Electrical audio signal processing systems and devices – Dereverberators
Reexamination Certificate
1997-04-23
2001-04-10
Chang, Vivian (Department: 2747)
Electrical audio signal processing systems and devices
Dereverberators
C379S406010
Reexamination Certificate
active
06215880
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a two-wire audio conference apparatus, and more particularly to a two-wire audio conference apparatus having a howling and echo prevention function of which are generated in a case where a speaker and a microphone are positioned near each other.
2. Description of Related Art
FIG. 1
is a block diagram of a conventional two-wire audio conference apparatus. In
FIG. 1
, an echo estimator
004
estimates an echo signal when a part of an output signal to a speaker
001
, i.e., a speaker output signal, goes around to a microphone
002
side. This estimated echo signal is subtracted from an input signal supplied from the microphone
002
, i.e., a microphone input signal, by a subtracter
003
such that echo is canceled.
On the other hand, an echo estimator
011
estimates an echo signal when a part of a transmission output signal to a hybrid circuit
009
goes around to a reception input signal from the hybrid circuit
009
. This estimated echo signal is subtracted from the reception input signal by a subtracter
010
such that echo is canceled.
A training controller
020
starts training in response to a trigger from an external trigger unit
019
. Training is performed to each of the echo estimators
004
and
011
. White noise generated from noise generators
015
and
014
are selected by selector
016
and
013
which are controlled by the training controller
020
and are outputted to the speaker
001
and the hybrid circuit
009
along the signal transmission directions, respectively. The echo estimators
011
and
004
refer to these signals to estimate in such a manner that generated echoes are canceled, respectively.
After sufficient estimation is completed, the echo estimator
011
is controlled to stop the estimating operation. The estimation operation of the echo estimator
004
is controlled by a bi-directional talk detector
017
to be described later. Also, the selectors
016
and
013
are reset to the original communication enable states.
A suppressing unit
018
compares the level of the reception input signal from the hybrid circuit
009
with the level of the microphone input signal. The suppressing unit
018
applies a predetermined fixed suppression between an input and an output on the side of lower level, but does not apply any suppression to the side of higher level. For example, it is assumed that level comparison between signals to input terminals a and c of the suppressing circuit
018
is performed and the level of the signal to the input terminal a is higher than that of the signal to the input terminal c. In this case, a suppression is applied to the signal to the input terminal c and the suppression applied signal is outputted from an output terminal b of the suppressing circuit
018
. On the contrary, any suppression is not applied to the signal to the input terminal a and the signal to the input terminal a is outputted from an output terminal d as it is. This suppression allows howling to be prevented and generation of remaining echo after the echo cancellation to be prevented.
In the estimating operation of the echo estimator
004
which processes echo on the acoustic side, the level of the output signal to the speaker
001
and the level of the input signal from the microphone
002
are compared by the bidirectional talk detector
017
. When the microphone input signal level is lower than the speaker output signal level, control is performed to activate the estimating operation of the echo estimator
004
.
An amplifier
007
amplifies and adjusts the transmission output signal to a line in level. An amplifier
012
amplifies and adjusts a signal to the speaker
001
in level.
A first problem of this training method is that it is necessary to generate a large level of noise in training. When there is a far end apparatus having the same structure, the processes of noise generation from the speaker and noise sending to the line must be executed twice. Also, in this method, there is a possibility that the training is erroneously performed in the near end apparatus when the far end apparatus performs the training at the same time. For instance, when a signal is inputted from the far end apparatus to the near end apparatus while the near end apparatus is being trained, the estimation result of the echo estimator
011
would include an error.
The second problem of the conventional technique is that the training process needs to be performed at the beginning of communication for estimation by an echo estimator. Also, it is desirable that the sound source used for the training is white noise and has a high level. For this reason, the high level of noise in the training is unpleasant to the people nearby. Also, in a case of erroneous training, howling sometimes is generated.
The third problem of the conventional technique is that there is sometimes a remaining echo after the echo cancellation because of insufficient suppression so that there is a case where an echo returns. As a result, an echo is amplified from the speaker of the far end apparatus, which can degrade communication quality degrade. This is a problem especially when the line delay is large.
SUMMARY OF THE INVENTION
An object of the present invention is to provide a method of accurately performing echo estimation without any training and a two-wire audio conference apparatus using the method.
Another object of the present invention is to provide a method of adaptively adjusting an attenuation quantity of a remaining echo such that high communication quality can be achieved and a two-wire audio conference apparatus for the same.
In order to achieve these objects an aspect of the present invention, a two-wire audio conference apparatus, includes a microphone, a speaker, a hybrid circuit for transmitting a transmission output signal inputted to its input terminal and for receiving a signal to output a reception input signal from its output terminal, a first subtracter connected to the microphone, a first echo estimating section responsive to a first activation control signal, for estimating a speaker output signal as a signal to be outputted from the speaker to supply a first echo cancellation quantity to the first subtracter, wherein the first subtracter subtracts the first echo cancellation quantity from a microphone input signal as a signal to be outputted from the microphone to output a first subtracted signal, a first circuit section provided between the first subtracter and the input terminal of the hybrid circuit, for generating a signal, which is attenuated in response to a first attenuation instruction and amplified with a predetermined amplification, from the first subtracted signal to output the generated signal as the transmission output signal to the hybrid circuit, a second subtracter connected to the output terminal of the hybrid circuit, a second echo estimating section responsive to a second activation control signal, for estimating the transmission output signal to supply a second echo cancellation quantity to the second subtracter, wherein the second subtracter subtracts the second echo cancellation quantity from the reception input signal to output a second subtracted signal, a second circuit section provided between the second subtracter and the speaker, for generating a signal, which is attenuated in response to a second attenuation instruction and amplified with a predetermined amplification, from the second subtracted signal to output the generated signal as the speaker output signal to the speaker, and a control section for monitoring the microphone input signal, the first subtracted signal, the transmission output signal, the reception input signal, the second subtracted signal and the speaker output signal, to output the first activation control signal to the first estimating section, the second activation control signal to the second estimating section, the first attenuation instruction to the first circuit section, and the second attenuation instruction to the second circuit secti
Chang Vivian
NEC Corporation
Sughrue Mion Zinn Macpeak & Seas, PLLC
LandOfFree
Two-wire conference apparatus with echo canceler without... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Two-wire conference apparatus with echo canceler without..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Two-wire conference apparatus with echo canceler without... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2504174