JPH11284550A - Voice input-output device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a voice input-output device which does not have voice disconnection, even in a simultaneous two-way speech communication state and also performs natural speech communication that does not have so many echo effects by improving a voice input-output device and switching attenuation quantity of a residual signal, based on the volume of an acoustic echo signal and transmission voice. SOLUTION: A conversation environment, in which voice disconnection does not exist even in a simultaneous two-way speech communication state and also which does not have so many echo effects, is provided such that, in a receiving state, when a loud speaker output signal is larger than a prescribed speaker output signal threshold and a residual signal subtracting a dummy echo signal from a transmitting signal is smaller than a prescribed residual echo threshold and for cases otherwise, an attenuator controller 211 sets the attenuation quantity of a transmitting attenuator 212 as 1st attenuation quantity and sets the attenuation quantity of the attenuator 212 to 2nd attenuation quantity that is smaller than the 1st attenuation quantity in cases except the above.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an audio input / output device, and more particularly, to an echo canceller for removing an acoustic echo signal from a loudspeaker output signal which is output from a loudspeaker, and an echo canceller for removing the acoustic echo signal. The present invention relates to an audio input / output device including an echo suppressor for suppressing a residual signal that cannot be removed.

[0002]

2. Description of the Related Art For example, in a voice communication system using a speaker and a microphone such as a speakerphone and a videophone,
In recent years, techniques related to echo cancellers that have been used for satellite communication and the like have been applied relatively inexpensively in recent years. Also, with the spread of mobile phones, the demand for devices equipped with the above-described echo canceller has been increasing. In particular, in a mobile phone or the like that requires strong miniaturization, the distance between the speaker and the microphone is short, and there is a high possibility that an acoustic echo signal that has entered the microphone from the speaker has a bad influence on the call state. Here, FIG. 4 shows an example of a conventional voice input / output device for removing the influence of an acoustic echo signal from an input voice. The voice input / output device includes a microphone 400, a speaker 401, an acoustic echo canceller 402, a non-linear processor 40,
3 is provided. In the above audio input / output device, an echo signal sneaking from the microphone 400 to the speaker 401 is suppressed by an acoustic echo canceller 402 including an adaptive filter and the like. However, when the adaptive filter does not converge sufficiently, for example, some residual signal is generated and transmitted as it is to the far end speaker. In the voice input / output device, a non-linear processor (echo suppressor) 403 for suppressing the residual signal is used to prevent the residual signal not removed by the acoustic echo canceller 402 from being transmitted to the far end speaker or the voice recognition device. The acoustic echo canceller 402 is provided at a subsequent stage. FIG. 5 shows an example of the characteristics of the nonlinear processor. This is called a center clipper method. In the center clipper method, the output is attenuated to 0 only when the residual signal of the acoustic echo canceller 402 is within the range of the threshold value ThL to ThH, and otherwise passes the signal as it is. This operation prevents the residual signal from being transmitted to the far end speaker.

[0003]

When the residual signal is suppressed using a non-linear processor as described above, not only the residual signal but also signals necessary for natural conversation such as background noise are removed. Frequent reception during double talk /
When switching between transmissions and switching to transmission / reception from silence, the nonlinear processor causes a sudden change in the background sound or the beginning of the speech, causing the speaker to feel uncomfortable. Was. In order to solve the problems in the conventional technology, the present invention improves the sound input / output device and switches the amount of attenuation of the residual signal based on the size of the acoustic echo signal and the transmitted sound, thereby achieving simultaneous transmission. It is an object of the present invention to provide a voice input / output device capable of performing a natural call with no sound interruption and less influence of an echo even in a two-way call state.

[0004]

