Telephonic communications – Echo cancellation or suppression
Reexamination Certificate
1999-09-01
2003-06-10
Isen, Forester W. (Department: 2644)
Telephonic communications
Echo cancellation or suppression
C379S406020, C379S406030, C379S406050, C379S406060, C379S406080, C379S406090, C379S406120, C379S406130
Reexamination Certificate
active
06577731
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to a method or apparatus for canceling echoes in the system using plural received signals and a single or plural transmission signals.
There are multi-channel echo canceling methods or apparatuses that remove echoes caused by received signals propagating through spatial acoustic paths in the system of transmitting plural received signals or a single or plural transmission signals. Two types of multi-channel echo canceling apparatuses including a cascade connection type and a linear coupled type have been proposed in the Technical Report of the Institute of Electronics, Information and Communication Engineers (IEICE) of Japan, Vol. 84, No. 330, pp. 7-14, CS-84-178 (hereinafter referred to as reference 1). According to the reference (1), the cascade connection type is inferior to the linear coupled type in echo suppression capability caused by the constitutional constraint. Here, the case where a linear coupled multi-channel echo canceling apparatus (echo canceller) is applied to a two-channel system, together with received signals and transmission signals, will be described below as a prior art.
FIG. 24
shows a linear coupled multi-channel echo canceller. The first loudspeaker
3
reproduces the first received signal
1
. The second loudspeaker
4
reproduces the second received signal
2
. The first echo
5
which is generated as the signal propagating via spatial acoustic paths from the first loudspeaker
3
to the first microphone
9
and the second echo
6
which is produced as the signal propagating through spatial acoustic paths from the second loudspeaker
4
to the first microphone
9
are added to the first transmission signal
12
or the voice of the talker
11
received with the first microphone
9
, so that a first mixed signal
14
is created. Similarly, the third echo
7
which is generated as the signal propagating the spatial acoustic paths from the first loudspeaker
3
to the second microphone
10
and the fourth echo
8
which is produced as the signal propagating through spatial acoustic paths from the second loudspeaker
4
to the second microphone
10
are added to the second transmission signal
13
or the voice of the talker
11
received with the second microphone
10
, so that a second mixed signal
15
is created.
In order to remove the echoes introduced into the first mixed signal
14
, the first adaptive filter
121
receives the first received signal
1
and then creates an echo replica
125
corresponding to the first echo
5
. The second adaptive filter
122
receives the second received signal
2
and then creates an echo replica
126
corresponding to the second echo
6
. The first subtracter
129
subtracts the echo replicas
125
and
126
corresponding to the first echo
5
and the second echo
6
, respectively, from the first mixed signal
14
. The first and second adaptive filters
121
and
122
are controlled to minimize the output of the first subtracter
129
. The output of the first subtracter
129
becomes the first output signal
16
of the echo canceller
100
.
In order to remove the echoes introduced into the second mixed signal
15
, the third adaptive filter
123
receives the first received signal
1
and then creates an echo replica
127
corresponding to the third echo
7
. The fourth adaptive filter
124
receives the second received signal
2
and then creates an echo replica
128
corresponding to the fourth echo
8
. The second subtracter
130
subtracts the replicas
127
and
128
corresponding to the third echo
7
and the fourth echo
8
from the second mixed signal
15
, respectively. The third and fourth adaptive filters
123
and
124
are controlled to minimize the output of the second subtracter
130
. The output of the second subtracter
130
becomes the second output signal
17
of the echo canceller
100
.
In the multi-channel teleconferencing system being one of important applications for multi-channel echo cancellation, plural microphones receives voices of talkers. Hence, received signals by each microphone can be approximately regarded as signals with an attenuation and a delay each depending on the distance between the talker and the microphone. The mutual correlation between received signals in different channels becomes very high. Let us now assume that the second received signal
2
which was generated as a delayed version of the first received signal
1
, an echo path which can be modeled as a finite impulse response filter, and an echo canceller based on linear combination with adaptive finite impulse response filters.
