer function amplitude value updater for updating the amplitude value of the transfer function when the amplitude value of the transmission signal on the frequency axis in which the noise component is suppressed is less than or equal to the amplitude value. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, the echo suppression method of the present invention divides the received signal on the time axis and calculates the received signal on the frequency axis in the first frequency division step and the transfer function between the transmitting speaker and the receiving microphone. A transfer function multiplication step of multiplying the reception signal on the frequency axis to calculate a pseudo echo signal on the frequency axis, and a step of dividing the transmission signal on the time axis to calculate a transmission signal on the frequency axis A frequency dividing step, and an echo suppressing step of suppressing an echo component of the transmission signal on the frequency axis received by the reception microphone based on the pseudo echo signal on the frequency axis and the transmission signal on the frequency axis. And a frequency synthesizing step of synthesizing a transmission signal on the time axis in which the echo component is suppressed from a transmission signal on the frequency axis in which the echo component is suppressed. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, the echo suppression method of the present invention divides the received signal on the time axis and calculates the received signal on the frequency axis in the first frequency dividing step, and divides the transmitted signal on the time axis, A second frequency division step of calculating a transmission signal on the frequency axis, a transfer function between the transmission speaker and the reception microphone is estimated to calculate a pseudo echo signal on the frequency axis, and the transmission signal on the frequency axis is calculated. An echo canceller step of removing the echo component of the transmission signal on the frequency axis by subtracting the pseudo echo signal on the frequency axis from the An echo suppression step of suppressing the echo component of the transmission signal on the frequency axis remaining without being removed by the echo canceller step based on the speech signal and the pseudo echo signal on the frequency axis; and the echo component. Has the structure of the suppressed transmission signal on the frequency axis and a frequency synthesizing step of synthesizing the transmission signals on the oppressed the time axis of the echo component. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, the echo suppression method of the present invention divides the received signal on the time axis and calculates the received signal on the frequency axis in the first frequency dividing step, and divides the transmitted signal on the time axis, A second frequency division step of calculating a transmission signal on the frequency axis, a transmission noise suppressing step of suppressing a noise component of the transmission signal on the frequency axis, and estimating a transfer function between the transmission speaker and the reception microphone. Then, the pseudo echo signal on the frequency axis is calculated, and the pseudo echo signal on the frequency axis is subtracted from the transmission signal on the frequency axis in which the noise component is suppressed to obtain the transmission signal on the frequency axis. An echo canceller step for removing an echo component, and an echo canceller step for removing the pseudo echo signal on the frequency axis and the transmission signal on the frequency axis on which the pseudo echo signal on the frequency axis is subtracted are removed. Without An echo suppression step of suppressing the echo component of the transmission signal on the frequency axis, and a transmission signal on the frequency axis in which the echo component is suppressed from the transmission signal on the frequency axis, on the time axis And a frequency synthesizing step of synthesizing a transmission signal. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, the echo suppression method of the present invention divides the received signal on the time axis and calculates the received signal on the frequency axis in the first frequency dividing step, and divides the transmitted signal on the time axis, A second frequency division step of calculating a transmission signal on the frequency axis, a reception noise suppression step of suppressing a noise component of the reception signal on the frequency axis, and a noise component of the transmission signal on the frequency axis. A transmitting noise suppressing step, calculating a pseudo echo signal on the frequency axis by estimating the transfer function between the transmitting speaker and the receiving microphone, An echo canceller step of subtracting a pseudo echo signal on the frequency axis to remove an echo component of the transmission signal on the frequency axis, and a transmission signal on the frequency axis remaining without being removed by the echo canceller step The eco And echo suppression process for suppressing component,
A frequency synthesizing step of synthesizing a transmission signal on the time axis in which the echo component is suppressed from a transmission signal on the frequency axis in which the echo component is suppressed. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Furthermore, the echo suppression method of the present invention divides the received signal on the time axis and calculates the received signal on the frequency axis in the first frequency division step, and divides the transmitted signal on the time axis, A second frequency division step of calculating a transmission signal on the frequency axis, a reception noise suppression step of suppressing a noise component of the reception signal on the frequency axis, and a noise component of the transmission signal on the frequency axis. Speaking noise suppression step, determining the voice component of the far end of the received signal on the frequency axis and the voice component of the near end of the transmitted signal on the frequency axis to determine whether it is in the double talk state And a double-talk determining step for estimating a transfer function between a transmitting speaker and a receiving microphone to calculate a pseudo echo signal, and calculating a pseudo echo signal on the frequency axis from the transmission signal on the frequency axis in which the noise component is suppressed. The frequency is subtracted from the echo signal An echo canceller step of removing the echo component of the above transmission signal, an echo suppression step of suppressing the echo component of the transmission signal on the frequency axis that remains without being removed in the echo cancellation step, and the echo component And a frequency synthesizing step for synthesizing the echo signal suppressed on the frequency axis from the echo signal suppressed on the frequency axis on the time axis. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

Further, the echo suppression method of the present invention divides the received signal on the time axis and calculates the received signal on the frequency axis in the first frequency dividing step, and divides the transmitted signal on the time axis, A second frequency division step of calculating a transmission signal on the frequency axis, a reception noise suppression step of suppressing a noise component of the reception signal on the frequency axis, and a noise component of the transmission signal on the frequency axis. Speaking noise suppressing step, detecting the voice component of the far end of the received signal on the frequency axis and the voice component of the near end of the transmitted signal on the frequency axis to determine whether or not the state is the double talk state. A double-talk determination step, a transfer function multiplication step of multiplying a transfer function between a transmitting speaker and a receiving microphone by a reception signal on the frequency axis to calculate a pseudo echo signal on the frequency axis, and the double-talk determination step. Based on the judgment of the process A transfer function updating step of updating a reaching function; an echo suppressing step of suppressing the echo component of the transmission signal on the frequency axis remaining without being removed in the echo canceller step; and a step of suppressing the echo component. And a frequency synthesizing step of synthesizing the transmission signal on the time axis in which the echo component is suppressed from the transmission signal on the frequency axis. With this configuration, the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal, and the echo component included in the reception signal on the frequency axis is suppressed. Therefore, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

The telephone set includes an echo suppressor according to the present invention, a reception section for receiving the encoded reception signal, a decoder for decoding the encoded reception signal into a reception signal on a time axis, and the time. A first amplifier that amplifies a reception signal on the axis, a reception speaker that converts the amplified reception signal on the time axis into voice and outputs the voice, and a voice of a near-end person is converted into a transmission signal on the time axis. And a second input amplifier for amplifying the transmission signal on the time axis, and an encoder for encoding the amplified transmission signal on the time axis,
And a transmitter for transmitting the encoded transmitter signal on the time axis. With this configuration, the noise component and the echo component included in the transmission signal can be suppressed and transmitted to the far-end party.

The videophone system includes the echo suppressor according to the present invention, a receiving section for receiving the far-end side video signal transmitted from the far-end party and the reception signal on the time axis, and the far-end side video signal on the screen. A video display unit displayed above, a first amplifier for amplifying a reception signal on the time axis from the far end person, and a reception signal on the time axis from the far end person for converting into a voice of the far end person. An output speaker, a microphone for inputting after converting the voice of the near end person into a transmission signal on the time axis, and a second amplifier for amplifying the transmission signal on the time axis, A camera is provided which captures an image of the near-end person side, converts it into a near-end side image signal and inputs the image, and a transmission unit which transmits the transmission signal on the time axis and the near-end side image signal. is doing. With this configuration, the noise component and the echo component included in the transmission signal can be suppressed and transmitted to the far-end party.

[0056]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of a first embodiment of an echo suppressor according to the present invention. This echo suppressor is different from other echo suppressors installed at a plurality of remote locations. It is applied to a communication system that is connected by a communication line and performs voice communication such as a telephone and a videophone system.

The echo suppressor 100 shown in FIG. 1 receives an input terminal 101 for inputting a received signal on the time axis transmitted from an echo suppressor at a remote place connected to a communication line, and an received signal on the time axis. A receiving speaker 112 that converts the voice of the far-end person into a voice and outputs the voice, a transmission microphone 113 that converts the voice of the near-end person into a transmission signal and inputs the voice, and a telephone at a remote place connected to a communication line. And an output terminal 102 for outputting a transmission signal on the time axis.

Here, the receiving signal on the time axis means a signal obtained by converting the voice of the far end person into a waveform of an electric signal. Further, the reception signal on the time axis is divided into predetermined frames, the frequency component included in the reception signal on the time axis is calculated for each divided frame, and the signal of the calculated frequency component is calculated on the frequency axis. It is called the reception signal above. Similarly, a signal obtained by converting the voice of the near end person into a waveform of an electric signal is called a transmission signal on the time axis. This transmission signal on the time axis is divided into predetermined frames, the frequency component included in the transmission signal on the time axis is calculated for each divided frame, and the signal of the calculated frequency component is It is called an on-axis transmission signal.

In the echo suppressor 100, the received signal on the time axis in which the noise signal of the far-end party is superimposed on the received far-end party's voice signal and the near-end party's side of the near-end person's voice signal. The transmission signal on the time axis in which the noise signal and the echo signal are superimposed is input. Therefore, in the following description, the noise signal on the far-end side and the voice signal of the far-end person, which are superimposed on the reception signal on the time axis, are respectively referred to as the noise component of the reception signal on the time axis and the noise signal on the time axis. It is called the voice component of the far end of the received signal. Further, the noise signal on the near-end side, the voice signal of the near-end person, and the echo signal, which are superimposed on the transmission signal on the time axis, are the noise component of the transmission signal on the time axis and the noise signal on the time axis, respectively. It is called the voice component of the near-end person of the transmission signal and the echo component of the transmission signal on the time axis.

The echo suppressor 100 further divides the received signal on the time axis received from the far end and calculates the frequency component of the received signal on the frequency axis by the first frequency divider 1.
03 and the second frequency divider 104 that divides the transmission signal on the time axis and calculates the frequency component of the transmission signal on the frequency axis.
And the transfer function of the echo path from when the voice sound of the far end speaker, which is amplified by the receiving speaker 112, is reflected by an object around the near end person including the near end person and reaches the transmitting microphone 113 (hereinafter, simply referred to as the receiving end. A transfer function multiplier for multiplying the frequency component of the reception signal on the frequency axis calculated by the first frequency divider 103 to calculate a pseudo echo signal on the frequency axis. 105
An echo suppression unit 106 for suppressing the echo component of the transmission signal on the frequency axis, and a frequency synthesizer 111 for synthesizing the transmission signal on the time axis from the transmission signal on the frequency axis in which the echo component is suppressed. Is equipped with.

Receiving speaker 112 and transmitting microphone 1
The transfer function between 13 is a function C (ω) = E (ω) / X obtained by the far-end voice signal X (ω) on the frequency axis and the echo signal E (ω) on the frequency axis. Say (ω). In this relational expression, ω represents the frequency.

The echo suppressor 106 also calculates the amplitude value of the frequency component of the pseudo echo signal on the frequency axis calculated by the transfer function multiplier 105.
7, a second amplitude calculator 107 for calculating the amplitude value of the frequency component of the reception signal on the frequency axis divided by the second frequency divider 104, and on the frequency axis calculated by the second amplitude calculator 107. An echo suppression coefficient calculator 109 for calculating an echo suppression coefficient based on the amplitude value of the frequency component of the pseudo echo signal and the amplitude value of the frequency component of the reception signal on the frequency axis calculated by the second amplitude calculator 108; An echo suppression coefficient multiplier 110 that multiplies the frequency component of the reception signal on the frequency axis divided by the first frequency divider 103 by the echo suppression coefficient, and the echo component included in the reception signal on the frequency axis It is supposed to suppress.