According to a first aspect of the present invention, a pseudo echo signal is generated by referring to a microphone, a speaker, and a speaker output signal output from the speaker. An acoustic echo canceller for subtracting the pseudo echo signal from the acoustic echo signal from which the speaker output signal wraps around the microphone, and an echo suppressor for suppressing a residual signal between the acoustic echo signal and the pseudo echo signal. An audio input / output device comprising: a residual signal level measuring means for measuring the level of the residual signal; and a speaker output signal level measuring means for measuring the level of the speaker output signal. Is a transmission attenuator that attenuates a transmission signal input from the microphone and transmitted to the far-end speaker, and a transmission attenuator received by the near-end speaker. A receiving attenuator for attenuating the received signal, and the transmission attenuation based on the level of the residual signal and the level of the speaker output signal measured by the residual signal level measuring means and the speaker output signal level measuring means, respectively. And an attenuator controller for controlling the attenuation of the receiving attenuator and the receiving attenuator, respectively. According to a second aspect of the present invention, in the audio input / output device according to the first aspect, the attenuator controller controls the transmission signal and the reception signal together with the level of the residual signal and the level of the speaker output signal. Controlling the attenuation of the transmission attenuator and the attenuation of the reception attenuator based on the magnitude relationship of the signals, wherein the transmission signal is in a receiving state smaller than the reception signal and the level of the speaker output signal is When the level of the residual signal is smaller than the predetermined threshold value of the residual signal, the attenuation amount of the transmission attenuator is controlled by the attenuator controller. Is set to a first attenuation, and in other cases in the receiving state, the attenuation of the transmission attenuator is set to a second attenuation smaller than the first attenuation.
The main point is that the above-mentioned residual signal is suppressed by setting the amount of attenuation to. According to a third aspect of the present invention, in the voice input / output device according to the second aspect, when the transmission signal is larger than the reception signal in a transmission state, the attenuator controller controls the reception attenuator. The gist is that the attenuation is set to the second attenuation and the speaker output signal is suppressed. The invention according to claim 4 is:
The audio input / output device according to claim 2 or 3,
In the intermediate state where the transmission signal and the reception signal are substantially the same, the attenuation of the transmission attenuator and the reception attenuator is set to half the second attenuation by the attenuator controller. Is the gist.

In the audio input / output device according to the first aspect, the attenuator control is performed based on the level of the speaker output signal that is the source of the acoustic echo signal and the level of the residual signal that has not been removed by the acoustic echo canceller. Since the attenuation of the transmission attenuator and the reception attenuator is set by the device, it is possible to provide the user with a natural communication environment with less influence of the echo. Further, according to the audio input / output device according to the second aspect, in a receiving state, the level of the speaker output signal is higher than the predetermined speaker output signal threshold value, and the level of the residual signal is If it is smaller than the predetermined residual signal threshold, in other words, if there is a received voice but no transmitted voice and it is not necessary to transmit a signal to the far-end talker, the attenuation of the transmission attenuator is used. The amount is for example -14d
B is set, and the residual signal is sufficiently suppressed. Otherwise, the attenuation of the transmission attenuator is set to, for example, about -6 dB in other cases, so that, for example, even when the double talk or the echo canceller is not converged, the natural communication state is maintained. The residual signal is suppressed while maintaining. According to the third aspect of the present invention, in the transmitting state, the attenuation of the receiving attenuator is set to, for example, about -6 dB by the attenuator controller, so that the speaker emits an excessive volume. Can be prevented. Further, according to the voice input / output device of the fourth aspect, in the intermediate state, the attenuation of the transmission attenuator and the reception attenuator is set to, for example, -3 dB by the attenuator controller, and the transmission voice and the reception voice are received. When there is no voice, the background noise is eliminated and the call state is prevented from becoming unnatural.

[0006]

Embodiments of the present invention will be described below with reference to the accompanying drawings to provide an understanding of the present invention.
Note that this embodiment is a specific example of the present invention, and does not limit the technical scope of the present invention. First, FIG. 1 shows a schematic configuration of a voice input / output device according to the present embodiment. An audio input / output device according to an embodiment of the present invention is used by being connected to, for example, a mobile phone or the like. The audio input / output device outputs a microphone M, a speaker S, and a speaker output signal 50 output from the speaker S. The acoustic echo canceller 10 generates a pseudo echo signal 104 by referring to the acoustic echo signal, and deducts the pseudo echo signal 104 from the acoustic echo signal from which the speaker output signal wraps around the microphone M.
And an echo suppressor 2 for suppressing a residual signal 105 between the acoustic echo signal and the pseudo echo signal 104.
This is the same as the prior art in that it has 0. The difference between the voice input / output device and the conventional technology is that a residual level monitor 108 for measuring the level of the residual signal 105 is used.
(Residual signal level measuring means) and a speaker signal level monitor 10 for measuring the level of the speaker output signal 50
6 (loudspeaker output signal level measuring means), wherein the echo suppressor 20 attenuates a transmission signal input from the microphone M and transmitted to the far end speaker, and a near end attenuator. A receiver attenuator 213 for attenuating a received signal received by the speaker, and the residual level monitor 108
And the transmission attenuator 212 based on the level of the residual signal 105 and the level of the speaker output signal 50 measured by the speaker signal level monitor 106, respectively.
And an attenuator controller 211 for controlling the amount of attenuation of the reception attenuator 213.

