SDR-based adaptive noise cancellation (ANC) system
The overall performance of an ANC system may be improved by configuring the ANC system to perform adaption in the frequency domain. The ANC systems may be configured to update an algorithm of an adaptive filter based, at least in part, on the first input signal, the second input signal, and a feedback signal that is based on an output of the adaptive filter. Updating may include changing parameters of the algorithm based on a SDR based, at least in part, on the first input signal. Updating may also include normalizing a step size and processing at least full band information for the input signal in a frequency domain to generate coefficient values for the algorithm. Updating may also include applying a frequency domain magnitude constraint on adaptive filter coefficients.
1. A method, comprising:
receiving a first input signal for reproduction through a transducer;
receiving a second input signal from a reference microphone or an error microphone; and
updating an algorithm for an adaptive filter based, at least in part, on the first input signal and the second input signal,
wherein the step of updating the algorithm for the adaptive filter comprises:
processing the first input signal in a frequency domain to generate coefficient values corresponding to a plurality of frequency bins for the algorithm; and
updating the coefficient values based on a step size, wherein the step size for a frequency bin is based, at least in part, on a value of the frequency bin and a value of at least one other frequency bin.
2. The method of claim 1 , wherein the step of updating the algorithm comprises normalizing the step size based, at least in part, on full band information of the second input signal.
3. The method of claim 1 , wherein the step of updating the algorithm comprises normalizing the step size based, at least in part, on the second input signal.
4. The method of claim 1 , wherein the step of updating the algorithm comprises normalizing the step size based, at least in part, on a maximum magnitude of full band information of the second input signal.
5. The method of claim 1 , wherein the step of updating the algorithm of the adaptive filter comprises calculating the set of coefficient values w m f for the plurality of frequency bins n for a transfer function (w(z)) based on:
W
m
f
(
n
+
1
)
=
(
1
-
μ
m
λ
m
)
W
m
f
(
n
)
+
μ
m
x
m
f
*
(
n
)
e
m
f
(
n
)
x
m
f
(
n
)
x
m
f
(
n
)
max
wherein μ m is the step size for frequency bin m,
λ m is signal leakage in frequency bin m,
x m f (n) is the second input signal from the reference microphone,
x m f *(n) is a conjugate of the second input signal from the reference microphone, and
e m f (n) is an error signal from an error microphone feedback signal.
6. The method of claim 1 , wherein the step of updating the algorithm for the adaptive filter comprises applying magnitude constraints at individual frequency bins.
7. The method of claim 6 , wherein the step of updating the algorithm of the adaptive filter comprises calculating the set of coefficient values w m f for the plurality of frequency bins n for a transfer function (w(z)) based on:
w
m
f
(
n
+
1
)
=
{
w
m
f
(
n
)
+
μ
m
x
m
f
*
(
n
)
e
m
f
(
n
)
x
m
f
(
n
)
x
m
f
(
n
)
max
w
m
f
(
n
)
<
limiter
m
f
(
1
-
λ
m
)
w
m
f
(
n
)
w
m
f
(
n
)
>
limiter
m
f
wherein μ m is the normalized step size,
λ m is signal leakage in frequency bin m,
x m f (n) is the second input signal from the reference microphone,
x m f *(n) is a conjugate of the second input signal from the reference microphone,
limiter m f is a magnitude limit for frequency bin m, and
e m f (n) is an error signal from an error microphone feedback signal.
8. The method of claim 6 , wherein the step of updating the algorithm of the adaptive filter comprises calculating the set of coefficient values w m f for the plurality of frequency bins n for a transfer function (w(z)) based on:
w
m
f
(
n
+
1
)
=
{
w
m
f
(
n
+
1
)
w
m
f
(
n
+
1
)
<
limiter
m
f
w
m
f
(
n
+
1
)
w
m
f
(
n
-
1
)
limiter
m
f
w
m
f
(
n
-
1
)
>
limiter
m
f
,
wherein limiter m f is a magnitude limit for frequency bin m.