Next, the operation of the echo suppressor will be described.

First, the amplitude value of the frequency component of the reception signal on the frequency axis divided by the first frequency divider is input to X1, the transfer function between the reception speaker and the transmission microphone is C, and the transmission microphone 113 is input. Let S be the amplitude value of the voice signal of the near-end person and X2 be the amplitude value of the frequency component of the transmission signal on the frequency axis divided by the second frequency divider.
Equation (3) is obtained. X2 = S + C × X1 (3)

Further, the reception speaker 112 and the transmission microphone 113 preset in the transfer function multiplier 105.
If the transfer function between them is C ', the echo suppression coefficient h is expressed by equation (4). h = (X2-C '* X1) / X2 (4)

Further, the transmission signal Y on the frequency axis is
Equation (5) is obtained. Y = h × X2 = X2-C ′ × X1 (5)

As can be seen by comparing equations (3) and (5), if C and C'are equal, the transmission signal Y on the frequency axis is the signal of the near end of the transmission signal on the frequency axis. It becomes the voice component S, and by multiplying by the echo suppression coefficient h, the echo component included in the transmission signal on the frequency axis can be suppressed.

As described above, first, in the echo suppressor 106, the first amplitude calculator 107 calculates the amplitude value C ′ × X1 of the frequency component of the pseudo echo signal on the frequency axis, and the second amplitude calculation. The amplitude value X2 of the transmission signal on the frequency axis is calculated by the instrument 108. Next, the coefficient calculator 109 calculates the echo suppression coefficient h based on the comparison judgment between the amplitude value C ′ × X1 of the pseudo echo signal on the frequency axis and the amplitude value X2 of the frequency component of the reception signal on the frequency axis. It Then, the echo suppression coefficient multiplier 110
The echo suppression coefficient h is added to the frequency component of the transmission signal on the frequency axis.
Is multiplied by and the echo component contained in the frequency component of the received signal on the frequency axis is suppressed.

Next, referring to the flowchart of FIG.
The operation of the echo suppressor of this embodiment will be described.

The received signal on the time axis transmitted from the echo suppressor at the remote place connected to the communication line is converted by the receiving speaker 112 into the voice of the far end person and output. The sound of the far-end person, which is expanded from the receiving speaker 112, propagates through the space on the side of the near-end person and the transmitting microphone 1
13, the voice of the near-end person and the noise of the near-end person side as an echo are transmitted by the transmitting microphone 113 on the time axis in which the voice signal of the near-end person, the noise signal of the near-end person side and the echo signal are superimposed. It is converted into a speech signal and input.

First, in the first frequency division step S120,
The first frequency divider 103 divides the received signal on the time axis transmitted from the telephone at the remote place into frequency components. To calculate the frequency component, time by Fourier transform −
Frequency conversion processing or subband division processing using a plurality of bandpass filters can be used. Then the second
In the frequency division step S121, the second frequency divider 104
Thus, the transmission signal on the time axis input by the transmission microphone 113 is divided into frequency components. Similarly, for the calculation of the frequency component, time-frequency conversion processing by Fourier transform or subband division processing by a plurality of bandpass filters can be used.

Then, in step S122a, the transfer function multiplier 105 calculates the reception signal on the frequency axis divided by the first frequency divider 103 and the preset transfer function between the reception speaker and the transmission microphone. The multiplication is performed to calculate the pseudo echo signal on the frequency axis. Next, in step S122b, the second amplitude calculator 108 calculates the amplitude value of the transmission signal on the frequency axis divided by the second frequency divider 104. The calculation of the amplitude value here may be the calculation of the amplitude value for each frequency component of the transmission signal on the frequency axis or the calculation of the amplitude value for each frequency band in which a plurality of frequency components are collected. .

Then, in step S122c, the amplitude value of the pseudo echo signal on the frequency axis calculated by the transfer function multiplier 105 is calculated. The calculation of the amplitude value here may be the calculation of the amplitude value for each frequency component of the pseudo echo signal on the frequency axis or the calculation of the amplitude value for each frequency band in which a plurality of frequency components are collected. . Next, in step S122d, the coefficient calculator 109 causes the first
The echo suppression coefficient is calculated based on the comparison judgment between the amplitude value of the pseudo echo signal on the frequency axis calculated by the amplitude calculator and the amplitude value of the transmission signal on the frequency axis calculated by the second amplitude calculator. It

Then, in step S122e, the echo suppression coefficient multiplier 110 causes the second frequency divider 104 to operate.
The received signal on the frequency axis divided by is multiplied by the echo suppression coefficient h calculated by the coefficient calculator 109.

In the flow chart shown in FIG.
Steps S122a to S surrounded by dotted lines
The echo suppression step S122 including the steps up to 122e.
Say. Next, in the frequency synthesizing step S123, the frequency synthesizer 111 synthesizes the transmission signal on the time axis from the transmission signal on the frequency axis multiplied by the echo suppression coefficient h, and transmits the echo suppressed transmission on the time axis. It is output as a speech signal. As a processing method for synthesizing the original transmission signal on the time axis from the frequency components of the transmission signal on the frequency axis, an inverse Fourier transform process or a subband synthesis process can be used.

As described above, according to the present embodiment, the echo suppression coefficient is calculated based on the amplitude value of the pseudo echo signal on the frequency axis and the amplitude value of the reception signal on the frequency axis. Since the echo component included in the received signal is suppressed on the frequency axis, sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

(Second Embodiment) FIG. 3 is a block diagram showing the configuration of a second embodiment of an echo suppressor according to the present invention. This echo suppressor is installed in a plurality of remote locations. The present invention is applied to a communication system that is connected to another echo suppression device that is used by a communication line and performs voice communication such as a telephone and a videophone system.

The echo suppressor 200 shown in FIG. 3 inputs an input terminal 201 for inputting a time-domain reception signal transmitted from a remote echo suppression device connected to a communication line, and a time-domain reception signal. A receiving speaker 216 that converts the voice of the far-end person into a voice and outputs the voice, a microphone 217 that converts the voice of the near-end person into a voice signal and inputs the voice signal, and echo suppression of a remote place connected to the communication line. The device is provided with an output terminal 202 for transmitting a transmission signal on the frequency axis.

The echo suppressor 200 further divides the received signal on the time axis received from the far end person, calculates the received signal on the frequency axis, and the first frequency divider 203, and transmits on the time axis. A second frequency divider 204 that divides a signal and calculates a transmission signal on the frequency axis, and a transmission on the frequency axis calculated by the second frequency divider 204 by estimating a pseudo echo signal on the frequency axis An echo canceller unit 205 for subtracting a pseudo echo signal on the frequency axis estimated from the signal,
An echo suppressor 210 that suppresses the echo component remaining without being removed by the echo canceller 205, and a frequency that synthesizes a transmission signal on the time axis from the frequency component of the transmission signal on the frequency axis in which the echo component is suppressed. And a synthesizer 215.

The echo canceller unit 205 estimates the transfer function between the receiving speaker 216 and the transmitting microphone 217 by updating the filter coefficient of the adaptive filter, and further updates the transfer function estimator 206.
And the frequency component of the reception signal on the frequency axis divided by the first frequency divider 203 and the transfer function estimated by the transfer function estimator 206 are multiplied to calculate the frequency component of the pseudo echo signal on the frequency axis. A first multiplier 207 that
A subtractor 208 for subtracting the frequency component of the pseudo echo signal on the frequency axis calculated by the first multiplier 207 from the frequency component of the transmission signal on the frequency axis divided by the second frequency divider 204; A second multiplier 209 for multiplying the frequency component of the above reception signal by the frequency component of the transmission signal on the frequency axis whose pseudo echo component has been subtracted by the subtractor 208 is provided. Based on the estimation, the frequency component of the pseudo echo component is subtracted from the frequency component of the transmission signal on the frequency axis input to the microphone 217 to remove the echo component.

The echo suppressing section 210 calculates the amplitude value of the frequency component of the pseudo echo signal on the frequency axis calculated by the echo canceller section 205. The first amplitude calculator 211.
A second amplitude calculator 212 for calculating the amplitude value of the frequency component of the transmission signal on the frequency axis on which the pseudo echo signal on the frequency axis is subtracted by the echo canceller unit 205;
The amplitude value of the frequency component of the pseudo echo signal on the frequency axis calculated by the amplitude calculator 211 and the second amplitude calculator 212
A coefficient calculator 213 that calculates an echo suppression coefficient based on the amplitude value of the frequency component of the transmission signal on the frequency axis after subtracting the pseudo echo signal on the frequency axis calculated by The echo canceller unit 2 is provided with an echo suppression coefficient multiplier 214 that multiplies the frequency component of the transmission signal on the frequency axis from which the signal is subtracted by the echo suppression coefficient.
In 05, the echo component remaining without being removed is suppressed.

Next, the estimation operation of the transfer function estimator 206 of the echo canceller section will be described. S (ω) = C (ω) X (ω) ... (6) Y (ω) = H _n (ω) X (ω) ... (7) E (ω) = S (ω) −Y (Ω) ・ ・ ・ (8) H _{n + 1} (ω) = H _n (ω) + αE (ω) X (ω) ・ ・ ・ (9)

Equations (6), (7), (8), and (9) are respectively the transmission signal S (ω) on the frequency axis and the first signal on the frequency axis.
Output Y (ω) of multiplier 207, subtracter 20 on the frequency axis
8 shows an output E (ω) of 8 and a transfer function H (ω) on the frequency axis of the echo canceling unit. Where C
(Ω) is the transfer function from the receiving speaker to the transmitting microphone, X (ω) is the receiving signal on the frequency axis, α is the step size parameter, subscript n represents the present, and subscript n +
1 represents a sample at the next time point on the time series. The formula (8) is a filter coefficient updating formula of the adaptive filter based on the LMS method, but may be based on another filter coefficient updating formula.

Transfer function C between speaker and microphone
For the estimation of (ω), the transfer function H (ω) of the adaptive filter included in the transfer function updater 206 and the transfer function C (ω) between the speaker and the microphone are obtained by updating the filter coefficient of the adaptive filter of Expression (9). And the difference is small. Therefore, after sufficiently converging by updating the filter coefficient of the adaptive filter, H (ω) becomes substantially equal to C (ω), and the echo component is present in E (ω) from Eqs. (6) to (8). It will not be included. Here, the transmission signal on the frequency axis input to the transmission microphone 217 is described as only the echo signal, but even when a voice signal other than the echo is actually included, the echo canceller unit 205 outputs only the echo signal. Operates to erase.

Further, in the echo canceller unit 205, the echo is erased when the filter coefficient of the adaptive filter is sufficiently updated, but when the filter coefficient is not sufficiently updated, or the speaker 216 transmits the voice. When the transfer function between the microphones 217 is changing, echoes are not sufficiently removed and are output from the echo canceller unit 205 as echo residuals. The echo suppressing section 210 installed in the latter stage of the echo canceller section 205 suppresses the echo residual that remains without being removed.

Next, referring to the flow chart of FIG.
The operation of the echo suppressor according to the second embodiment will be described.