[0007] The attenuator controller 211 controls the transmission attenuator 212 and the receiver based on the level of the residual signal 105 and the level of the speaker output signal 50, as well as the magnitude relationship between the transmission signal and the reception signal. This is for controlling the attenuation amounts Gtx and Grx of the attenuator 213, respectively, wherein the transmission signal is in a receiving state smaller than the reception signal,
The level of the speaker output signal 50 measured by the speaker output signal level monitor 106 is larger than a predetermined speaker output signal threshold, and the level of the residual signal 105 measured by the residual level monitor 108 is Is smaller than the residual signal threshold value, the attenuator controller 211 sets the attenuation Gtx of the transmission attenuator 212 to, for example, -14 dB (first attenuation), and in the reception state. In other cases, the attenuation amount Gtx of the transmission attenuator 212 is controlled by the attenuator controller 211.
Is set to, for example, -6 dB (second attenuation), and the above-mentioned residual signal 105 is suppressed.

Next, details of the voice input / output device will be described. In the voice input / output device, a reception signal 40 from a far-end speaker received by the near-end speaker is input to the echo suppressor 20. In the echo suppressor 20, the instantaneous level of the received signal 40 is averaged for a predetermined time by the level detector 202, and the received signal level b2 is averaged.
(204) is calculated. On the other hand, a transmission signal (microphone input signal) 3 input from the microphone M and transmitted to the far-end talker side
0 is input to the echo suppressor 20 via the acoustic echo canceller 10. In the echo suppressor 20, the instantaneous level of the transmission signal 30 is averaged for a predetermined time by a level detector 201, and the transmission signal level b0 is averaged.
(203) is calculated. The transmission signal level b calculated by the two level detectors 201 and 202
0 (203) and the reception signal level b2 (204) are output to the transmission / reception determination unit 205, respectively. The transmission / reception determination unit 205 determines a transmission state (mainly a state in which the reception signal is attenuated) and a reception state (mainly transmission state) in accordance with the magnitude of the transmission signal level b0 (203) and the reception signal level b2 (204). Is determined, and a corresponding determination signal L1 (206) is output to the attenuator controller 211. The discrimination signal L1 is set to 0 in the transmission state, and set to 1 in the reception state. Also, the transmission signal level b0 from the level detector 201
Based on (203), the transmitted voice detector 207 determines whether the transmitted signal is voiced or unvoiced, and outputs a corresponding determination signal Nt (208) to the attenuator controller 211. You. When the transmission signal is in a voiced state, the discrimination signal Nt is 1, and when the transmission signal is unvoiced, the discrimination signal Nt is 0. Similarly, the level detector 202
Based on the reception signal level b2 (204) from the receiver, the reception voice detector 209 determines whether the reception signal 40 is in a voiced state or an unvoiced state, and the corresponding discrimination signal Nr (210) is attenuator control. Output to the unit 211.
When the reception signal is in the voiced state, the discrimination signal Nr is set to 1; when the reception signal is in the unvoiced state, the discrimination signal Nt is set to 0.

The acoustic echo canceller 10 generates a pseudo echo signal 104 by the adaptive filter 103 using the reception signal (speaker output signal) 50 attenuated by the reception attenuator 213 as a reference signal. Then, based on the transmission signal 30 input from the microphone M, the pseudo echo signal 1
By subtracting 04, the acoustic echo signal sneaking into the speaker S from the microphone M is suppressed. The residual signal of the acoustic echo canceller 10, that is, the residual signal 105 between the acoustic echo signal and the pseudo echo signal 104 is suppressed by the echo suppressor 20. The magnitude of the speaker output signal 50 is determined by the speaker signal level monitor 106.
And the speaker output signal detection flag Tm
(101) is generated. If the speaker output signal detection flag Tm is larger than a predetermined speaker output signal threshold value, Tm = 1, and if smaller, Tm = 0. The residual signal 105 of the acoustic echo canceller 10 is measured by a residual level monitor 108, and a residual detection flag Rm (102) is generated. If the residual detection flag Rm is smaller than the threshold value, Rm = 1, and if smaller, Rm = 0. In the audio input / output device according to the present embodiment, the discrimination signal L
1, Nt, Nr, the speaker output signal flag Tm (101), and the residual detection flag discrimination signal R
The attenuation Gtx of the transmission attenuator 212 and the attenuation Grx of the reception attenuator 213 are controlled by the attenuator controller 211 of the echo suppressor 211 based on m (102). The discrimination signals L1, Nt, NR, and the flag T
FIG. 2 shows an example of the attenuation amounts Gtx and Grx of the transmission attenuator 212 and the reception attenuator 213 set by the attenuator controller 211 based on m and Rm, and FIG. 3 shows an operation example thereof.

