Digital transceiver with multi-rate processing

• US 6,778,599 B1
• Filed: 03/09/2000
• Issued: 08/17/2004
• Est. Priority Date: 03/09/2000
• Status: Expired due to Term

• Current Assignee: STMicroelectronics NV
• Original Assignee: Tioga Technologies Incorporated
• Inventors: Doron, Ayelet Shahar
• Primary Examiner(s): Ghebretinsae, Temesghen
• Application Number:
  US09/521,337
• Time in Patent Office: 1,622 Days
• Field of Search: 375/219, 375/229, 375/230, 375/231, 375/232, 375/233, 379/406.8, 370/286, 370/290
• US Class Current: 375/232
• CPC Class Codes: H03H 2021/0096 with input-sampling frequency and output-delivery frequency which differ, e.g. extrapolation; anti-aliasing
• CPC Class Codes: H03H 21/0012 Digital adaptive filters
• CPC Class Codes: H04L 2025/03509 fractionally spaced H04L2025/03515 takes precedence

Abstract

A digital data receiver includes a front end, coupled to receive an input signal transmitted over a communication channel, the signal representing a stream of transmitted symbols with a given input symbol rate, and to generate, responsive to the input signal, a stream of received samples at a processing rate that is a non-integer rational multiple of the input symbol rate. A feed-forward equalizer receives input samples at the processing rate, responsive to the received samples, and generates forward-equalized samples at the input symbol rate. Decision circuitry processes the forward-equalized samples so as to generate a stream of estimated symbols at the input symbol rate.

Claims

1. A digital data receiver, comprising:

a front end, coupled to receive an input signal transmitted over a communication channel, the signal representing a stream of transmitted symbols with a given input symbol rate, and to generate, responsive to the input signal, a stream of received samples at a processing rate that is a non-integer rational multiple of the input symbol rate;

a feed-forward equalizer, coupled to receive input samples at the processing rate, responsive to the received samples, and to generate forward-equalized samples at the input symbol rate, the feed-forward equalizer comprising;

a first plurality of taps, arranged in series to receive a succession of the input samples at the processing rate;

a second plurality of registers operative to store filter coefficients in a number of groups, such that each group contains one filter coefficient for each of the first plurality of taps;

a third plurality of multipliers, coupled to multiply the samples in the taps by the corresponding filter coefficients so as to generate respective tap outputs for each one of the groups;

one or more summers, coupled to sum the tap outputs for each one of the groups, thus generating respective group equalized partial sums; and

at least one switch, operative to select among the groups at the input symbol rate, so that the equalized partial sums are output in alternation as the forward-equalized samples; and

decision circuitry, coupled to process the forward-equalized samples so as to generate a stream of estimated symbols at the input symbol rate.

2. A receiver according to claim 1, wherein the input signal comprises a Digital Subscriber Line (DSL) signal.

3. A receiver according to claim 1, wherein the feed-forward equalizer comprises an adaptive filter, and wherein the filter coefficients are adjusted automatically responsive to an error signal based on the stream of estimated symbols.

4. A digital data receiver, comprising:

a front end, coupled to receive an input signal transmitted over a communication channel, the signal representing a stream of transmitted symbols with a given input symbol rate, and to generate, responsive to the input signal, a stream of received samples at a processing rate that is a non-integer rational multiple of the input symbol rate;

a feed-forward equalizer, coupled to receive input samples at the processing rate, responsive to the received samples, and to generate forward-equalized samples at the input symbol rate, the feed-forward equalizer comprising;

a first plurality of taps, arranged in series to receive a succession of the input samples at the processing rate;

a second plurality of registers operative to store filter coefficients in a number of groups, such that each group contains one filter coefficient for each of the first plurality of taps;

a third plurality of multipliers, coupled to multiply the samples in the taps by the corresponding filter coefficients so as to generate respective tap outputs for each one of the groups;

one or more summers, coupled to sum the tap outputs for each one of the groups, thus generating respective group equalized partial sums; and

at least one switch, operative to select among the groups at the input symbol rate, so that the equalized partial sums are output in alternation as the forward-equalized samples; and

decision circuitry, coupled to process the forward-equalized samples so as to generate a stream of estimated symbols at the input symbol rate,

wherein the processing rate is substantially equal to 1.5 times the input symbol rate.