The received signal on the frequency axis transmitted from the echo suppressor at the remote place connected to the communication line is converted into the voice of the far end person which is amplified by the receiving speaker 216 and output. The sound of the far-end person, which is expanded from the receiving speaker 216, propagates in the space of the near-end person side and the transmitting microphone 2
17, and the voice of the near-end person and the noise of the near-end person side as an echo, and on the frequency axis where the voice signal of the near-end person, the noise signal of the near-end person side, and the echo signal are superimposed by the transmission microphone 217. Is converted into a transmission signal and input.

Here, the operation of calculating the frequency component of the reception signal on the frequency axis processed in the first frequency division step S220 is performed by the operation of the reception signal on the frequency axis processed in the first frequency division step S120. The operation is the same as the operation of calculating the frequency component, and the operation of calculating the frequency component of the transmission signal on the frequency axis processed in the second frequency division step S221 is processed in the second frequency division step S121. The operation is the same as the operation of calculating the frequency component of the transmission signal on the frequency axis. Further, the operation of suppressing the echo component processed in the echo suppression step S223 is the same as the operation of suppressing the echo component processed in the echo suppression step S122 described above, and further, the frequency processed in the frequency synthesis step S224. The operation of synthesizing the transmission signal on the time axis from the frequency component of the transmission signal on the axis is performed by the above-described frequency synthesizing step S12.
The operation is the same as the operation of synthesizing the transmission signal on the time axis from the frequency component processed in 3, and since the operation has been described in the first embodiment, the description in this embodiment will be omitted.

The echo canceling step S222 includes the steps S222a to S222d described below. In step S222a, the transfer function between the reception speaker 216 and the transmission microphone 217 estimated by the transfer function estimator 206 by the multiplier 207 and the frequency of the reception signal on the frequency axis divided by the first frequency divider 203. The frequency component of the pseudo echo signal on the frequency axis is calculated by multiplying by the component.

Next, in step S222b, the second subtractor 208 multiplies the transfer function by the frequency component of the transmission signal on the frequency axis divided by the second frequency divider 204 to obtain the reception signal on the frequency axis. The frequency component, that is, the frequency component of the pseudo echo signal on the frequency axis is subtracted. Then, in step S222c, the second multiplier 209 subtracts the frequency component of the pseudo echo signal on the frequency axis from the frequency component of the transmission signal on the frequency axis,
The frequency components of the received signal on the frequency axis are multiplied and output to the transfer function estimator 206. Then, step S22
In 2d, the transfer function estimator 206 causes step 22
Based on the multiplication result multiplied by 2c, it is determined whether or not the estimated transfer function is updated.

As described above, according to the present embodiment, the echo suppression coefficient is calculated based on the amplitude value of the pseudo echo signal on the frequency axis and the amplitude value of the reception signal on the frequency axis. By suppressing the echo component included in the received signal on the frequency axis on the frequency axis, it is possible to obtain sufficient echo suppression performance with a smaller processing calculation amount than the processing calculation amount on the time axis.

(Embodiment 3) FIG. 5 is a block diagram showing the configuration of an echo suppressor according to a third embodiment of the present invention. This echo suppressor is installed in a plurality of remote locations. The present invention is applied to a communication system that is connected to another echo suppressor through a communication line and performs voice communication such as a telephone and a videophone system.

The echo suppressor 300 shown in FIG.
An input terminal 301 for receiving a reception signal on the time axis from an echo suppressor at a remote place connected to a communication line, and a reception signal on the time axis received from the far end person, converted into a voice of the far end person. And a receiving speaker 313 that outputs the same, and a transmitting microphone 3 that converts the voice of the near end person into a voice signal and inputs the voice signal.
14 and an output terminal 302 for transmitting a transmission signal on the time axis to an echo suppressing device at a remote place connected to a communication line.

The echo suppressor 300 further divides the received signal on the time axis received from the far-end person, and calculates the frequency component of the received signal on the frequency axis.
03, and a second frequency divider 304 that divides the transmission signal on the time axis and calculates the frequency component of the transmission signal on the frequency axis.
And a noise suppression unit 305 that suppresses a noise component included in the frequency component of the transmission signal on the frequency axis divided by the second frequency divider 304 and emphasizes the voice component of the near-end person,
The frequency component of the pseudo echo signal on the frequency axis is estimated, and the frequency component of the pseudo echo signal on the estimated frequency axis is subtracted from the frequency component of the transmission signal on the frequency axis calculated by the second frequency divider 304. Echo canceller section 31
0, an echo suppression unit 310 that suppresses the echo component remaining without being removed by the echo canceller unit 205, and a transmission signal on the frequency axis from the frequency component of the transmission signal on the frequency axis in which the echo component is suppressed. And a frequency synthesizer 312 for synthesizing.

Next, referring to the flow chart of FIG.
The operation of the echo suppressor according to the third embodiment will be described.

The reception signal on the frequency axis transmitted from the remote telephone connected to the communication line is the reception speaker 21.
It is converted into the voice of the far end person who is made loud by 6 and is output. The sound of the far-end person, which is amplified from the receiving speaker 216, propagates through the space on the side of the near-end person and reaches the transmitting microphone 217. Thus, the voice signal of the near-end person, the noise signal of the near-end person and the echo signal are converted into a transmission signal on the frequency axis and input.

Here, the operation of calculating the frequency component of the reception signal on the frequency axis processed in the first frequency division step S320 is the same as the above-mentioned first frequency division step S120, S220.
The operation is the same as the operation of calculating the frequency component of the reception signal on the frequency axis processed in step S3, and the second frequency division step S3.
The operation of calculating the frequency component of the transmission signal on the frequency axis processed in 21 is performed by the second frequency division step S121,
The operation is the same as the operation of calculating the frequency component of the transmission signal on the frequency axis processed in S221. Further, the operation of removing the echo component superimposed on the frequency-domain received signal processed in the echo canceling step S323 is superimposed on the frequency-axis received signal processed in the echo canceling step S222. The operation is the same as the operation for removing the echo component, and the operation for suppressing the echo component processed in the echo suppressing step S324 is the same as the echo suppressing step S12 described above.
2. The operation is the same as the operation of suppressing the echo component processed in S223. Furthermore, the operation of synthesizing the transmission signal on the time axis from the frequency components of the transmission signal on the frequency axis processed in the frequency synthesis step S323 is the same as the above-described frequency synthesis step S1.
23, which is the same as the operation of synthesizing the transmission signal on the time axis from the frequency components of the transmission signal on the frequency axis processed in S224, and the operation in the first and second embodiments Since it has been described, the description in this embodiment will be omitted.

First, in step S322a, the amplitude calculator 306 calculates the amplitude value of the frequency component of the transmission signal on the frequency axis. The calculation of the amplitude value here may be the calculation of the amplitude value for each frequency component or the calculation of the amplitude value for each frequency band in which a plurality of frequency components are collected. Next, in step S322b, the noise amplitude value estimation step, the noise amplitude estimator 307 causes the amplitude calculator 30
Based on the amplitude value of the frequency component of the transmission signal on the frequency axis calculated by 6, the amplitude value of the frequency component of the noise signal is estimated.

Next, in step S322c, the coefficient calculator 308 causes the amplitude value of the frequency component of the transmission signal on the frequency axis calculated in step S322a and the amplitude value of the frequency component of the noise signal estimated in step S322b. Based on and, the noise suppression coefficient is calculated. Next, in step S322d, the noise suppression coefficient calculated in step S322c is multiplied by the frequency component of the transmission signal on the frequency axis, and the frequency component of the noise suppressed transmission signal on the frequency axis is transmitted to the echo canceller unit 310. Is output. The steps from S322a to S322d are collectively referred to as a transmission noise suppressing step S322.

Next, a method of calculating the noise suppression coefficient which is processed in step S322b included in the transmission noise suppression step S322 will be described.

The voice component is s (i) and the noise component is n (i).
Then, the transmission signal x (i) on the time axis input to the transmission microphone by superimposing the noise component on the voice component is expressed by the equation (10). x (i) = s (i) + n (i) (10)

Further, when the transmission signal on the time axis is Fourier-transformed into the transmission signal on the frequency axis, equation (11) is obtained. X (ω) = S (ω) + N (ω) (11)

Here, the noise is a stationary signal, the voice component and the noise component are uncorrelated, energy addition is established, and the Fourier transform S ′ of the estimated signal s ′ (i) of the voice signal is obtained.
Assume that (ω) is represented by equation (12). S '(ω) = | S' (ω) | exp {jarg (X (ω))} = {| X (ω) |-| N (ω) |} exp {jarg (X (ω))}・ ・ (12)

Transforming equation (12), equation (1
It becomes like 3). S '(ω) = H (ω) X (ω) (13)

Here, H (ω) in the equation (13) is H (ω) = {| X (ω) | − | N (ω) |} / | X (ω) | (14) Is. Using equation (13) and equation (14) with subscript k representing the number for each frequency band, equation (13) is transformed into equation (1).
Expression (14) is expressed as Expression (16) as in Expression 5). _{S k = H k X k ···} (15) H k = {| X k | - | N k |} / | X k | ··· (16)

Expressions (15) and (16) are obtained by subtracting the amplitude value of the frequency component of the noise signal from the amplitude value of the frequency component of the transmission signal on the frequency axis, By dividing by the amplitude value of the component and multiplying the frequency component of the transmission signal on the frequency axis for each frequency band as the noise suppression coefficient, it is possible to calculate the transmission signal on the frequency axis in which the noise component is suppressed. Shows.

Next, referring to the flow chart of FIG.
The noise amplitude value estimation step will be described in more detail. The noise amplitude value estimating step further includes an amplitude value comparing step S701, a noise amplitude value updating step S702, and a noise amplitude value replacing step S70.
Includes 3 and 3.

In FIG. 7, the symbol k indicates a predetermined frequency unit in which the amplitude value of the frequency component of the noise signal is calculated, or a frequency band number in which a plurality of divided frequencies are collected. The symbol i indicates a time-series number in which the amplitude value of the frequency component of the noise signal is calculated, and the amplitude value of the frequency component of the noise signal is repeatedly estimated for each frame, which is determined in advance.

First, in the amplitude value comparison step S701, the amplitude value X _i of the frequency component of the input reception signal on the frequency axis is input.
(K) is compared with the amplitude value N _i-1 (k) of the frequency component of the noise signal estimated in the previous frame. When the amplitude value N _i-1 (k) of the frequency component of the noise signal estimated in the immediately preceding frame is smaller than the amplitude value X _i (k) of the frequency component of the reception signal on the frequency axis, the noise amplitude update Proceed to step S702. On the other hand, when the amplitude value N _i-1 (k) of the frequency component of the noise signal estimated in the previous frame is equal to or larger than the amplitude value X _i (k) of the frequency component of the reception signal on the frequency axis, the noise amplitude Proceed to the replacement step S703.

Next, in the noise amplitude updating step S703,
The amplitude value N _i-1 (k) of the frequency component of the noise signal estimated in the immediately preceding frame is multiplied by a predetermined update amount Lup to obtain the estimated amplitude value of the frequency component of the updated noise signal. N _i (k). Here, the update amount Lup is a value larger than 1, and is set in advance so as to be an update amount of, for example, about 0.1 dB to 3 dB. When the estimated amplitude value N _i-1 (k) of the frequency component of the noise signal is smaller than the amplitude value X _i (k) of the frequency component of the reception signal on the frequency axis, N
_i (k) = N _i−1 (k) * Lup, and the estimated noise amplitude value is gradually increased when the noise level is increasing,
It is also possible to estimate fluctuating noise.
Moreover, the reason why the estimated amplitude value of the frequency component of the noise signal is gradually increased is to prevent excessive reaction to the voice signal in the voice section.

