EP0884886A2
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Echo canceler employing multiple step gains
Assignee: Oki Electric Industry Co Ltd
•
Filing: 1998-04-30
•
Granted: —
•
Expires: 2018-04-30
Abstract
Abstract According to a first aspect of the invention, in an echo canceler having an adaptive filter, the tap coefficients of the adaptive filter are partitioned into at least two segments, and different step gains are assigned to different segments. The step gains are used in adjusting the tap coefficients, and determine the sensitivity of the adjustment to noise in the residual signal left after echo cancellation. Comparatively small step gains are assigned to noise-sensitive segments. According to a second aspect of the invention, the step gain is switched according to the level of local background noise. <IMAGE>
Inventors
- 👤 Masashi Takada
- 👤 Yoshihiro Ariyama
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-11-05 19:09
CPC Classifications
H; H04; H04M; H04M9/00; H04M9/08; H04M9/082
Claims Analysis
Total Claims: 17
Claim 1
Independent Claim
112 words
An echo cancellation method employing an adaptive
filter (10) to generate an echo replica signal, subtracting
the echo replica signal from a local input signal in which
local background noise is present, thereby canceling echo,
and adjusting tap coefficients of the adaptive filter
according to a residual signal left after echo cancellation,
using an algorithm in which a step gain determines
adjustment sensitivity to the residual signal, comprising
the steps of: partitioning said tap coefficients into at least two
segments, which are affected to different degrees by said
local background noise; and assigning at least two different step gains to
respective segments, for use in adjusting the tap
coefficients in said segments.
Claim 2
Dependent Claim
45 words
The method of claim 1, wherein said step of
partitioning further comprises the steps of: calculating smoothed absolute values of said tap
coefficients; calculating a noise power of said residual signal; and comparing said smoothed absolute values of said tap
coefficients with said noise power.
Claim 3
Dependent Claim
38 words
The method of claim 1, further comprising the steps of: classifying said local background noise into different
levels; and selecting different step gains for assignment to said
segments, corresponding to said different levels of said
local background noise;
Claim 4
Dependent Claim
50 words
The method of claim 3, further comprising the steps of: canceling said local background noise from said
residual signal; and choosing between using said residual signal before
cancellation of said local background noise, and using said
residual signal after cancellation of said local background
noise, for adjusting said tap coefficients.
Claim 5
Dependent Claim
142 words
The method of claim 4, wherein said local background
noise is classified into low, intermediate, and high levels,
and said step of choosing comprises: choosing said residual signal before cancellation of
said local background noise when said local background noise
is classified as low; choosing said residual signal after cancellation of
said local background noise when said local background noise
is classified as high; and choosing said residual signal before cancellation of
said local background noise for adjusting the tap
coefficients in a first one of said segments, and choosing
said residual signal after cancellation of said local
background noise for adjusting the tap coefficients in a
second one of said segments, the second one of said segments
being more sensitive to said local background noise than the
first one of said segments, when said local background noise
is classified as intermediate.
Claim 6
Dependent Claim
123 words
An echo canceler, having an adaptive filter (10) that
generates an echo replica signal, subtracting said echo
replica signal from a local input signal in which local
background noise is present, thereby canceling echo, and
adjusting tap coefficients of the adaptive filter according
to a residual signal left after echo cancellation, using an
algorithm in which a step gain determines adjustment
sensitivity to the residual signal, comprising: a partitioner (16) for partitioning said tap
coefficients into at least two segments, which are affected
to different degrees by said local background noise; and a step gain generator (18) coupled to said partitioner
(16), for assigning at least two different step gains to
respective segments, for use in adjusting the tap
coefficients in said segments.
Claim 7
Dependent Claim
53 words
The echo canceler of claim 6, further comprising: a noise power calculator (15) for calculating a noise
power of said residual signal; an absolute smoother (17) for calculating smoothed
absolute values of said tap coefficients; and a partitioner (16) for comparing said smoothed absolute
values of said tap coefficients with said noise power.
Claim 8
Dependent Claim
55 words
The echo canceler of claim 6, further comprising a
noise level classifier (20), coupled to said step gain
generator (18), for classifying said local background noise
into different levels, and causing said step gain generator
to select different step gains for assignment to said
segments, corresponding to said different levels of said
local background noise.