5. A digital data receiver, comprising:

a front end, coupled to receive an input signal transmitted over a communication channel, the signal representing a stream of transmitted symbols with a given input symbol rate, and to generate, responsive to the input signal, a stream of received samples at a processing rate that is a non-integer rational multiple of the input symbol rate;

a feed-forward equalizer, coupled to receive input samples at the processing rate, responsive to the received samples, and to generate forward-equalized samples at the input symbol rate, the feed-forward equalizer comprising;

a first plurality of taps, arranged in series to receive a succession of the input samples at the processing rate;

a second plurality of registers operative to store filter coefficients in a number of groups, such that each group contains one filter coefficient for each of the first plurality of taps;

a third plurality of multipliers, coupled to multiply the samples in the taps by the corresponding filter coefficients so as to generate respective tap outputs for each one of the groups;

one or more summers, coupled to sum the tap outputs for each one of the groups, thus generating respective group equalized partial sums; and

at least one switch, operative to select among the groups at the input symbol rate, so that the equalized partial sums are output in alternation as the forward-equalized samples;

decision circuitry, coupled to process the forward-equalized samples so as to generate a stream of estimated symbols at the input symbol rate; and

a training switch, coupled to substitute a sequence of known training symbols for the estimated symbols during an initial adjustment of the filter coefficients of the equalizer.

6. A digital data receiver, wherein the receiver is coupled to a transmitter, which is adapted to transmit an output signal over the channel based on a stream of output symbols, and wherein the receiver comprises:

a front end, coupled to receive an input signal transmitted over a communication channel, the signal representing a stream of transmitted symbols with a given input symbol rate, and to generate, responsive to the input signal, a stream of received samples at a processing rate that is a non-integer rational multiple of the input symbol rate;

a feed-forward equalizer, coupled to receive input samples at the processing rate, responsive to the received samples, and to generate forward-equalized samples at the input symbol rate, the feed-forward equalizer comprising;

a first plurality of taps, arranged in series to receive a succession of the input samples at the processing rate;

a second plurality of registers operative to store filter coefficients in a number of groups, such that each group contains one filter coefficient for each of the first plurality of taps;

a third plurality of multipliers, coupled to multiply the samples in the taps by the corresponding filter coefficients so as to generate respective tap outputs for each one of the groups;

one or more summers, coupled to sum the tap outputs for each one of the groups, thus generating respective group equalized partial sums; and

at least one switch, operative to select among the groups at the input symbol rate, so that the equalized partial sums are output in alternation as the forward-equalized samples;

decision circuitry, coupled to process the forward-equalized samples so as to generate a stream of estimated symbols at the input symbol rate;

an echo canceller, coupled to generate, responsive to the output symbols, echo cancellation values at the processing rate; and

an adder, coupled to sum the received samples with the echo cancellation values, thus generating the input samples to the feed-forward equalizer at the processing rate.

7. A digital data transceiver, comprising:

a transmitter, which is adapted to transmit an output signal over a communication channel based on a stream of output symbols at an output symbol rate;

an echo canceller, coupled to generate, responsive to the output symbols, echo cancellation values at a predetermined processing rate; and

a receiver, which comprises;

a front end, coupled to receive an input signal over the communication channel, the signal representing a stream of transmitted symbols with a given input symbol rate, and to generate, responsive thereto, a stream of input samples at the processing rate;

an adder, coupled to sum the input samples with the echo cancellation values, thus generating echo-canceled samples at the processing rate;

an equalizer, coupled to filter the echo-canceled samples so as to generate equalized samples at the input symbol rate; and

decision circuitry, coupled to process the equalized samples so as to generate a stream of estimated symbols at the input symbol rate,

wherein the processing rate is a non-integer rational multiple of at least one of the input symbol rate and the output symbol rate.

8. A transceiver according to claim 7, wherein the input and output signals comprise Digital Subscriber Line (DSL) signals.

9. A transceiver according to claim 7, wherein the equalizer comprises:

a first plurality of taps, arranged in series to receive a succession of the echo-canceled samples at the processing rate;

a second plurality of registers operative to store a first number of groups of filter coefficients, such that each group contains one filter coefficient for each of the first plurality of taps;

a third plurality of multipliers, coupled to multiply the samples in the taps by the corresponding filter coefficients so as to generate respective tap outputs for each one of the groups;

one or more summers, coupled to sum the tap outputs for each one of the groups, thus generating respective group equalized partial sums; and

at least one switch, operative to select among the groups at the input symbol rate so that the partial sums are output in alternation as the equalized samples at the input symbol rate.