Next, in the noise amplitude value replacing step S702, the estimated amplitude value N _i (k) of the frequency component of the noise signal is
Amplitude value X _i (k) of the frequency component of the transmission signal on the frequency axis
To the estimated amplitude value N _i (k) of the frequency component of the updated noise signal. When the amplitude value N _i-1 (k) of the frequency component of the noise signal is larger than the amplitude value X _i (k) of the frequency component of the reception signal on the frequency axis, N _i (k) = X _i
The reason (k) is to detect the noise level in the non-voice section between syllables.

As described above, according to the present embodiment, after suppressing the noise component contained in the received signal on the frequency axis, the amplitude value of the frequency component of the pseudo echo signal on the frequency axis and the frequency value on the frequency axis are obtained. Since the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the reception signal of, and the echo component included in the reception signal is suppressed on the frequency axis,
Even in the presence of ambient noise, the residual echo does not increase, and sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

(Embodiment 4) FIG. 8 is a block diagram showing the configuration of an echo suppressor according to a fourth embodiment of the present invention. The echo suppressor is installed at a plurality of remote locations. The present invention is applied to a communication system that is connected to another echo suppressor by a communication line and performs voice communication such as a telephone and a videophone system.

The echo suppressor 400 shown in FIG.
An input terminal 401 for receiving a reception signal on the time axis from an echo suppressing device at a remote place connected to a communication line, and a reception signal on the time axis received from the far end person, converted into a voice of the far end person. And the output speaker 414, and the transmission microphone 4 for converting the voice of the near end person into a voice signal and inputting the voice signal.
15 and an output terminal 402 for transmitting a transmission signal on a time axis to an echo suppressing device at a remote place connected to a communication line.

The echo suppressor 400 further divides the received signal on the time axis received from the far end and calculates the frequency component of the received signal on the frequency axis by the first frequency divider 4.
03, and a second frequency divider 404 that divides the transmission signal on the time axis and calculates the frequency component of the transmission signal on the frequency axis.
And the noise component included in the frequency component of the received signal on the frequency axis divided by the first frequency divider 403 is suppressed, and the voice component of the far-end party included in the frequency component of the received signal on the frequency axis is emphasized. The receiving noise suppressing section 405, and the second
A transmission noise suppression unit 410 that suppresses a noise component included in the frequency component of the transmission signal on the frequency axis divided by the frequency divider 404 and emphasizes the voice component of the near-end person, a reception speaker 414, and transmission. The frequency component of the pseudo echo signal on the frequency axis is calculated from the multiplication of the transfer function between the microphones 415 and the frequency component of the reception signal on the frequency axis, and the transmission on the frequency axis divided by the second frequency divider 404 is performed. An echo canceller unit 411 that removes the frequency component of the pseudo echo signal on the frequency axis from the frequency component of the signal to remove the echo component, and an echo suppression unit 412 that suppresses the echo component that remains without being removed by the echo canceller unit 411. And a frequency synthesizer 4 for synthesizing the transmission signal on the frequency axis from the frequency component of the transmission signal on the frequency axis in which the echo component is suppressed.
13 and 13.

The reception noise suppressing section 405 includes a frequency divider 4
Amplitude calculator 406 for calculating the amplitude value of the frequency component of the reception signal on the frequency axis divided by 03, and the amplitude of the noise component included in the frequency component of the reception signal on the frequency axis calculated by the amplitude calculator 406 A noise amplitude estimator 407 that estimates a value, an amplitude value of a frequency component of the reception signal on the frequency axis calculated by the amplitude calculator 406, and the amplitude value of the reception signal on the frequency axis estimated by the noise amplitude estimator 407 are included in the reception signal. A coefficient calculator 408 for calculating a noise suppression coefficient for suppressing the noise component included in the received signal on the frequency axis based on the estimated amplitude value of the frequency component of the noise component, and the frequency divider 4
A multiplier 409 that multiplies the frequency component of the reception signal on the frequency axis calculated by 03 by the noise suppression coefficient calculated by the coefficient calculator 408 to suppress the noise component included in the reception signal on the frequency axis. The noise component included in the frequency component of the transmission signal on the frequency axis is suppressed, and the near-end person's voice component is emphasized.

Next, referring to the flow chart of FIG.
The operation of the echo suppressor according to the fourth embodiment will be described.

The reception signal on the time axis transmitted from the remote telephone connected to the communication line is the reception speaker 414.
Is converted into the voice of the far-end person and output. The voice of the far-end person, which is amplified from the receiving speaker 414, propagates through the space of the near-end person side and reaches the transmitting microphone 415, and as the echo, together with the voice of the near-end person and the noise of the near-end person side, the transmitting microphone 415. Thus, the voice signal of the near-end person, the noise signal of the near-end person and the echo signal are converted into a transmission signal on the frequency axis and input.

Here, the operation of calculating the frequency component of the reception signal on the frequency axis processed in the first frequency division step S420 is the same as the above-mentioned first frequency division step S120, S22.
0, the same operation as calculating the frequency component of the reception signal on the frequency axis processed in S320, and calculating the frequency component of the transmission signal on the frequency axis processed in the second frequency division step S421. The operation is the same as the operation of calculating the frequency component of the transmission signal on the frequency axis processed in the above-described second frequency division steps S121, S221, and S321. Furthermore, the operation of removing the echo component superimposed on the received signal on the frequency axis processed in the echo cancellation step S424 is performed by the echo cancellation step S22 described above.
2, the operation is the same as the operation of removing the echo component superimposed on the received signal on the frequency axis processed in S323, and the operation of suppressing the echo component processed in the echo suppression step S425 is the above-described echo suppression step. S122, S223,
The operation is the same as the operation of suppressing the echo component processed in S324. Furthermore, the operation of synthesizing the transmission signal on the time axis from the frequency components of the transmission signal on the frequency axis processed in the frequency synthesis step S426 is the same as the above-described frequency synthesis step S12.
The operation is the same as the operation of synthesizing the transmission signal on the frequency axis on the time axis from the frequency components of the transmission signal on the frequency axis processed in S3, S223, and S325. Since the operation has been described in the embodiment, the description in this embodiment will be omitted.

Further, the receiving noise suppressing step S422 and the transmitting noise suppressing step S423 suppress the noise component contained in the transmitting signal on the frequency axis of the transmitting noise suppressing step S322 described in the third embodiment. Similar to the behavior,
In the reception noise suppression step S422, the noise component included in the frequency component of the reception signal on the frequency axis is suppressed, and in the transmission noise suppression step S423, the noise component included in the frequency component of the transmission signal on the frequency axis is suppressed. To do. By arranging the reception noise suppressing step S422 and the transmission noise suppressing step S423 before the echo canceling step S424, it is possible to prevent a delay in updating the filter coefficient of the adaptive filter due to the inclusion of noise components.

As described above, according to the present embodiment, after suppressing the noise component contained in the reception signal on the frequency axis, the amplitude value of the frequency component of the pseudo echo signal on the frequency axis and the frequency value on the frequency axis are obtained. The echo suppression coefficient is calculated based on the amplitude value of the frequency component of the reception signal of, and the echo component included in the reception signal is suppressed on the frequency axis.
Even in the presence of ambient noise, the residual echo does not increase, and sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

(Embodiment 5) FIG. 10 is a block diagram showing the configuration of an echo suppressor according to a fifth embodiment of the present invention. This echo suppressor is installed in a plurality of remote locations. The present invention is applied to a communication system that is connected to another echo suppressor through a communication line and performs voice communication such as a telephone and a videophone system.

The echo suppressor 500 shown in FIG. 10 has an input terminal 501 for receiving a received signal on the time axis from an echo suppressor at a remote place connected to a communication line, and an input terminal 501 on the time axis received from a far-end person. A receiving speaker 519 for converting the received signal into the amplified voice of the far end and outputting the same, and a transmitting microphone 52 for converting the sound of the near end into a voice signal for input.
0, and an output terminal 502 for transmitting a transmission signal on a time axis to an echo suppressing device at a remote place connected to a communication line.

The echo suppressor 500 further divides the received signal on the time axis received from the far end and calculates the frequency component of the received signal on the frequency axis by the first frequency divider 5.
03, and the second frequency divider 504 for dividing the transmission signal on the time axis and calculating the frequency component of the transmission signal on the frequency axis.
On the frequency axis divided by the first frequency divider 503, the reception noise suppressing unit 505 suppressing noise components included in the frequency components of the reception signal on the frequency axis divided by the first frequency divider 503, and on the frequency axis divided by the second frequency divider 504. A transmission noise suppression unit 511 that suppresses a noise component included in the frequency component of the transmission signal, and a double talk that determines whether or not the reception signal on the frequency axis and the transmission signal on the frequency axis include a voice component. Judgment device 510
And a transfer function and a reception signal on the frequency axis in a section until the voice sound of the far end speaker, which is amplified from the receiving speaker 519, is reflected by an object around the near end person including the near end person and reaches the transmitting microphone 520. The echo component included in the transmission signal on the frequency axis input to the transmission microphone 520 is estimated from the multiplication of ## EQU1 ## and is estimated from the transmission signal on the frequency axis in which the noise component is suppressed by the transmission noise suppression unit 511. Echo canceller unit 51 for removing the echo component and removing the echo component
6, an echo suppression unit 517 that suppresses the echo component remaining without being removed by the echo canceller unit 411, and a transmission signal on the time axis where the echo component is suppressed from the transmission signal on the frequency axis where the echo component is suppressed. And a frequency synthesizer 518 for synthesizing speech signals.

The double talk determiner 510 compares the amplitude value of the reception signal on the frequency axis calculated by the amplitude calculator 506 of the reception noise suppressor with the noise amplitude estimation value estimated by the noise amplitude estimator 507. Further, the amplitude value of the transmission signal on the frequency axis by the amplitude calculator 512 of the transmission noise suppressing unit is compared with the noise amplitude estimation value estimated by the noise amplitude estimator 513, and the reception signal of the reception signal on the frequency axis is compared. Whether the far-end voice component and the near-end voice component of the transmission signal on the frequency axis are simultaneously included, and whether the far-end voice signal includes the far-end voice component. It is designed to judge the heel.

Next, the operation of the echo suppressor of the fifth embodiment will be described with reference to the flowchart of FIG.

The reception signal on the time axis transmitted from the echo suppressing device at the remote place connected to the communication line is converted into the voice of the far end person and output by the reception speaker 519. The sound of the far-end person, which is expanded from the receiving speaker 519, propagates in the space of the near-end person side and the transmission microphone 52.
On the time axis where the voice signal of the near-end person and the noise signal of the near-end person side and the echo signal are superposed by the transmitting microphone 520 together with the voice of the near-end person and the noise of the near-end side as an echo. Is converted into a transmission signal and input.

