JPH088789A - Voice communication system - Google Patents

Abstract

PURPOSE:To improve effects of echo suppression and noise suppression by using the output of a noise cancel means for a signal to be used for the learning of an echo in an echo cancel means. CONSTITUTION:An echo canceler 11 performs the decision of a receiving speech/ transmitting speech in a receiving speech/transmitting speech decision part 12 and performs the learning of an adaptive filter 13 only in the case of the receiving speech and a non-transmitting speech. As a signal to be used in the learning of the adaptive filter 13, the outputs of a noise canceler 14 as an input signal and an error signal are used. By this constitution, a noise elimination characteristic in which the state of a reception signal is taken into consideration is imparted to the noise canceler 14, and the effects of echo suppression and noise suppression can be improved not so as to get the influence of the noise canceler 14 in the echo canceler 11 and enter the influence of the echo canceler 11 in the noise canceler 14.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to improvement of an echo suppression system in a voice communication system.

[0002]

2. Description of the Related Art In a low bit rate speech coder such as CELP, the coding quality is deteriorated by noise in an environment where background noise is large, such as inside a car. To prevent this, a technique has been proposed that suppresses noise without using a special reference voice. This is called a noise canceller. As a conventional example of this, for example, "Study of background noise reduction method in mobile communication" (Miki et al., IEICE)
There is the autumn 1993 national convention B-287).

On the other hand, an echo canceller which enables a hands-free call on a car phone is also an effective technique for improving call quality. In the hands-free call mode, a speaker for reception and a microphone for transmission are installed on a dashboard or the like, separately from the handset, and the speaker and the microphone are used as a transmitter / receiver for communication. When this hands-free call mode is used, the talker can talk without holding a handset, and the call can be made without one-handed driving even during a call, which is effective in terms of safety during driving. However, during such a hands-free call, the voice received from the speaker may sneak into the microphone and be sent to the other party to generate an acoustic echo. In order to prevent the deterioration of the call quality due to this echo, an echo canceller that cancels the wraparound of the received voice is required.

As a conventional example of the echo canceller, an adaptive filter is used to learn the relationship between the received signal and the echo mixed in the transmission input signal, and the learned contents are used from the received signal which is the input of the echo canceller to generate a pseudo echo. Is generated and subtracted from the transmission input signal.

As described above, when making a hands-free call from the inside of an automobile, a noise canceller is required as a countermeasure against background noise, and an echo canceller is required as a countermeasure against echo.

3 to 5 show a conventional communication system including a noise canceller and an echo canceller.

FIG. 3 shows a configuration including only the noise canceller shown in the above document. Noise canceller 11
The signal input to 0 is a transmission input signal, and the state determination unit 112 refers to this input to determine whether this input is noisy or noiseless, and at the same time detects whether it is voiced or unvoiced. Next, in the input parameter calculation unit 111, parameters such as the variance and correlation coefficient of noise and speech are estimated according to the determination result, and the coefficient of the noise cancellation filter 113 is determined based on these parameters, Noise is canceled by operating the filter 113. If it is determined that there is no noise, filtering is not performed.

FIG. 4 shows a communication system including a noise canceller and an echo canceller, which cancels noise first and then cancels echo. The noise canceller 110 has the same function as the noise canceller 110 of FIG. The input of the echo canceller 120 is the reception signal (hereinafter referred to as input A) and the output of the noise canceller 110 of the transmission input signal (hereinafter referred to as input B). Refer to the inputs A and B in the reception / sending determination unit 121. Listen /
The transmission filter is determined, and the adaptive filter 122 is learned using the input A and the input B only in the case of receiving and not transmitting. The echo is canceled by applying a reception signal to the adaptive filter 122 using the coefficient of the adaptive filter 122 obtained by this learning to generate a pseudo echo and subtracting it from the input B.

In FIG. 5, after canceling the echo first,
It is a communication system that cancels noise. The noise canceller 110 and the echo canceller 120 are the same as those in FIG. 4 except that the input of the noise canceller 110 is the output of the echo canceller 120 of the transmission input signal and the input of the echo canceller 120 is the reception signal and the transmission input signal.

