US 10,542,153 B2 · Multi-channel residual echo suppression

Abstract

Audio systems and methods for suppressing residual echo are provided. First and second audio program content signals are received, and a residual signal from an echo canceler is received. A first spectral mismatch is determined based at least upon a cross power spectral density of the first program content signal and the residual signal. A second spectral mismatch is determined based at least upon a cross power spectral density of the second program content signal and the residual signal. The residual signal is filtered to reduce residual echo, based at least upon the first spectral mismatch, the second spectral mismatch, a spectral density of the first program content signal, a spectral density of the second program content signal, and a spectral density of the residual signal.

Claims16

Dependent claims

A method of suppressing residual echo, comprising:

receiving a residual signal from an echo cancelation subsystem;
receiving a first program content signal;
determining a first spectral mismatch based at least upon a cross power spectral density of the first program content signal and the residual signal;
at least one of retrieving, from a memory, an initial first spectral mismatch based upon a selected sound stage configuration, and storing, in the memory, the determined first spectral mismatch in association with the selected sound stage configuration;
receiving a second program content signal;
determining a second spectral mismatch based at least upon a cross power spectral density of the second program content signal and the residual signal;
and
controlling a filter to filter the residual signal based upon the first spectral mismatch, the second spectral mismatch, a spectral density of the first program content signal, a spectral density of the second program content signal, and a spectral density of the residual signal.

5 more elements

A residual echo suppression system, comprising:

a plurality of content signal inputs, each of the plurality of content signal inputs configured to receive an audio signal;
a residual input configured to receive a residual signal from an echo cancelation system;
a memory;
a processor configured to determine a plurality of spectral mismatches, each of the plurality of spectral mismatches based at least upon a cross power spectral density involving at least one of the audio signals, the processor further configured to at least one of retrieve, from the memory, an initial first spectral mismatch based upon a selected sound stage configuration and store, in the memory, at least one spectral mismatch from the plurality of determined spectral mismatches in association with the selected sound stage configuration;
and
a filter configured to filter the residual signal based at least upon the plurality of spectral mismatches to provide an output signal.

3 more elements

An audio system comprising:

a plurality of inputs, each of the plurality of inputs to receive an audio signal;
a signal processing subsystem coupled to the plurality of inputs and configured to provide a processed signal based upon at least a portion of the plurality of audio signals to a loudspeaker for rendering an acoustic signal;
a microphone input to receive a microphone signal having an echo component of the acoustic signal;
an echo cancelation subsystem configured to provide an estimated echo signal from the plurality of audio signals, to receive the microphone signal, and to subtract the estimated echo signal from the microphone signal to provide a residual signal;
and
a residual echo suppression subsystem configured to determine a plurality of spectral mismatches, each of the plurality of spectral mismatches based at least upon a cross power spectral density of the residual signal and one of the plurality of audio signals, and to filter the residual signal based at least upon the plurality of spectral mismatches to provide an output signal, wherein the plurality of spectral mismatches is calculated in accord with the equation
H
p
⁢
⁢
f
⁡
(
k
)
=
max
⁢
{
1
-
β
⁢
Δ
⁢
⁢
H
H
⁡
(
k
)
⁢
S
p
⁢
⁢
p
⁡
(
k
)
⁢
Δ
⁢
⁢
H
⁡
(
k
)
S
r
⁢
⁢
r
⁡
(
k
)
+
p
,
H
min
}
,
where ΔHH(k) is a complex conjugate transpose of a spectral mismatch of a frequency bin k, Spp(k) is a cross-power spectral density of the frequency bin k, ΔH(k) is the spectral mismatch of the frequency bin k, Srr is a power spectral density of the residual signal, β is a scaling factor, and p is a regularization factor.

65 more elements

Dependencies are inferred from the claim text.

Multi-channel residual echo suppression

Life of this patent

Multi-channel residual echo suppression

Life of this patent

Abstract

Claims16