Here, the operation of calculating the frequency component of the reception signal on the frequency axis processed in the first frequency division step S520 is the same as the above-mentioned first frequency division step S120, S22.
0, S320, and S420 are the same as the operation of calculating the frequency component of the received signal on the frequency axis, and
The operation of calculating the frequency component of the transmission signal on the frequency axis processed in the second frequency division step S521 is the same as the above-described second frequency division step S121, S221, S321, S421.
The operation is the same as the operation of calculating the frequency component of the transmission signal on the frequency axis processed in (1). Further, the operation of removing the echo component superimposed on the reception signal on the frequency axis processed in the echo cancellation step S525 is performed by the reception signal on the frequency axis processed in the echo cancellation steps S222, S323, and S424 described above. The operation is the same as the operation of removing the superimposed echo component, and the operation of suppressing the echo component processed in the echo suppression step S526 is the same as the operation of suppressing the echo component processed in the echo suppression steps S122, S223, S324, and S425. It is the same as the operation of suppressing. Further, the operation of synthesizing the transmission signal on the time axis from the frequency components of the transmission signal on the frequency axis processed in the frequency synthesis step S527 is performed by the above-described frequency synthesis steps S123, S223.
The operation is the same as the operation of synthesizing the transmission signal on the time axis from the frequency components of the transmission signal on the frequency axis processed in S325 and S426, and the operation is performed in the first to fourth embodiments. Since it has been described, the description in this embodiment will be omitted.

Next, the operation of the double talk determiner 510 will be described with reference to the flowchart of FIG.

In step S801, the amplitude value X1 (k) of the frequency component of the reception signal on the frequency axis and the amplitude value a × N1 of the frequency component of the noise signal multiplied by a predetermined constant a.
(K) is compared, and the amplitude value X1 of the frequency component of the reception signal on the frequency axis larger than the amplitude value a × N1 (k) of the frequency component of the noise signal multiplied by the predetermined constant a is calculated.
When (k) is detected, it is determined that the voice signal of the far end person is detected. Here, the symbol k is a frequency band number. Also, by multiplying by a predetermined constant a,
This is to prevent a reaction due to the influence of minute fluctuations in the noise signal.

Similarly, in step S802, the magnitude of the amplitude value X2 (k) of the transmission signal on the frequency axis and the amplitude value b × N2 (k) of the noise signal multiplied by a predetermined constant b are calculated. By comparison, the amplitude value X2 (k) of the frequency component of the transmission signal on the frequency axis, which is larger than the amplitude value b × N2 (k) of the frequency component of the noise signal multiplied by the predetermined constant b, is detected. At times, it is determined that the voice signal of the near end person has been detected. Further, it is to prevent the reaction from being influenced by the minute fluctuation of the noise signal by multiplying the predetermined constant b.

Next, in step S803, the far-end voice component is detected in the received signal on the frequency axis in step S801, and the near-end voice component is detected in the transmitted signal on the frequency axis in step S802. When
A flag is set to signal the decision to the process of updating the transfer function. On the other hand, when the voice component of the far-end party is not detected in the reception signal on the frequency axis in step S801, or the voice component of the far-end party is detected in the reception signal on the frequency axis in step S801, and step S8
When the voice component of the near-end person is not detected in the transmission signal on the frequency axis at 02, the flag is reset to notify the determination to the step of updating the next transfer function. The steps S801 to S804 are collectively referred to as a double talk determination step S524.

As described above, according to the present embodiment, after suppressing the noise component contained in the reception signal on the frequency axis on the frequency axis, the amplitude value of the frequency component of the pseudo echo signal on the frequency axis is suppressed. And the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the received signal on the frequency axis, and the echo component included in the received signal on the frequency axis is suppressed. However, the residual echo does not increase, and sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

(Embodiment 6) FIG. 13 is a block diagram showing the configuration of an echo suppressor according to a sixth embodiment of the present invention. This echo suppressor is installed in a plurality of remote locations. The present invention is applied to a communication system that is connected to another echo suppressor through a communication line and performs voice communication such as a telephone and a videophone system.

Echo suppressor 600 shown in FIG.
Is an input terminal 601 for receiving a received signal on the time axis from an echo suppressor at a remote place connected to a communication line, and a voice of the far-end person who has amplified the received signal on the time axis received from the far-end person. A receiver speaker 624 for converting and outputting to a voice, a transmitting microphone 625 for converting a voice of a near end person into a voice signal and inputting the voice, and a transmitter on a time axis for transmitting to an echo suppressor at a remote place connected to a communication line. And an output terminal 602 for transmitting a signal.

The echo suppressor 600 further divides the received signal on the time axis received from the far end person and calculates the frequency component of the received signal on the frequency axis.
03, and a second frequency divider 604 that divides the transmission signal on the time axis and calculates the frequency component of the transmission signal on the frequency axis.
A transmission noise suppression unit 605 that suppresses a noise component included in the frequency component of the reception signal on the frequency axis divided by the first frequency divider 603; and on the frequency axis divided by the second frequency divider 604. Transmission noise suppressing section 610 for suppressing the noise component contained in the frequency component of the transmission signal of, the voice component of the far end included in the reception signal on the frequency axis, and the near component included in the transmission signal on the frequency axis. A double-talk determiner 615 that detects the voice component of the proband and determines whether or not the voice is in the double-talk state, a reception signal on the frequency axis in which the noise signal is suppressed by the reception noise suppression unit 605, and transmission noise suppression The transfer function updater 617 that updates the transfer function based on the comparison judgment with the transmission signal on the frequency axis where the noise signal is suppressed by the unit 610, and the frequency at which the noise signal is suppressed by the reception noise suppression unit 605. A transfer function multiplier 616 that calculates a pseudo echo signal on the frequency axis by multiplying the above reception signal by a transfer function, and an echo that suppresses the echo component included in the transmission signal on the frequency axis that remains without being removed. Suppression unit 618
And a frequency synthesizer 623 for synthesizing the transmission signal on the time axis from the frequency component of the transmission signal on the frequency axis in which the echo component is suppressed.

Further, the echo suppressing unit 618 calculates the amplitude value of the pseudo echo signal on the frequency axis calculated by the transfer function multiplier 616, and the noise suppressing unit 610 suppresses the noise component. A second amplitude calculator 620 for calculating the amplitude value of the transmitted signal on the frequency axis,
An echo suppression coefficient calculator 621 that calculates an echo suppression coefficient based on a comparison determination between the amplitude value of the pseudo echo signal on the frequency axis and the amplitude value of the reception signal on the frequency axis, and the noise suppression unit 6
An echo suppression coefficient multiplier 62 that multiplies the received signal on the frequency axis where the noise signal is suppressed by 10 by the echo suppression coefficient
2 is provided, and the echo component included in the received signal on the frequency axis is suppressed.

Further, the bull talk determiner 615 provided in the echo suppressor 600 estimates the amplitude of the received signal on the frequency axis calculated by the amplitude calculator 606 of the received noise suppressor 605 and the noise amplitude estimator 607. The comparison between the noise amplitude estimated value and the noise amplitude estimated value estimated by the noise amplitude estimator 612 is performed between the amplitude value of the transmission signal on the frequency axis by the amplitude calculator 611 of the transmission noise suppression unit 610 and the noise amplitude estimated value estimated by the noise amplitude estimator 612. Based on the comparison judgment, it is determined whether the far-end voice component of the reception signal on the frequency axis and the near-end voice component of the transmission signal on the frequency axis are detected at the same time, and on the frequency axis. It is determined whether or not only the voice component of the far-end party of the received voice signal is detected, and the determination result is output to the transfer function updater 617. When the far-end voice component of the received signal on the frequency axis and the near-end voice component of the transmitted signal on the frequency axis are detected simultaneously, and the far-end voice component of the received signal on the frequency axis Is detected, the transfer function update operation becomes unstable, so the transfer function updater 61
7 prevents the transfer function from being updated.

Next, the operation of the echo suppressor of the sixth embodiment will be described with reference to the flowchart of FIG.

The reception signal on the frequency axis transmitted from the echo suppressor at a remote place connected to the communication line is converted into the voice of the far end person and output by the reception speaker 624. The voice of the far-end person, which is expanded from the receiving speaker 624, propagates in the space of the near-end person side and the transmitting microphone 6
25, and the voice of the near-end person and the noise of the near-end person side as an echo, along with the voice signal of the near-end person, the noise signal of the near-end person side, and the echo signal are superposed by the transmission microphone 625. Is converted into a transmission signal and input.

Here, the operation of calculating the frequency component of the reception signal on the frequency axis processed in the first frequency division step S620 is the same as the above-mentioned first frequency division step S120, S22.
0, S320, S420, and S520, the same operation as calculating the frequency component of the reception signal on the frequency axis, and the frequency of the transmission signal on the frequency axis processed in the second frequency division step S621. The operation of calculating the component is performed by the above-described second frequency division step S121, S221, S32.
The operation is the same as the operation of calculating the frequency component of the transmission signal on the frequency axis processed in S1, S421, and S521. Furthermore, the operation of determining the double-talk state processed in the double-talk determination step S624 is the same as the operation of determining the double-talk state processed in the double-talk determination step S524 described above, and further, in the echo suppression step S627. The operation of suppressing the echo component to be processed is performed by the above-described echo suppression steps S122, S223, S324, S425, S.
This is the same as the operation of suppressing the echo component processed in 526. Furthermore, the operation of synthesizing the transmission signal on the time axis from the frequency components of the transmission signal on the frequency axis processed in the frequency synthesis step S628 is performed by the above-described frequency synthesis step S123,
The operation is the same as the operation of synthesizing the transmission signal on the time axis from the frequency components of the transmission signal on the frequency axis processed in S223, S325, S426, and S527, and is the same as that of the first to fifth embodiments. Since the operation has been described in the form, the description in the present embodiment will be omitted.

Next, the operation of the transfer function updating step S625 will be described with reference to the flowchart in FIG.

In the transfer function updating step S625, the amplitude value X2 (k) of the transmission signal on the frequency axis, the reception speaker 624, and the transmission microphone 625 are determined based on the comparison determination in the double talk determination step S624. The product H (k) of the transfer function H (k) and the amplitude value X1 (k) of the received signal on the frequency axis
Amplitude value comparison step S901 for comparing and determining X1 (k)
And the amplitude value X2 (k) of the transmission signal on the frequency axis is the transfer function H (k) between the receiving speaker 624 and the transmission microphone 625 and the amplitude value X1 of the transmission signal on the frequency axis.
A first transfer function amplitude value updater S902 that updates the amplitude value of the transfer function when it is larger than the product H (k) × X1 (k) of (k), and the amplitude value X2 ( k)
Is smaller than the product H (k) × X1 (k) of the transfer function H (k) between the receiving speaker 624 and the transmitting microphone 625 and the amplitude value X1 (k) of the transmitting signal on the frequency axis. And a second transfer function amplitude value updater S903 that sometimes updates the amplitude value of the transfer function. Here, the symbol k is the number of the frequency band put together when calculating the amplitude value.

Next, the steps after the transfer function updating in the transfer function updating step S625 will be described.

First, in the amplitude value comparing step S901, the amplitude value H (k) of the transfer function preset in the transfer function multiplier 616 and the amplitude value X1 of the reception signal on the frequency axis are set.
The product with (k) is the amplitude value X2 of the transmission signal on the frequency axis.
The transfer function is not updated when it is determined to be equal to (k). The product of the amplitude value H (k) of the transfer function preset in the transfer function multiplier 616 and the amplitude value X1 (k) of the reception signal on the frequency axis is the amplitude of the transmission signal on the frequency axis. If it is determined that the value is not more than the value X2 (k), the process proceeds to the first transfer function amplitude value updating step S902. On the other hand, the amplitude value H of the transfer function preset in the transfer function multiplier 616
When it is determined that the product of (k) and the amplitude value X1 (k) of the reception signal on the frequency axis is larger than the amplitude value X2 (k) of the transmission signal on the frequency axis, the transfer function amplitude value updating step is performed. move on.