As described above, in the conventional example, the input to the noise canceller is only the transmission input signal or the echo canceller output. Further, the noise canceller and the echo canceller only make their own judgments in their respective judgment units and do not share the results.

[0011]

As described above, in the conventional method, when canceling only noise, the noise canceller has only the transmission input signal or a signal corresponding thereto, so that the noise canceling characteristic of the noise canceller is reduced. Did not consider the state of the received signal at all. Therefore, when the echo is mixed in the transmission input signal, the echo may be erroneously determined as noise or voice, and an error may occur in the state determination and the parameter determination.

When canceling the echo after canceling the noise first, since the filter for canceling the noise changes with time, when learning the echo,
There is a problem that accurate learning cannot be performed due to the influence of the noise canceller. Also, the noise canceller cannot use the echo canceling effect of the echo canceller.

When canceling the noise after canceling the echo first, many errors remain in the output of the echo canceller during learning by the echo canceller, which is erroneously determined to be noise or voice, and the state determination and parameter estimation are performed. Error occurs.

Therefore, according to the present invention, the noise canceller is provided with noise removal characteristics in consideration of the state of the received signal so that the echo canceller is not affected by the noise canceller and the noise canceller is not affected by the echo canceller.
The purpose is to improve the effects of echo suppression and noise suppression.

[0015]

In order to achieve the above object, the present invention provides a reproducing means for reproducing a received signal, a transmitting means for generating a transmitting signal from a transmitting input signal, and a noise mixed in the transmitting signal. The noise canceling means for suppressing and the relationship between the received signal and the echo of the received signal mixed in the transmitted signal are learned to generate a pseudo echo estimated from the received signal, and the pseudo signal is generated from the transmitted signal. In a voice communication system having echo canceling means for suppressing echo by subtracting echo, the output of the noise canceling means is used in place of the transmission signal as a signal used for learning of echo by the echo canceling means. And

In addition to the output of the noise canceling means and the received signal, the output of the echo canceling means and the received signal may be used as the signal used for learning the echo by the echo canceling means.
It is characterized in that it has a learning signal switching means for switching both.

The learning signal switching means is characterized in that the learning signal is switched by the output signal of the echo canceling means or by the output signal of the echo canceling means and the reception signal.

[0018]

According to the present invention, the noise canceling characteristics in consideration of the state of the received signal are given to the noise canceller as described above, and whether or not the learning is performed by the echo canceling means is reflected in the noise canceling means. it can.

[0019]

BEST MODE FOR CARRYING OUT THE INVENTION The parts of the voice communication system according to the present invention relating to the invention will be described in detail below with reference to the accompanying drawings.

FIG. 1 is a block diagram showing a first embodiment of the present invention.

This example is a communication system in which echo is first canceled and then noise is canceled.

Inputs to the echo canceller 11 are a reception signal (hereinafter referred to as input A) and a transmission input signal (hereinafter referred to as input B).
Is called).

The echo canceller 11 makes a receiving / sending determination in the receiving / sending determining unit 12 with reference to the input A and the input B, and learns the adaptive filter 13 only when receiving and not transmitting. Then, using the adaptive filter coefficients learned in this way, the adaptive filter 1 is applied to the received signal.
Echo is canceled by applying 3 to generate a pseudo echo and subtracting it from the input B.

The noise canceller 14 cancels noise from the output of the echo canceller 11.

In this example of the voice communication system, when the input signal A and the output of the noise canceller 14 as an error signal are used as the signal used for learning of the adaptive filter 13, the noise by the noise canceller 14 is echoed to the level after cancellation. Can be fully canceled.

FIG. 2 is a block diagram showing a second embodiment of the present invention.

This example is a communication system in which echo is canceled first and then noise is canceled.

The echo canceller 21 makes a reception / transmission determination in the reception / transmission determination unit 22 with reference to the input A and the input B, and learns the adaptive filter 23 only in the case of reception and non-transmission. Then, using the adaptive filter coefficients learned in this way, the adaptive filter 2 is applied to the received signal.
Echo is canceled by applying 3 to generate a pseudo echo and subtracting it from the input B.

