Techniques for updating filter coefficients of an adaptive filter
A technique for updating filter coefficients of an adaptive filter includes filtering a signal with an adaptive filter, whose filter coefficients are grouped into filter blocks. In this case a number of the filter blocks is less than or equal to a number of the filter coefficients. During each update period, the filter coefficients for less than all of the filter blocks are updated based on a network echo path impulse response.
1. A method for updating filter coefficients of an adaptive filter, comprising:
filtering a signal with an adaptive filter whose filter coefficients are grouped into filter blocks, wherein a number of the filter blocks is less than or equal to a number of the filter coefficients; and
updating, for each input sample of the signal, the filter coefficients for at least one of the filter blocks but less than all of the filter blocks based on a network echo path impulse response.
2. The method of claim 1 , further comprising:
segmenting the adaptive filter into filter blocks based on values of the filter coefficients.
3. The method of claim 1 , wherein the updating further comprises:
updating the filter coefficients for the filter blocks that exhibit a higher energy more frequently than the filter coefficients for the filter blocks that exhibit a lower energy.
4. The method of claim 1 , wherein the updating further comprises:
updating the filter coefficients for the filter blocks that exhibit a larger absolute value more frequently than the filter coefficients for the filter blocks that exhibit a lower absolute value.
5. The method of claim 1 , wherein the updating further comprises:
updating the filter coefficients for the filter blocks that exhibit a higher magnitude more frequently than the filter coefficients for the filter blocks that exhibit a lower magnitude.
6. The method of claim 1 , wherein a length of the filter blocks is the same.
7. The method of claim 1 , wherein a length of at least two of the filter blocks is different.
8. The method of claim 1 , wherein the filter blocks includes first, second, and third blocks, and wherein the filter coefficients in the first block have a higher energy, absolute value, or magnitude than the filter coefficients in the second and third blocks and the updating further comprises:
updating, for each input sample of the signal, the filter coefficients for the first block;
updating, for odd input samples of the signal, the filter coefficients for the second block; and
updating, for even input samples of the signal, the filter coefficients for the third block.
9. The method of claim 1 , wherein the adaptive filter implements an affine projection (AP) algorithm, a normalized least mean squares (NLMS) algorithm, a least mean squares (LMS) algorithm, or a recursive least squares (RLS) algorithm during the updating of the filter coefficients.
10. The method of claim 1 , further comprising:
grouping the filter coefficients into the filter blocks based on the network echo path impulse response.
11. The method of claim 1 , wherein the signal is an echo signal and the adaptive filter is configured to reduce the echo signal.
12. A communication device, comprising:
a control module; and
an adaptive filter in communication with the control module, wherein the adaptive filter is configured to filter a signal and is segmented into filter blocks, and wherein a number of the filter blocks is less than or equal to a number of filter coefficients for the adaptive filter, the filter coefficients are grouped into the filter blocks based on a network echo path impulse response, and, for each input sample of the signal, the filter coefficients for at least one of the filter blocks but less than all of the filter blocks are updated.
13. The communication device of claim 12 , wherein the filter coefficients for the filter block that exhibits a highest energy is updated more frequently than the filter coefficients for the filter blocks that exhibit a lower energy.
14. The communication device of claim 12 , wherein the filter coefficients for the filter block that exhibits a larger magnitude is updated more frequently than the filter coefficients for the filter blocks that exhibit a smaller magnitude.
15. The communication device of claim 12 , wherein the filter coefficients for the filter block that exhibits a highest absolute value is updated more frequently than the filter coefficients for the filter blocks that exhibit a lower absolute value.
16. The communication device of claim 12 , wherein a length of the filter blocks is the same.
17. The communication device of claim 12 , wherein a length of at least two of the filter blocks is different.
18. The communication device of claim 12 , wherein the signal is an echo signal and the communication device is an echo canceller.
19. A method for updating filter coefficients of an adaptive filter of an echo canceller, comprising:
filtering an echo signal with an adaptive filter whose filter coefficients are grouped into filter blocks based on a network echo path impulse response, wherein a number of the filter blocks is less than or equal to a number of the filter coefficients; and
updating, for each input sample of the signal, the filter coefficients for at least one of the filter blocks but less than all of the filter blocks based on a metric associated with the filter coefficients.
20. The method of claim 19 , wherein a length of at least two of the filter blocks is different.