Then, the first transfer function amplitude value updating step S9
In 02, the product of the amplitude value H (k) of the transfer function preset in the transfer function multiplier 616 and the amplitude value X1 (k) of the reception signal on the frequency axis is the amplitude of the transmission signal on the frequency axis. When it is determined that the value is smaller than the value X2 (k), the product k1 × H (k) of the amplitude value H (k) of the transfer function and a positive constant k1 larger than 1 determined in advance is calculated and updated. Stored as a transfer function. On the other hand, in the transfer function amplitude value updating step, the product of the amplitude value H (k) of the transfer function preset in the transfer function multiplier 616 and the amplitude value X1 (k) of the reception signal on the frequency axis is the frequency axis. Amplitude value X of the transmitted signal above
When it is determined that it is larger than 2 (k), the product k2 × H (k) of the amplitude value H (k) of the transfer function and a predetermined positive constant k2 smaller than 1 is calculated and updated. It is held as a transfer function.

Next, the transfer function multiplier 616 calculates a pseudo echo signal on the frequency axis based on the updated transfer function, and then the pseudo echo signal on the frequency axis calculated by the transfer function multiplier 616. Based on the above, the echo suppression unit 618 suppresses the echo component included in the transmission signal on the frequency axis.

As described above, according to the present embodiment, after suppressing the noise component contained in the received signal on the frequency axis, the amplitude value of the frequency component of the pseudo echo signal on the frequency axis and the frequency value on the frequency axis are obtained. Since the echo suppression coefficient is calculated based on the amplitude value of the frequency component of the reception signal of, and the echo component included in the reception signal is suppressed on the frequency axis,
Even in the presence of ambient noise, the residual echo does not increase, and sufficient echo suppression performance can be obtained with a smaller processing calculation amount than the processing calculation amount on the time axis.

(Embodiment 7) FIG. 11 shows the eighth embodiment of the present invention.
2 is a block diagram showing the configuration of the embodiment of FIG.
Reference numeral 101 denotes a receiving unit 1102 that receives a reception signal on the time axis transmitted from the far end, a decoder 1103 that decodes the received encoded signal, and a noise component included in the transmission signal on the frequency axis. An echo suppressor 1104 for suppressing an echo component and a first amplifier 1105 for amplifying a received signal
, A receiving speaker 1106 for converting a reception signal on the time axis into a voice of the far end person and outputting the same, and a transmission microphone 110 for converting a voice signal of the near end person into a reception signal on the time axis and inputting.
7 and the second amplifier 120 for amplifying the transmission signal on the time axis
8, an encoder 1109 that encodes the amplified transmission signal on the time axis, and a transmission unit 1110 that transmits the transmission signal on the time axis.

Next, the operation of the eighth embodiment of the echo suppressor of the invention will be described.

The received signal received by wireless or wire is converted by the receiving section 1102 into a processable signal. The encoded signal which is the output of the receiving unit 1102 is decoded by the decoder 1103. The output signal of the decoder 1103 becomes one input of the echo suppressor 1104. The received signal once input to the echo suppressor 1104 is amplified by the first amplifier 1105,
The speaker 1106 radiates the sound into the space. To the microphone 1107, the sound of the far end person, the sound of the near end person, and the like output from the speaker 1106 are input.
The transmission signal input to the microphone 1107 is amplified by the amplifier 1108 and becomes the other input of the echo suppressor 1104. The signal whose echo is suppressed by the echo suppressor 1104 is encoded by the encoder 1109 and transmitted by the transmitter 1110.

As described above, according to the present embodiment, after suppressing the noise component included in the received signal on the frequency axis, the amplitude value and the frequency axis of the pseudo echo signal on the frequency axis on the frequency axis are suppressed. The echo suppression coefficient is calculated based on the amplitude value of the received signal above, and the echo component included in the received signal is suppressed on the frequency axis, so that the residual echo increases even in the presence of ambient noise. Without doing so, the echo component can be suppressed with a smaller processing calculation amount than the processing calculation amount on the time axis, and a clear transmission signal on the frequency axis that does not include the echo component can be transmitted.

(Embodiment 8) FIG. 12 is a block diagram showing an echo suppressor according to an eighth embodiment of the present invention. The videophone system 1201 is a far-end party transmitted from a far-end party. A receiving unit 1202 that receives the side video signal and the reception signal on the time axis, a video display unit 1203 that displays the far-end side video signal on the screen, and a noise component and an echo included in the transmission signal on the time axis. Echo suppressor 1204 for suppressing the component and first amplifier 1205 for amplifying the received signal
And a receiving speaker 1206 for converting the received signal into the voice of the far-end and outputting the same, a transmitting microphone 1207 for converting the voice of the near-end into the transmitting signal on the time axis and inputting, and the near-end side A camera 1208 for converting a captured image into a video signal and inputting it, a second amplifier 1209 for amplifying and outputting a transmission signal on the time axis, a video signal on the near-end person side, and a transmission on the time axis And a transmitter 1210 for transmitting a signal.

Next, the operation of the eighth embodiment of the echo suppressor of the invention will be described.

The signal received from the far-end party via a wireless or wired communication line is converted by the receiving unit 1202 into a far-end side image signal that can be processed and a reception signal on the time axis. The far-end party side video signal converted by the receiving unit 1202 is displayed by the video display unit 1203.
Displayed on the screen. On the other hand, the received signal on the time axis converted by the receiving unit 1202 is the echo suppressor 120.
4 is input. Furthermore, the reception signal on the time axis is amplified by the first amplifier 1205, converted into the voice of the far end person by the reception speaker 1206, and is amplified in the space of the near end person. The amplified sound of the far-end person propagates in the space on the near-end person side, is reflected by an object around the near-end person including the near-end person, and reaches the transmitting microphone. It is converted into a transmission signal on the time axis and input together with the noise around the near-end person. The input transmission signal on the time axis is amplified by the second amplifier 1209 and input to the echo suppressor 1204. Echo suppressor 1204
The transmission signal on the time axis in which the echo is suppressed by is input to the transmission unit 1210. On the other hand, a video image taken around the near-end person by the camera 1208 is converted into a video signal,
It is input to the transmission unit 1210. The transmission signal and the video signal on the time axis are converted into a transmittable signal by the transmission unit 1210 and transmitted to the far end person through the communication line.

As described above, according to the present embodiment, after suppressing the noise component contained in the received signal on the frequency axis, the amplitude value and frequency axis of the pseudo echo signal on the frequency axis on the frequency axis are suppressed. The echo suppression coefficient is calculated based on the amplitude value of the received signal above, and the echo component included in the received signal is suppressed on the frequency axis, so that the residual echo increases even in the presence of ambient noise. Without doing so, the echo component can be suppressed with a smaller processing calculation amount than the processing calculation amount on the time axis, and a clear transmission signal on the frequency axis that does not include the echo component can be transmitted.

[0158]

As described above, according to the present invention, the noise suppression processing for suppressing the noise component contained in the reception signal on the frequency axis and the transmission signal on the frequency axis, and the transmission signal on the frequency axis. Echo removal processing to remove the echo component contained in
By performing echo suppression processing that suppresses the echo component of the transmission signal remaining on the frequency axis that cannot be removed in the frequency domain, it is possible to adequately follow fluctuations in the transfer function between the reception speaker and the transmission microphone. In addition, by executing the noise component suppression in the preceding stage of the echo removal processing, even when ambient noise is present, without increasing the residual echo,
It is possible to provide an echo suppressor that suppresses an echo component in a short time.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a first embodiment of an echo suppressor according to the present invention.

FIG. 2 is a flowchart showing the operation of the echo suppressor according to the first embodiment.

FIG. 3 is a block diagram showing a configuration of a second embodiment of an echo suppressor according to the present invention.

FIG. 4 is a flowchart showing an operation of the echo suppressor according to the second embodiment.

FIG. 5 is a block diagram showing a configuration of a third embodiment of an echo suppressor according to the present invention.

FIG. 6 is a flowchart showing an operation of the echo suppressor according to the third embodiment.

FIG. 7 is a flowchart showing the operation of a noise estimation step according to the third embodiment.

FIG. 8 is a block diagram showing a configuration of a fourth embodiment of an echo suppressor according to the present invention.

FIG. 9 is a flowchart showing the operation of the echo suppressor according to the fourth embodiment.

FIG. 10 is a block diagram showing a configuration of a fifth embodiment of an echo suppressor according to the present invention.

FIG. 11 is a flowchart showing an operation of the echo suppressor according to the fifth embodiment.

FIG. 12 is a flowchart showing the operation of the double talk determination process of the fifth embodiment.

FIG. 13 is a block diagram showing a configuration of a sixth embodiment of an echo suppressor according to the present invention.

FIG. 14 is a flowchart showing the operation of the echo suppressor according to the sixth embodiment.

FIG. 15 is a flowchart showing the operation of a transfer function amplitude value updating step according to the sixth embodiment.

FIG. 16 is a block diagram showing a configuration of a seventh embodiment of an echo suppressor according to the present invention.

FIG. 17 is a block diagram showing the configuration of an eighth embodiment of an echo suppressor according to the present invention.

FIG. 18 is a block diagram showing the configuration of a conventional acoustic echo canceller.

FIG. 19 is a block diagram showing a configuration of a conventional noise suppression device.

[Explanation of symbols]

100, 200, 300, 400, 500, 600 Echo suppressors 101, 201, 301, 401, 501, 601 Input terminals 102, 202, 302, 402, 502, 602 Output terminals 112, 216, 313, 414, 519, 624, 1
106, 1206 Earpiece speakers 113, 217, 314, 415, 520, 625, 1
107, 1207 transmission microphone 103, 203, 303, 403, 503, 603 frequency divider 104, 204, 304, 404, 504, 604 frequency divider 111, 215, 312, 413, 518, 623 frequency synthesizer 106, 210, 311, 412, 517, 618 Echo suppression section 107, 211, 619 First amplitude calculator 108, 212, 620 Second amplitude calculator 109, 213, 621 Coefficient calculator 110, 214, 622 Echo suppression coefficient multiplier 205, 310, 411, 516 Echo canceller section 206, 617 Transfer function updater 105, 616 Transfer function multiplier 208 Subtractor 305, 410, 511, 610 Transmission noise suppression section 405, 505, 610 Reception noise suppression section 306, 406, 506, 512, 606, 611 amplitude Calculators 308, 408, 508, 514, 608, 613 Coefficient calculators 307, 407, 507, 513, 607, 612 Noise amplitude estimators 309, 409, 509, 515, 609, 614 Noise suppression coefficient multipliers 510, 615 Double talk determiner 1101 Telephone 1102, 1202 Receiver 1103 Decoder 1104, 1204 Echo suppressor 1105, 1108, 1205, 1209 Amplifier 1109 Encoder 1110, 1210 Transmitter 1201 Videophone system 1203 Video display 1208 Camera S701, S801, S802, S901 Amplitude value comparison step S702 Noise amplitude value update step S703 Noise amplitude value replacement step S803, S804 Flag setting step S902, S903 Transfer function amplitude value update step S120, S220, S320 , S420, S520,
S620 First frequency division step S121, S221, S321, S421, S521,
S621 Second frequency division step S422, S522 Received noise suppression step S322, S423, S523 Transmitted noise suppression step S524, S624 Double talk determination step S222, S323, S424, S525 Echo cancellation step S122, S223, S324, S425, S526,
S626 Echo suppression steps S123, S223, S325, S426, S527,
S628 Frequency synthesis process