10. A transceiver according to claim 7, wherein the output symbol rate is substantially equal to the input symbol rate.

11. A transceiver according to claim 7, wherein the equalizer comprises an adaptive filter, having filter coefficients that are adjusted automatically responsive to an error signal based on the estimated symbols.

12. A digital data transceiver, comprising:

a transmitter, which is adapted to transmit an output signal over a communication channel based on a stream of output symbols at an output symbol rate;

an echo canceller, coupled to generate, responsive to the output symbols, echo cancellation values at a predetermined processing rate; and

a receiver, which comprises;

a front end, coupled to receive an input signal over the communication channel, the signal representing a stream of transmitted symbols with a given input symbol rate, and to generate, responsive thereto, a stream of input samples at the processing rate;

an adder, coupled to sum the input samples with the echo cancellation values, thus generating echo-canceled samples at the processing rate;

an equalizer, coupled to filter the echo-canceled samples so as to generate equalized samples at the input symbol rate; and

decision circuitry, coupled to process the equalized samples so as to generate a stream of estimated symbols at the input symbol rate,

wherein the processing rate is a non-integer rational multiple of at least one of the input symbol rate and the output symbol rate, and

wherein the processing rate is substantially equal to 1.5 times the input symbol rate.

13. A digital data transceiver, comprising:

a transmitter, which is adapted to transmit an output signal over a communication channel based on a stream of output symbols at an output symbol rate;

an echo canceller, coupled to generate, responsive to the output symbols, echo cancellation values at a predetermined processing rate; and

a receiver, which comprises;

a front end, coupled to receive an input signal over the communication channel, the signal representing a stream of transmitted symbols with a given input symbol rate, and to generate, responsive thereto, a stream of input samples at the processing rate;

an adder, coupled to sum the input samples with the echo cancellation values, thus generating echo-canceled samples at the processing rate;

an equalizer, coupled to filter the echo-canceled samples so as to generate equalized samples at the input symbol rate; and

decision circuitry, coupled to process the equalized samples so as to generate a stream of estimated symbols at the input symbol rate,

wherein the processing rate is a non-integer rational multiple of at least one of the input symbol rate and the output symbol rate,

wherein the echo canceller comprises;

a first plurality of taps, arranged in series to receive a succession of the output symbols at the output symbol rate;

a second plurality of registers operative to store a first number of groups of filter coefficients, such that each group contains one filter coefficient for each of the first plurality of taps;

a third plurality of multipliers, coupled to multiply the samples in the taps by the corresponding filter coefficients so as to generate respective tap outputs for each one of the groups;

one or more summers, coupled to sum the tap outputs for each one of the groups, thus generating respective group echo cancellation partial sums; and

at least one switch, operative to select among the groups at the processing rate so that the partial sums are output in alternation as the echo cancellation values at the processing rate.

14. A digital data transceiver, comprising:

a transmitter, which is adapted to transmit an output signal over a communication channel based on a stream of output symbols at an output symbol rate;

an echo canceller, coupled to generate, responsive to the output symbols, echo cancellation values at a predetermined processing rate; and

a receiver, which comprises;

a front end, coupled to receive an input signal over the communication channel, the signal representing a stream of transmitted symbols with a given input symbol rate, and to generate, responsive thereto, a stream of input samples at the processing rate;

an adder, coupled to sum the input samples with the echo cancellation values, thus generating echo-canceled samples at the processing rate;

an equalizer, coupled to filter the echo-canceled samples so as to generate equalized samples at the input symbol rate; and

decision circuitry, coupled to process the equalized samples so as to generate a stream of estimated symbols at the input symbol rate,

wherein the processing rate is a non-integer rational multiple of at least one of the input symbol rate and the output symbol rate, and

wherein the output symbol rate is different from the input symbol rate.