Next, the operation of the voice input / output device will be described in detail with reference to FIGS. State A is a transmission state, in which a transmission signal 30 is present. At this time, the attenuation of the transmission attenuator 212 is set to 0 dB and the attenuation of the reception attenuator 213 is set to -6 dB by the attenuator controller 211. As a result, the volume of the receiving speaker is slightly reduced, but not at a level that hinders conversation. State B is a transmission state, but the transmission signal 30 is in a silent state. At this time, the transmission attenuator 2 is controlled by the attenuator controller 211.
12, the attenuation of the receiving attenuator 213 is set to -3 dB, and both are balanced. still,
In the transition from the state A to the state B, the change of each attenuator is gradual. State C is a receiving state, but the receiving voice 40 is a silent state. The attenuator 21 at this time
The settings of 2, 213 are the same as in the above state B. The intermediate state according to the present invention corresponds to the above state B or state C. State D is the receiving state, but the speaker output signal 50
Is smaller than a predetermined speaker output signal threshold, and the residual signal 105 is larger than a predetermined residual signal threshold. This state D indicates the head timing of the received
It often occurs at the time of double talk in which both transmitted voice and received voice exist. At this time, the attenuation of the transmission attenuator 212 is set to -6 dB, and the attenuation of the reception attenuator 213 is set to 0 dB by the attenuator controller 211. In this state D, since the speaker output signal 50 is small, the echo signal included in the residual signal 105 is very small. State E is a receiving state, in which the speaker output signal 50 is larger than a predetermined speaker output signal threshold and the residual signal 105 is larger than a predetermined residual signal threshold. This state E occurs, for example, when the operation of the echo canceller 10 starts. At this time, the attenuation of the transmission attenuator 212 is set to -6 dB, and the attenuation of the reception attenuator 213 is set to 0 dB by the attenuator controller 211. State F is a receiving state, and the speaker output signal 5
0 is smaller than a predetermined speaker output signal threshold, and the residual signal 105 is smaller than a predetermined residual signal threshold. This state F occurs when both the transmitted voice and the received voice are low. At this time, attenuator controller 2
11, the attenuation of the transmission attenuator 212 is set to -6 dB, and the attenuation of the reception attenuator 213 is set to 0 dB. In this state F, the residual signal 105 is hardly transmitted to the far end speaker. State G is a receiving state, and the speaker output signal 5
0 is greater than a predetermined speaker output signal threshold and the residual signal 105 is smaller than a predetermined residual signal threshold. In this state G, there is no transmitted voice and only received voice exists, and the adaptive filter 103 used in the echo canceller 10 converges to a predetermined level, and the residual signal 105 of the echo canceller 10 becomes a predetermined level. This is the case when it becomes smaller. At this time, the attenuation of the transmission attenuator 212 by the attenuator controller 211 is −14 dB,
The attenuation of the receiving attenuator 213 is set to 0 dB, and the residual signal 105 from the acoustic echo canceller 10 is sufficiently suppressed. As described above, in the voice input / output device according to the present embodiment, the echo attenuation is set by setting the attenuation of the transmission attenuator and the reception attenuator based on the magnitudes of the speaker output signal and the residual signal. Therefore, it is possible to provide a conversation environment in which the sound is not interrupted even in the simultaneous two-way communication state and the influence of the echo is small.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS In the above embodiment, the voice input / output device according to the present invention was used for the pre-processing of the telephone. However, the present invention is not limited to this. Further, the attenuation of the transmission attenuator and the reception attenuator in each state is not limited to the above example. Such a voice input / output device is also an example of the voice input / output device in the present invention.