Claims (26)

[Claims] 1. A transmission speaker for converting a reception signal on a time axis received from a communication line into a voice of a far end person and outputting the voice.
A reception microphone that converts the voice of the near-end person into a transmission signal on the time axis and inputs the same; a first frequency divider that divides the reception signal on the time axis and calculates the reception signal on the frequency axis; A transfer function multiplier for calculating a pseudo echo signal on the frequency axis by multiplying the transfer function between the transmission speaker and the reception microphone by the reception signal on the frequency axis, and dividing the transmission signal on the time axis. A second frequency divider for calculating a transmission signal on the frequency axis, and the frequency received by the reception microphone based on the pseudo echo signal on the frequency axis and the transmission signal on the frequency axis. An echo suppressor that suppresses an echo component of an on-axis transmission signal, and a transmission signal on the time axis in which the echo component is suppressed is synthesized from a transmission signal on the frequency axis in which the echo component is suppressed. Frequency synthesizer Echo suppressing apparatus characterized by obtaining. 2. A transmitting speaker for converting a received signal on a time axis received from a communication line into a voice of a far end person and outputting the voice.
A reception microphone that converts the voice of the near-end person into a transmission signal on the time axis and inputs the same; a first frequency divider that divides the reception signal on the time axis and calculates the reception signal on the frequency axis; A second frequency divider that divides the transmission signal on the time axis and calculates the transmission signal on the frequency axis, and estimates the transfer function between the transmission speaker and the reception microphone to estimate the transmission function on the frequency axis. An echo canceller unit for calculating a pseudo echo signal, subtracting the pseudo echo signal on the frequency axis from the transmission signal on the frequency axis to remove an echo component of the transmission signal on the frequency axis, and the frequency axis Of the transmission signal on the frequency axis remaining without being removed by the echo canceller based on the transmission signal on the frequency axis and the pseudo echo signal on the frequency axis from which the pseudo echo signal is subtracted Suppresses the echo component An echo suppressor, and a frequency synthesizer for synthesizing a transmission signal on the time axis in which the echo component is suppressed from a transmission signal on the frequency axis in which the echo component is suppressed, Echo suppressor. 3. A transmission speaker for converting a reception signal on a time axis received from a communication line into a voice of a far end person and outputting the voice.
A reception microphone that converts the voice of the near-end person into a transmission signal on the time axis for input, and a first frequency divider that divides the reception signal on the time axis and calculates the reception signal on the frequency axis. A second frequency divider that divides the transmission signal on the time axis and calculates the transmission signal on the frequency axis; and a transmission noise suppression unit that suppresses noise components of the transmission signal on the frequency axis. And a pseudo echo signal on a frequency axis is calculated by estimating a transfer function between the transmitting speaker and the receiving microphone, and a pseudo echo signal on the frequency axis is calculated from the transmission signal on the frequency axis in which the noise component is suppressed. An echo canceller unit for subtracting a pseudo echo signal to remove an echo component of a transmission signal on the frequency axis, the frequency axis from which the pseudo echo signal on the frequency axis and the pseudo echo signal on the frequency axis are subtracted Based on the sending signal above An echo suppressor for suppressing the echo component of the transmission signal on the frequency axis remaining without being removed by the echo canceller, and the echo component from the transmission signal on the frequency axis in which the echo component is suppressed. And a frequency synthesizer for synthesizing the suppressed transmission signal on the time axis. 4. A transmission speaker for converting a reception signal on a time axis received from a communication line into a voice of a far end person and outputting the voice.
A reception microphone that converts the voice of the near-end person into a transmission signal on the time axis for input, and a first frequency divider that divides the reception signal on the time axis and calculates the reception signal on the frequency axis. A second frequency divider that divides the transmission signal on the time axis and calculates the transmission signal on the frequency axis, and a reception noise suppression unit that suppresses noise components of the reception signal on the frequency axis, A transmission noise suppression unit that suppresses a noise component of the transmission signal on the frequency axis, a transfer function between the transmission speaker and the reception microphone is estimated to calculate a pseudo echo signal on the frequency axis, and An echo canceller unit for subtracting a pseudo echo signal on the frequency axis from a transmission signal on the frequency axis in which a noise component is suppressed to remove an echo component of the transmission signal on the frequency axis, and the echo canceller unit. Remains without being removed by An echo suppressor for suppressing the echo component of the transmission signal on the frequency axis, and a transmission on the time axis with the echo component suppressed from the transmission signal on the frequency axis with the echo component suppressed An echo suppressor comprising: a frequency synthesizer for synthesizing signals. 5. A transmission speaker for converting a reception signal on a time axis received from a communication line into a voice of a far end person and outputting the voice.
A reception microphone that converts the voice of the near-end person into a transmission signal on the time axis for input, and a first frequency divider that divides the reception signal on the time axis and calculates the reception signal on the frequency axis. A second frequency divider that divides the transmission signal on the time axis and calculates the transmission signal on the frequency axis, and a reception noise suppression unit that suppresses noise components of the reception signal on the frequency axis, A transmission noise suppressing unit for suppressing a noise component of the transmission signal on the frequency axis, a voice component of the far end of the reception signal on the frequency axis, and a voice of the near end of the transmission signal on the frequency axis. A double-talk determiner that detects a component and determines whether or not it is in a double-talk state, a transfer function between the transmitting speaker and the receiving microphone is estimated to calculate a pseudo echo signal, and the noise component is suppressed. From the transmission signal on the frequency axis to the frequency axis Of the echo signal of the transmission signal on the frequency axis by subtracting the pseudo echo signal of, and the echo component of the transmission signal on the frequency axis remaining without being removed by the echo canceller unit And a frequency synthesizer for synthesizing a transmission signal on the time axis in which the echo component is suppressed from a transmission signal on the frequency axis in which the echo component is suppressed, Characteristic echo suppressor. 6. A transmitting speaker for converting a received signal on a time axis received from a communication line into a voice of a far end person and outputting the voice.
A reception microphone that converts the voice of the near-end person into a transmission signal on the time axis for input, and a first frequency divider that divides the reception signal on the time axis and calculates the reception signal on the frequency axis. A second frequency divider that divides the transmission signal on the time axis and calculates the transmission signal on the frequency axis, and a reception noise suppression unit that suppresses noise components of the reception signal on the frequency axis, A transmission noise suppressing unit for suppressing a noise component of the transmission signal on the frequency axis, a voice component of the far end of the reception signal on the frequency axis, and a voice of the near end of the transmission signal on the frequency axis. On the frequency axis by multiplying the transfer function between the transmitting speaker and the receiving microphone and the receiving signal on the frequency axis by detecting the component and determining whether or not the state is the double talk state. A transfer function multiplier for calculating a pseudo echo signal; A transfer function updater for updating the transfer function based on the determination of the Routh determiner, and an echo suppressor for suppressing the echo component of the transmission signal on the frequency axis remaining without being removed by the echo canceller. An echo suppressor comprising: a frequency synthesizer for synthesizing a transmission signal on the time axis in which the echo component is suppressed from a transmission signal on the frequency axis in which the echo component is suppressed. 7. The echo suppressor includes a first amplitude calculator that calculates an amplitude value of the pseudo echo signal on the frequency axis,
A second amplitude calculator for calculating an amplitude value of the transmission signal on the frequency axis, and a second amplitude calculator based on a comparison determination of the amplitude value of the pseudo echo signal on the frequency axis and the amplitude value of the transmission signal on the frequency axis. And a coefficient calculator for calculating an echo suppression coefficient, and a multiplier for multiplying the transmission signal on the frequency axis by the echo suppression coefficient to suppress the echo component of the transmission signal on the frequency axis. The echo suppressor according to claim 1. 8. The echo suppressor includes a first amplitude calculator that calculates an amplitude value of a pseudo echo signal on the frequency axis,
A second amplitude calculator for calculating the amplitude value of the transmission signal on the frequency axis from which the pseudo echo signal on the frequency axis is subtracted; the amplitude value of the pseudo echo signal on the frequency axis and the frequency value on the frequency axis A coefficient calculator that calculates an echo suppression coefficient based on a comparison judgment with the amplitude value of the transmission signal on the frequency axis on which the pseudo echo signal is subtracted, and the frequency on which the pseudo echo signal on the frequency axis is subtracted A multiplier for multiplying the amplitude value of the transmission signal on the axis by the echo suppression coefficient and suppressing the echo component of the transmission signal on the frequency axis that remains without being removed. The echo suppressor according to any one of claims 2 to 6. 9. The echo canceller unit includes a transfer function updater for updating a transfer function between the transmitting speaker and the receiving microphone, and a received signal on the frequency axis multiplied by the transfer function to obtain the frequency axis. A first multiplier for calculating the pseudo echo signal, and a pseudo echo signal on the frequency axis is subtracted from the transmission signal on the frequency axis to remove the echo component of the transmission signal on the frequency axis. A second multiplier for multiplying the reception signal on the frequency axis by the transmission signal on the frequency axis from which the echo component has been removed and outputting the result to the transfer function updater. The echo suppressor according to claim 2. 10. The echo canceller unit includes a transfer function updater for updating a transfer function between the transmitting speaker and the receiving microphone, and a received signal on a frequency axis multiplied by the transfer function to multiply the frequency on the frequency axis. Of the pseudo echo signal on the frequency axis, and the pseudo echo signal on the frequency axis is subtracted from the transmission signal on the frequency axis in which noise components are suppressed. A subtractor for removing the echo component, and a second multiplier for multiplying the reception signal on the frequency axis by the transmission signal on the frequency axis from which the echo component is removed and outputting the result to the transfer function updater. The echo suppressor according to claim 3, further comprising: 11. The echo canceller unit includes a transfer function updater for updating a transfer function between the transmitting speaker and the receiving microphone, and the transfer function for the receiving signal on the frequency axis in which noise components are suppressed. A first multiplier for calculating a pseudo echo signal on the frequency axis by multiplication, and a pseudo echo signal on the frequency axis for subtracting the pseudo echo signal on the frequency axis from the transmission signal on the frequency axis; A subtracter that removes the echo component, and a second multiplier that multiplies the reception signal on the frequency axis and the transmission signal on the frequency axis from which the echo component is removed, and outputs the product to the transfer function updater. The echo suppressor according to claim 4, further comprising: 12. The echo canceller unit includes a transfer function updater for updating a transfer function between the transmitting speaker and the receiving microphone, and the transfer function for the receiving signal on the frequency axis in which noise components are suppressed. A first multiplier that multiplies to calculate a pseudo echo signal on the frequency axis; and a frequency multiplier that subtracts the pseudo echo signal on the frequency axis from a transmission signal on the frequency axis in which noise components are suppressed. A subtracter for removing the echo component of the above transmission signal, a reception signal on the frequency axis and a transmission signal on the frequency axis from which the echo component has been removed are multiplied to the transfer function updater. 6. A second multiplier for outputting is provided.
The echo suppressor according to. 13. The reception noise suppression unit includes an amplitude calculator that calculates an amplitude value of a reception signal on the frequency axis and a noise amplitude estimator that estimates an amplitude value of a noise component of the reception signal on the frequency axis. And a coefficient calculator that calculates a noise suppression coefficient based on a comparison determination of the amplitude value of the reception signal on the frequency axis and the amplitude value of the noise component of the reception signal on the frequency axis,