15. A digital data transceiver, comprising:

a transmitter, which is adapted to transmit an output signal over a communication channel based on a stream of output symbols at an output symbol rate;

an echo canceller, coupled to generate, responsive to the output symbols, echo cancellation values at a predetermined processing rate; and

a receiver, which comprises;

a front end, coupled to receive an input signal over the communication channel, the signal representing a stream of transmitted symbols with a given input symbol rate, and to generate, responsive thereto, a stream of input samples at the processing rate;

an adder, coupled to sum the input samples with the echo cancellation values, thus generating echo-canceled samples at the processing rate;

an equalizer, coupled to filter the echo-canceled samples so as to generate equalized samples at the input symbol rate; and

decision circuitry, coupled to process the equalized samples so as to generate a stream of estimated symbols at the input symbol rate,

wherein the processing rate is a non-integer rational multiple of at least one of the input symbol rate and the output symbol rate, and

wherein the echo canceller comprises an adaptive filter, having filter coefficients that are adjusted automatically responsive to an error signal based on the input samples.

16. A method for processing an input signal received over a communication channel, in the presence of an echo signal based on a stream of output symbols provided at an output symbol rate, the input signal representing a stream of transmitted symbols with a given input symbol rate, the method comprising:

determining echo cancellation values at a predetermined processing rate, responsive to the stream of output symbols;

sampling the input signal to generate a stream of received samples at the processing rate;

summing the received samples with the echo cancellation values, thus generating echo-canceled samples at the processing rate;

equalizing the echo-canceled samples so as to generate a stream of forward-equalized samples at the input symbol rate; and

applying a decision process to the forward-equalized samples so as to generate a stream of estimated symbols at the input symbol rate,

wherein the processing rate is a non-integer rational multiple of at least one of the input symbol rate and the output symbol rate.

17. A method according to 16, wherein the input signal comprises a Digital Subscriber Line (DSL) signal.

18. A method according to claim 16, wherein processing the received samples comprises equalizing the received samples adaptively, using filter coefficients that are adjusted automatically responsive to an error signal derived from the estimated symbols.

19. A method according to claim 16, wherein the output symbol rate is substantially equal to the input symbol rate.

20. A method for processing an input signal received over a communication channel, in the presence of an echo signal based on a stream of output symbols provided at an output symbol rate, the input signal representing a stream of transmitted symbols with a given input symbol rate, the method comprising:

determining echo cancellation values at a predetermined processing rate, responsive to the stream of output symbols;

sampling the input signal to generate a stream of received samples at the processing rate;

summing the received samples with the echo cancellation values, thus generating echo-canceled samples at the processing rate;

equalizing the echo-canceled samples so as to generate a stream of forward-equalized samples at the input symbol rate; and

applying a decision process to the forward-equalized samples so as to generate a stream of estimated symbols at the input symbol rate,

wherein the processing rate is a non-integer rational multiple of at least one of the input symbol rate and the output symbol rate, and

wherein the processing rate is substantially equal to 1.5 times the input symbol rate.

21. A method for processing an input signal received over a communication channel, in the presence of an echo signal based on a stream of output symbols provided at an output symbol rate, the input signal representing a stream of transmitted symbols with a given input symbol rate, the method comprising:

determining echo cancellation values at a predetermined processing rate, responsive to the stream of output symbols;

sampling the input signal to generate a stream of received samples at the processing rate;

summing the received samples with the echo cancellation values, thus generating echo-canceled samples at the processing rate;

equalizing the echo-canceled samples so as to generate a stream of forward-equalized samples at the input symbol rate; and

applying a decision process to the forward-equalized samples so as to generate a stream of estimated symbols at the input symbol rate,

wherein the processing rate is a non-integer rational multiple of at least one of the input symbol rate and the output symbol rate, and

wherein processing the received samples comprises equalizing the received samples adaptively, using filter coefficients that are adjusted automatically responsive to an error signal derived from the estimated symbols, and

wherein equalizing the received samples adaptively comprises substituting a sequence of known training symbols for the estimated symbols during an initial adjustment of the filter coefficients.