[0012]

As described above, in the audio input / output device according to the first aspect, the level of the speaker output signal which is the source of the acoustic echo signal and the level of the residual signal not removed by the acoustic echo canceller are reduced. Based on this, the attenuation of the transmission attenuator and the reception attenuator is set by the attenuator controller, so that the user can be provided with a natural communication environment with less influence of echo. Further, according to the audio input / output device according to the second aspect, in a receiving state, the level of the speaker output signal is higher than the predetermined speaker output signal threshold value, and the level of the residual signal is If it is smaller than the predetermined residual signal threshold, in other words, if there is a received voice but no transmitted voice and it is not necessary to transmit a signal to the far-end talker, the attenuation of the transmission attenuator is used. The amount is set to, for example, about −14 dB, and the residual signal is sufficiently suppressed. Otherwise, the attenuation of the transmission attenuator is set to, for example, about -6 dB in other cases, so that, for example, even when the double talk or the echo canceller is not converged, the natural communication state is maintained. The residual signal is suppressed while maintaining. According to the third aspect of the present invention, in the transmitting state, the attenuation of the receiving attenuator is set to, for example, about -6 dB by the attenuator controller, so that the speaker emits an excessive volume. Can be prevented. Further, according to the voice input / output device of the fourth aspect, in the intermediate state, the attenuation of the transmission attenuator and the reception attenuator is set to, for example, -3 dB by the attenuator controller, and the transmission voice and the reception voice are received. When there is no voice, the background noise is eliminated and the call state is prevented from becoming unnatural.

[Brief description of the drawings]

FIG. 1 is a diagram showing a schematic configuration of a voice input / output device according to an embodiment of the present invention.

FIG. 2 is a diagram showing an example of setting of an attenuator in the audio input / output device.

FIG. 3 is a diagram showing an operation example of an attenuator of the audio input / output device.

FIG. 4 is a diagram showing an example of a conventional voice input / output device.

FIG. 5 is a diagram illustrating a center clipper type non-linear processor.

[Explanation of symbols]

 M ... Microphone S ... Speaker 10 ... Acoustic echo canceller 20 ... Echo suppressor 106 ... Speaker signal level monitor 108 ... Residual level monitor 211 ... Attenuator controller 212 ... Transmission attenuator 213 ... Reception attenuator

Claims (4)

[Claims] 1. A quasi-echo signal is generated by referring to a microphone, a loudspeaker, and a loudspeaker output signal output from the loudspeaker. And a sound input / output device including an acoustic echo canceller for subtracting the residual signal and an echo suppressor for suppressing a residual signal between the acoustic echo signal and the pseudo echo signal. And a speaker output signal level measuring means for measuring a level of the speaker output signal, wherein the echo suppressor converts a transmission signal input from the microphone and transmitted to the far-end talker side. A transmission attenuator for attenuating, a reception attenuator for attenuating a reception signal received at the near end speaker, and the residual signal level measuring means And an attenuator controller for controlling the attenuation of the transmission attenuator and the attenuation of the reception attenuator based on the level of the residual signal and the level of the speaker output signal respectively measured by the speaker output signal level measuring means. An audio input / output device, comprising: 2. The attenuator controller according to claim 1, wherein said attenuator controller attenuates said transmission attenuator and said reception attenuator based on a magnitude relationship between said transmission signal and said reception signal together with a level of said residual signal and a level of a speaker output signal. Control each amount
In the receiving state where the transmission signal is smaller than the reception signal, the level of the speaker output signal is larger than the predetermined threshold value of the speaker output signal, and the level of the residual signal is the predetermined level of the residual signal. If the threshold value is smaller than the threshold value, the attenuation amount of the transmission attenuator is set to the first attenuation amount by the attenuator controller. 2. The audio input / output device according to claim 1, wherein the attenuation is set to a second attenuation smaller than the first attenuation, and the residual signal is suppressed. 3. In a transmission state where the transmission signal is larger than the reception signal, the attenuation of the reception attenuator is set to the second attenuation by the attenuator controller, and the speaker output signal is set. 3. The audio input / output device according to claim 2, wherein the audio input / output device is suppressed. 4. In an intermediate state where the transmission signal and the reception signal are substantially the same, the attenuation of the transmission attenuator and the reception attenuator is set to half of the second attenuation by the attenuator controller. The audio input / output device according to claim 2 or 3, wherein the audio input / output device is set to a value.