The noise canceller 24 cancels noise in the output of the echo canceller 21.

In this voice communication system, input A
At the same time, the learning signal switching unit 25 switches the error signal used for learning of the adaptive filter 23. That is, the subtraction output of the echo canceller 21 is used as an error signal until the echo canceller 21 finishes learning to the noise level. After that, by switching the error signal to the output of the noise canceller 24, it is possible to sufficiently cancel the echo up to the level after the noise cancellation by the noise canceller 24.

Whether or not the echo canceller 21 has finished learning up to the noise level is determined by the echo canceller 21.
Or determine the decrease of the average level of the subtraction output of
Alternatively, the judgment is made by observing the change in the cancellation amount by comparing the ratio of the average level of the input A and the average level of the subtraction output of the echo canceller 21.

[0032]

As described above, according to the present invention, the reproducing means for reproducing the received signal, the transmitting means for generating the transmitting signal from the transmitting input signal, and the noise canceling means for suppressing the noise mixed in the transmitting signal. , Learn the relationship between the received signal and the echo of the received signal mixed in the transmitted signal, generate a pseudo echo estimated from the received signal, and suppress the echo by subtracting this pseudo echo from the transmitted signal In a voice communication system having an echo canceling means, a signal used for learning an echo by the echo canceling means is used as an output of the noise canceling means and a received signal.

In addition to using the noise canceling means output and the received signal as the signal used for the echo learning by the echo canceling means, the echo canceling means output and the received signal may be used, and the learning signal switching means is provided. Switch between the two as the learning progresses.

The progress of learning, which is the reference for switching the learning signal of the learning signal switching means, is judged by either the output signal of the echo cancellation means or the ratio of the output signal of the echo cancellation means and the reception signal.

As a result, the noise canceller which gives consideration to the state of the received signal is given to the noise canceller, the influence of the noise canceller does not enter the echo canceller, and the influence of the echo canceller does not enter the noise canceller, so that the echo suppression and the noise suppression are suppressed. You can improve the effect.

[Brief description of drawings]

FIG. 1 is a block diagram showing a first embodiment of the present invention.

FIG. 2 is a block diagram showing a second embodiment of the present invention.

FIG. 3 is a block diagram showing a conventional example of a voice communication system using a noise canceller.

FIG. 4 is a block diagram showing a conventional example of a voice communication system using a noise canceller and an echo canceller.

FIG. 5 is a block diagram showing another conventional example of a voice communication system using a noise canceller and an echo canceller.

[Explanation of symbols]

 1, 101 Speakers 2, 102 Microphones 11, 21, 111 Echo cancellers 12, 22, 112 Receiving / transmitting judging sections 13, 23, 113 Adaptive filters 14, 24, 114 Noise cancellers 16, 26, 116 Subtractors 117 Parameter calculating section 118 state determination unit 119 noise cancellation filter

Claims (4)

[Claims] 1. A reproducing means for reproducing a reception signal, a transmission means for generating a transmission signal from a transmission input signal, a noise canceling means for suppressing noise mixed in the transmission signal, and a reception signal and a transmission signal. An echo canceling unit that suppresses an echo by learning a relationship between the mixed reception signal and an echo, generating a pseudo echo estimated from the reception signal, and subtracting the pseudo echo from the transmission signal. In the voice communication system having the above, the voice communication system is characterized in that the output of the noise canceling means is used in place of the transmission signal for the signal used for learning the echo by the echo canceling means. 2. The echo canceling means output and the received signal may be used in addition to the noise canceling means output and the received signal as the signals used for the echo learning by the echo canceling means. The voice communication system according to claim 1, further comprising learning signal switching means for switching. 3. The voice communication system according to claim 2, wherein the learning signal switching unit switches the learning signal by using an output signal of the echo canceling unit. 4. The voice communication system according to claim 2, wherein the learning signal switching means switches the learning signal with an output signal of the echo canceling means and a reception signal.