9. The method of claim 1 , further comprising:
generating an adaptive noise cancellation (ANC) signal using the adaptive filter; and
generating an output signal for output to the transducer comprising a combination of the first input signal and the adaptive noise cancellation (ANC) signal.
10. An apparatus, comprising:
an audio controller configured to perform steps comprising:
receiving a first input signal for reproduction through a transducer;
receiving a second input signal from a reference microphone or an error signal from an error microphone; and
updating an algorithm for an adaptive filter based, at least in part, on the first input signal and the second input signal,
wherein the step of updating the algorithm for the adaptive filter comprises:
processing the first input signal in a frequency domain to generate coefficient values corresponding to a plurality of frequency bins for the algorithm; and
updating the coefficient values based on a step size, wherein the step size for a frequency bin is based, at least in part, on a value of the frequency bin and a value of at least one other frequency bin.
11. The apparatus of claim 10 , wherein the step of updating the algorithm comprises normalizing the step size based, at least in part, on full band information of the second input signal.
12. The apparatus of claim 10 , wherein the step of updating the algorithm comprises normalizing the step size based, at least in part, on a maximum magnitude of full band information of the second input signal.
13. The apparatus of claim 10 , wherein the step of updating the algorithm of the adaptive filter comprises calculating the set of coefficient values w m f for the plurality of frequency bins n for a transfer function (w(z)) based on:
W
m
f
(
n
+
1
)
=
(
1
-
μ
m
λ
m
)
W
m
f
(
n
)
+
μ
m
x
m
f
*
(
n
)
e
m
f
(
n
)
x
m
f
(
n
)
x
m
f
(
n
)
max
wherein μ m is the step size for frequency bin m,
λ m is signal leakage in frequency bin m,
x m f (n) is the second input signal from the reference microphone,
x m f *(n) is a conjugate of the second input signal from the reference microphone,
limiter m f is a magnitude limit for frequency bin m, and
e m f (n) is an error signal from an error microphone feedback signal.
14. The apparatus of claim 10 , wherein the step of updating the algorithm for the adaptive filter comprises applying magnitude constraints at individual frequency bins.
15. The apparatus of claim 14 , wherein the step of updating the algorithm of the adaptive filter comprises calculating the set of coefficient values w m f for the plurality of frequency bins n for a transfer function (w(z)) based on:
w
m
f
(
n
+
1
)
=
{
w
m
f
(
n
)
+
μ
m
x
m
f
*
(
n
)
e
m
f
(
n
)
x
m
f
(
n
)
x
m
f
(
n
)
max
w
m
f
(
n
)
<
limiter
m
f
(
1
-
λ
m
)
w
m
f
(
n
)
w
m
f
(
n
)
>
limiter
m
f
wherein μ m is the normalized step size,
λ m is signal leakage in frequency bin m,
x m f (n) is the second input signal from the reference microphone,
x m f *(n) is a conjugate of the second input signal from the reference microphone, and
e m f (n) is an error signal from an error microphone feedback signal.
16. The apparatus of claim 14 , wherein the step of updating the algorithm of the adaptive filter comprises calculating the set of coefficient values w m f for the plurality of frequency bins n for a transfer function (w(z)) based on:
w
m
f
(
n
+
1
)
=
{
w
m
f
(
n
+
1
)
w
m
f
(
n
+
1
)
<
limiter
m
f
w
m
f
(
n
+
1
)
w
m
f
(
n
-
1
)
limiter
m
f
w
m
f
(
n
-
1
)
>
limiter
m
f
,
wherein limiter m f is a magnitude limit for frequency bin m.
17. The apparatus of claim 10 , wherein the audio controller is further configured to perform the steps of:
generating an adaptive noise cancellation (ANC) signal using the adaptive filter; and
generating an output signal for output to the transducer comprising a combination of the first input signal and the adaptive noise cancellation (ANC) signal.