7. A multiplier for multiplying the received signal on the frequency axis by the noise suppression coefficient to suppress the noise component of the received signal on the frequency axis, according to any one of claims 4 to 6. The echo suppressor described. 14. The amplitude value comparator for judging whether or not the amplitude value of the noise component of the reception signal on the frequency axis is larger than the amplitude value of the reception signal on the frequency axis. And, when it is determined that the amplitude value of the noise component of the reception signal on the frequency axis is less than or equal to the amplitude value of the reception signal on the frequency axis, the frequency axis is set to the amplitude value of the noise component of the reception signal on the frequency axis. When it is determined that the noise amplitude value interchanger that replaces the amplitude value of the receiving signal on the above, and the amplitude value of the noise component of the receiving signal on the frequency axis is larger than the amplitude value of the receiving signal on the frequency axis, The noise suppression apparatus according to claim 13, further comprising a noise amplitude value updater that updates the amplitude value of the noise component of the received signal on the frequency axis. 15. The transmission noise suppressing unit calculates an amplitude value of a transmission signal on the frequency axis, and an amplitude calculator on the frequency axis based on the amplitude value of the transmission signal on the frequency axis. Of a noise amplitude estimator for estimating the amplitude value of the noise signal on the frequency axis included in the transmission signal of, and the amplitude value of the transmission signal on the frequency axis and the amplitude value of the noise signal on the frequency axis The coefficient calculator that calculates a noise suppression coefficient based on a comparison determination, and the multiplier that suppresses the noise component of the transmission signal on the frequency axis by multiplying the transmission signal on the frequency axis by the noise suppression coefficient The echo suppressor according to any one of claims 3 to 6, further comprising: 16. The noise amplitude estimator compares and determines whether or not the amplitude value of the noise component of the transmission signal on the frequency axis is larger than the amplitude value of the transmission signal on the frequency axis. A comparator and the amplitude of the noise component of the transmission signal on the frequency axis when it is determined that the amplitude value of the noise component of the transmission signal on the frequency axis is less than or equal to the amplitude value of the transmission signal on the frequency axis. A noise amplitude value transposition device that replaces the amplitude value of the transmission signal on the frequency axis with a value, and the amplitude value of the noise component of the transmission signal on the frequency axis is greater than the amplitude value of the transmission signal on the frequency axis. 16. The echo suppressor according to claim 15, further comprising: a noise amplitude value updater that updates the amplitude value of the noise component of the transmission signal on the frequency axis when it is determined that it is also large. 17. The double talk determiner compares an amplitude value of a reception signal on the frequency axis with an amplitude estimation value of a noise component of the reception signal on the frequency axis to obtain a reception signal on the frequency axis. A first amplitude value comparator for determining whether or not a voice component of the far-end person is detected, an amplitude value of a transmission signal on the frequency axis and an estimated amplitude value of a noise component of the transmission signal on the frequency axis. And a second amplitude value comparator for determining whether or not the voice component of the near-end person is detected in the transmission signal on the frequency axis, and the voice of the far-end person for the reception signal on the frequency axis. A first flag setter that sets a first flag when a component is detected and a voice signal of a near end is not detected in a transmission signal on the frequency axis, and outputs the first flag to the echo canceller unit;
When the far-end voice component is not detected in the reception signal on the frequency axis, and when the near-end voice component is detected in the transmission signal on the frequency axis, or in the reception signal on the frequency axis A second flag is set when a voice component of the far end is detected and a voice component of the near end is detected in the transmission signal on the frequency axis and is output to the echo canceller unit. The echo suppressor according to claim 5 or 6, further comprising: 18. The transfer function updater is configured such that an amplitude value of a reception signal on the frequency axis in which the noise component is suppressed is smaller than an amplitude value of a transmission signal on the frequency axis in which the noise component is suppressed. An amplitude value comparator for determining whether or not it is larger,
When the amplitude value of the reception signal on the frequency axis with the noise component suppressed is larger than the amplitude value of the transmission signal on the frequency axis with the noise component suppressed, the amplitude value of the transfer function is updated. A first transfer function amplitude value updater, and when the amplitude value of the reception signal on the frequency axis in which the noise component is suppressed is less than or equal to the amplitude value of the transmission signal on the frequency axis in which the noise component is suppressed, A second for updating the amplitude value of the transfer function
The echo suppressor according to claim 6, further comprising a transfer function amplitude value updater. 19. A first frequency division step of dividing a reception signal on the time axis and calculating a reception signal on the frequency axis, and the reception signal on the frequency axis in a transfer function between a transmitting speaker and a reception microphone. A transfer function multiplication step of multiplying to calculate a pseudo echo signal on the frequency axis; a second frequency division step of dividing the transmission signal on the time axis to calculate a transmission signal on the frequency axis; An echo suppression step of suppressing an echo component of the transmission signal on the frequency axis received by the receiving microphone based on the pseudo echo signal and the transmission signal on the frequency axis, and the echo component is suppressed. And a frequency synthesizing step of synthesizing the transmission signal on the time axis in which the echo component is suppressed from the transmission signal on the frequency axis. 20. A first frequency division step of dividing a reception signal on the time axis to calculate a reception signal on the frequency axis, and dividing a transmission signal on the time axis to generate a transmission signal on the frequency axis. A second frequency division step of calculating, a transfer function between the transmitting speaker and the receiving microphone is estimated to calculate a pseudo echo signal on the frequency axis, and the pseudo echo signal on the frequency axis is calculated from the transmission signal on the frequency axis. An echo canceller process for removing an echo component of a transmission signal on the frequency axis by subtracting a signal, a transmission signal on the frequency axis and a transmission signal on the frequency axis from which a pseudo echo signal on the frequency axis is subtracted An echo suppression step of suppressing the echo component of the transmission signal on the frequency axis remaining without being removed by the echo canceller step based on the pseudo echo signal, and an echo suppression step on the frequency axis in which the echo component is suppressed. Talk A frequency synthesizing step of synthesizing a transmission signal on the time axis in which the echo component is suppressed from a signal. 21. A first frequency division step of dividing a reception signal on the time axis to calculate a reception signal on the frequency axis, and dividing a transmission signal on the time axis to generate a transmission signal on the frequency axis. A second frequency division step of calculating, a transmission noise suppression step of suppressing a noise component of the transmission signal on the frequency axis, a transfer function between a transmission speaker and a reception microphone is estimated, and a pseudo echo on the frequency axis is calculated. An echo canceller step of calculating a signal and subtracting a pseudo echo signal on the frequency axis from a transmission signal on the frequency axis in which the noise component is suppressed to remove an echo component of the transmission signal on the frequency axis. On the frequency axis remaining without being removed in the echo canceller process based on the pseudo echo signal on the frequency axis and the transmission signal on the frequency axis from which the pseudo echo signal on the frequency axis is subtracted. Of the transmitted signal An echo suppression step of suppressing the echo component, and a frequency synthesizing step of synthesizing a transmission signal on the time axis with the echo component suppressed from a transmission signal on the frequency axis with the echo component suppressed. An echo suppression method comprising: 22. A first frequency division step of dividing a reception signal on the time axis to calculate a reception signal on the frequency axis, and dividing a transmission signal on the time axis to generate a transmission signal on the frequency axis. A second frequency division step of calculating, a reception noise suppression step of suppressing a noise component of the reception signal on the frequency axis, a transmission noise suppression step of suppressing a noise component of the transmission signal on the frequency axis, The transfer function between the speaker and the receiving microphone is estimated to calculate the pseudo echo signal on the frequency axis, and the pseudo echo signal on the frequency axis is converted from the transmission signal on the frequency axis in which the noise component is suppressed. An echo canceller process for removing the echo component of the transmission signal on the frequency axis by subtraction, and an echo suppression for suppressing the echo component of the transmission signal on the frequency axis that remains without being removed by the echo canceller process. Process and A frequency synthesizing step of synthesizing a transmission signal on the time axis in which the echo component has been suppressed from a transmission signal on the frequency axis in which the echo component has been suppressed. 23. A first frequency division step of dividing a reception signal on the time axis to calculate a reception signal on the frequency axis, and dividing a transmission signal on the time axis to generate a transmission signal on the frequency axis. A second frequency division step of calculating, a reception noise suppression step of suppressing a noise component of the reception signal on the frequency axis, a transmission noise suppression step of suppressing a noise component of the transmission signal on the frequency axis, A double-talk determining step of determining whether or not a far-end voice component of the received signal on the frequency axis and a near-end voice component of the transmitted signal on the frequency axis are determined to determine whether or not a double-talk state exists; The pseudo echo signal is calculated by estimating the transfer function between the talking speaker and the receiving microphone, and the frequency is calculated by subtracting the pseudo echo signal on the frequency axis from the transmission signal on the frequency axis in which the noise component is suppressed. Echo component of on-axis transmission signal is removed An echo canceller step, an echo suppression step of suppressing the echo component of the transmission signal on the frequency axis remaining without being removed by the echo canceller step, and a transmission on the frequency axis in which the echo component is suppressed. And a frequency synthesizing step of synthesizing a transmission signal on the time axis in which the echo component is suppressed from the speech signal. 24. A first frequency division step of dividing a reception signal on the time axis to calculate a reception signal on the frequency axis, and dividing a transmission signal on the time axis to generate a transmission signal on the frequency axis. A second frequency division step of calculating, a reception noise suppression step of suppressing a noise component of the reception signal on the frequency axis, a transmission noise suppression step of suppressing a noise component of the transmission signal on the frequency axis, A double-talk determining step of determining whether or not a far-end voice component of the received signal on the frequency axis and a near-end voice component of the transmitted signal on the frequency axis are determined to determine whether or not a double-talk state exists; A transfer function multiplication step of multiplying a transfer function between the talk speaker and the receiving microphone by a received signal on the frequency axis to calculate a pseudo echo signal on the frequency axis; and the transfer based on the determination of the double talk determination step. Transfer function updater to update function From an echo suppression step of suppressing the echo component of the transmission signal on the frequency axis remaining without being removed by the echo canceller step, and a transmission signal on the frequency axis in which the echo component is suppressed. And a frequency synthesizing step of synthesizing the transmission signal on the time axis in which the echo component is suppressed. 25. An echo suppressor according to any one of claims 1 to 18, a receiver for receiving the encoded reception signal, and a reception signal on the time axis for the encoded reception signal. A decoder that decodes the received signal on the time axis, a first amplifier that amplifies the received signal on the time axis, a receiving speaker that converts the amplified received signal on the time axis into voice, and outputs the voice of the near-end person. A transmission microphone for converting into a transmission signal on the time axis for input, a second amplifier for amplifying the transmission signal on the time axis, and an encoder for encoding the amplified transmission signal on the time axis And a transmitter for transmitting the encoded transmitter signal on the time axis. 26. The echo suppressor according to any one of claims 1 to 18, a receiving unit that receives a far-end side video signal transmitted from a far-end person and a reception signal on a time axis, and A video display unit for displaying the far-end side video signal on the screen, a first amplifier for amplifying the reception signal on the time axis from the far end person, and a reception signal on the time axis from the far end person for far A receiving speaker that converts and outputs the sound of the end person, a microphone that converts the sound of the near end into a transmission signal on the time axis and inputs, and a transmission signal on the time axis. A second amplifier that amplifies, a camera that captures an image on the near-end person side, converts it into a near-end side image signal, and inputs the signal; a transmission that transmits the transmission signal on the time axis and the near-end side image signal A videophone system characterized by having a section.