It is now assumed that x
1
(n) is a first received signal
1
at the time n, x
2
(n) is a second received signal
2
at the time n, and d(n) is an echo introduced into the mixed signal
14
. When the time differences between the first received signals
1
and the second received signals
2
is n
d
(natural number), X
2
(n) is expressed by:
x
2
(
n
)=
x
1
(
n−n
d
) (1)
For simplification, let us now assume that all the spatial acoustic paths ranging from the first and second loudspeakers
3
and
4
to the first and second microphones
9
and
10
are equal in the impulse response length (N). The echo d(n) mixed in the mixed signal
14
is the sum of the echo
5
and the echo
6
and is expressed by:
d
(
n
)
=
∑
i
=
0
N
-
1
h
1
,
i
x
1
(
n
-
i
)
+
∑
i
=
0
N
-
1
h
2
,
i
x
2
(
n
-
i
)
(
2
)
where h
1,j
is an impulse response sample value of a spatial acoustic path ranging from the loudspeaker
3
to the microphone
9
, h
2,i
is an impulse response sample value of a spatial acoustic path ranging from the loudspeaker
4
to the microphone
9
, and i is an integer value between 0 and N−1.
When x
2
(n) is eliminated by substituting the equation (1) into (2), the following equation (3) is obtained:
d
(
n
)
=
∑
i
=
0
n
d
-
1
h
1
,
i
x
1
(
n
-
i
)
+
∑
i
=
n
d
N
-
1
(
h
1
,
i
+
h
2
,
i
-
n
d
)
x
1
(
n
-
1
)
+
∑
i
=
N
-
n
d
N
-
1
h
2
,
i
x
1
(
n
-
n
d
-
i
)
(
3
)
The echo replica d(n) hat created by each of the adaptive filters
121
and
122
is expressed by:
d
^
(
n
)
=
∑
i
=
0
N
-
1
w
1
,
i
(
n
)
x
1
(
n
-
i
)
+
∑
i
=
0
N
-
1
w
2
,
i
(
n
)
x
2
(
n
-
i
)
(
4
)
where w
1,i
(n) is the i-th filter coefficient of the adaptive filter
121
and w
2,i
(n) is the i-th filter coefficient of the adaptive filter
122
.
When x
2
(n) is eliminated by substituting the equation (1) into the equation (4), the following equation (5) is obtained:
d
^
(
n
)
=
∑
i
=
0
n
d
-
1
w
1
,
i
(
n
)
x
1
(
n
-
i
)
+
∑
i
=
n
d
N
-
1
{
w
1
,
i
(
n
)
+
w
2
,
i
-
n
d
(
n
)
}
x
1
(
n
-
i
)
+
∑
i
=
N
-
n
d
N
-
1
w
2
,
i
(
n
)
x
1
(
n
-
n
d
-
i
)
(
5
)
The residual echo e(n) is expressed by:
e
(
n
)
=
d
(
n
)
-
d
^
(
n
)
=
∑
i
=
0
n
d
-
1
{
h
1
,
i
-
w
1
,
i
(
n
)
}
x
1
(
n
-
i
)
+
∑
i
=
n
d
N
-
1
{
h
1
,
i
+
h
2
,
i
-
n
d
-
w
1
,
i
(
n
)
-
w
2
,
i
-
n
d
(
n
)
}
x
1
(
n
-
i
)
+
∑
i
=
N
-
n
d
N
-
1
{
h
2
,
i
-
w
2
,
i
(
n
)
}
x
1
(
n
-
n
d
-
i
)
(
6
)
The condition of completely eliminating echoes from the equation (6) is expressed by:
h
1,i
=w
1,i
(
n
)
i
=0
, . . . ,n
d
−1
h
1,i
+h
2,i−n
d
=w
1,i
(
n
)+
w
2,i−n
d
(
n
)
i=n
d
, . . . ,N−
1
h
2,i
=w
2,i
(
n
)
i=N−n
d
, . . . ,N−
1 (7)
The following expression is uniquely determined by the equation (7).
w
1,0
(
n
), . . . ,
w
1−n
−1
(
n
)and
w
2,N−n
d
(
n
), . . . ,
w
2,N−1
(
n
) [1]
However, it is clear that the following expression has an unlimited number of solutions.
i w
1,n
d
(
n
), . . . ,
w
1,N−1
(
n
) and
w
2,0
, . . . ,w
2,N−n
d
&mi
Dickstein , Shapiro, Morin & Oshinsky, LLP
Isen Forester W.
Singh Ramnandan
LandOfFree
Method and apparatus of cancelling multi-channel echoes does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus of cancelling multi-channel echoes, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus of cancelling multi-channel echoes will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3128802