Claim 9
Dependent Claim
22 words
The echo canceler of claim 8, further comprising a
noise canceler (19) for canceling said local background
noise from said residual signal.
Claim 10
Dependent Claim
43 words
The echo canceler of claim 9, wherein said noise level
classifier (20) also chooses between using said residual
signal before cancellation of said local background noise,
and using said residual signal after cancellation of said
local background noise, for adjusting said tap coefficients.
Claim 11
Dependent Claim
140 words
The echo canceler of claim 10, wherein said noise level
classifier (20) classifies said local background noise into
low, intermediate, and high levels, chooses said residual
signal before cancellation of said local background noise
when said local background noise is classified as low,
chooses said residual signal after cancellation of said
local background noise when said local background noise is
classified as high, and chooses said residual signal before
cancellation of said local background noise for adjusting
the tap coefficients in a first one of said segments and
said residual signal after cancellation of said local
background noise for adjusting the tap coefficients in a
second one of said segments, the second one of said segments
being more sensitive to said local background noise than the
first one of said segments, when said local background noise
is classified as intermediate.
Claim 12
Dependent Claim
101 words
An echo cancellation method employing an adaptive
filter (10) to generate an echo replica signal, subtracting
the echo replica signal from a local input signal in which
local background noise is present, thereby canceling echo,
and adjusting tap coefficients of the adaptive filter
according to a residual signal left after echo cancellation,
using an algorithm in which a step gain determines
adjustment sensitivity to the residual signal, comprising
the steps of: classifying said local background noise into different
levels; and selecting different step gains for use in adjusting
said tap coefficients, corresponding to said different
levels of said local background noise.
Claim 13
Dependent Claim
50 words
The method of claim 12, further comprising the steps
of: canceling said local background noise from said
residual signal; and choosing between using said residual signal before
cancellation of said local background noise, and using said
residual signal after cancellation of said local background
noise, for adjusting said tap coefficients.
Claim 14
Dependent Claim
112 words
The method of claim 13, wherein said local background
noise is classified into low, intermediate, and high levels,
and said step of choosing comprises: choosing said residual signal before cancellation of
said local background noise when said local background noise
is classified as low; choosing said residual signal after cancellation of
said local background noise when said local background noise
is classified as high; and choosing said residual signal before cancellation of
said local background noise for adjusting some of said tap
coefficients, and choosing said residual signal after
cancellation of said local background noise for adjusting
some other of said tap coefficients, when said local
background noise is classified as intermediate.
Claim 15
Dependent Claim
116 words
An echo canceler, having an adaptive filter (10) that
generates an echo replica signal, subtracting said echo
replica signal from a local input signal in which local
background noise is present, thereby canceling echo, and
adjusting tap coefficients of the adaptive filter according
to a residual signal left after echo cancellation, using an
algorithm in which a step gain determines adjustment
sensitivity to the residual signal, comprising: a noise level classifier (20) for classifying said
local background noise into different levels; and a step gain generator (18), coupled to said noise level
classifier (20), for selecting different step gains for use
in adjusting said tap coefficients, corresponding to said
different levels of said local background noise.
Claim 16
Dependent Claim
58 words
The echo canceler of claim 15, further comprising a
noise canceler (19) for canceling said local background
noise from said residual signal, wherein said noise level
classifier (20) chooses between using said residual signal
before cancellation of said local background noise, and
using said residual signal after cancellation of said local
background noise, for adjusting said tap coefficients.
Claim 17
Dependent Claim
110 words
The echo canceler of claim 16, wherein said noise level
classifier (20) classifies said local background noise into
low, intermediate, and high levels, chooses said residual
signal before cancellation of said local background noise
when said local background noise is classified as low,
chooses said residual signal after cancellation of said
local background noise when said local background noise is
classified as high, and chooses said residual signal before
cancellation of said local background noise for adjusting
some of said tap coefficients and said residual signal after
cancellation of said local background noise for adjusting
some other of said tap coefficients, when said local
background noise is classified as intermediate.
References Cited (Backward Citations)
No citation records indexed.
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.