22. A method for processing an input signal received over a communication channel, in the presence of an echo signal based on a stream of output symbols provided at an output symbol rate, the input signal representing a stream of transmitted symbols with a given input symbol rate, the method comprising:

determining echo cancellation values at a predetermined processing rate, responsive to the stream of output symbols;

sampling the input signal to generate a stream of received samples at the processing rate;

summing the received samples with the echo cancellation values, thus generating echo-canceled samples at the processing rate;

equalizing the echo-canceled samples so as to generate a stream of forward-equalized samples at the input symbol rate; and

applying a decision process to the forward-equalized samples so as to generate a stream of estimated symbols at the input symbol rate,

wherein the processing rate is a non-integer rational multiple of at least one of the input symbol rate and the output symbol rate, and

wherein processing the received samples comprises equalizing the received samples adaptively, using filter coefficients that are adjusted automatically responsive to an error signal derived from the estimated symbols, and

wherein equalizing the received samples adaptively comprises;

providing at least first and second groups of filter coefficients for use in equalizing the received samples;

initially determining values of the filter coefficients in the first group; and

subsequently determining values of the filter coefficients in the second group responsive to the values determined in the first group.

23. A method for processing an input signal received over a communication channel, in the presence of an echo signal based on a stream of output symbols provided at an output symbol rate, the input signal representing a stream of transmitted symbols with a given input symbol rate, the method comprising:

determining echo cancellation values at a predetermined processing rate, responsive to the stream of output symbols;

sampling the input signal to generate a stream of received samples at the processing rate;

summing the received samples with the echo cancellation values, thus generating echo-canceled samples at the processing rate;

equalizing the echo-canceled samples so as to generate a stream of forward-equalized samples at the input symbol rate; and

applying a decision process to the forward-equalized samples so as to generate a stream of estimated symbols at the input symbol rate,

wherein the processing rate is a non-integer rational multiple of at least one of the input symbol rate and the output symbol rate, and

wherein determining the echo cancellation values comprises determining the values adaptively, using filter coefficients that are adjusted automatically responsive to the echo-canceled samples.

24. A method for processing an input signal received over a communication channel, in the presence of an echo signal based on a stream of output symbols provided at an output symbol rate, the input signal representing a stream of transmitted symbols with a given input symbol rate, the method comprising:

determining echo cancellation values at a predetermined processing rate, responsive to the stream of output symbols;

sampling the input signal to generate a stream of received samples at the processing rate;

summing the received samples with the echo cancellation values, thus generating echo-canceled samples at the processing rate;

equalizing the echo-canceled samples so as to generate a stream of forward-equalized samples at the input symbol rate; and

applying a decision process to the forward-equalized samples so as to generate a stream of estimated symbols at the input symbol rate,

wherein the processing rate is a non-integer rational multiple of at least one of the input symbol rate and the output symbol rate, and

wherein the output symbol rate is different from the input symbol rate.

25. An adaptive digital filter for processing an input signal having an input sample rate, comprising:

a first plurality of taps, arranged in series to receive a succession of samples of the input signal at the input sample rate;

a second plurality of registers operative to store a number of groups of filter coefficients, which are automatically adjustable during operation of the filter, such that each group contains one filter coefficient for each of the plurality of taps;

a third plurality of multipliers, coupled to multiply the samples in the taps by the corresponding filter coefficients so as to generate respective tap outputs for each one of the groups;

one or more summers, coupled to sum the tap outputs for each one of the groups, thus generating respective group partial sums; and

at least one switch, operative to select among the groups at a selection rate that is substantially equal to the product of the input sample rate times the number of groups divided by a selected integer, so as to generate an output signal comprising a sequence of the partial sums having an output sample rate that is a non-integer rational fraction of the input sample rate; and

an error signal input line, coupled to provide to the filter an error signal derived from the output signal for use in automatic adjustment of the filter coefficients.

26. A filter according to claim 25, wherein the third plurality of multipliers comprises a number of multipliers for each of the taps that is equal to the number of groups, such that one of the multipliers is assigned respectively to generate each of the tap outputs.

27. A filter according to claim 25, wherein the one or more summers comprise a single summer.

28. A filter according to claim 27, wherein the at least one switch comprises a respective switch for each of the first plurality of taps, which switch is coupled to select the filter coefficient by which to multiply the sample in the tap.

29. A filter according to claim 25, wherein the number of groups is chosen to be one of the values two and three, and wherein the selected integer is equal to the other of the values.

30. A filter according to claim 25, wherein the filter comprises a feed-forward equalizer, and the output signal comprises equalized signal values, and wherein the error signal is responsive to a difference between the equalized signal values and estimated symbols derived from the equalized signal values.

31. An adaptive digital filter for processing an input signal having an input sample rate, comprising:

a first plurality of taps, arranged in series to receive a succession of samples of the input signal at the input sample rate;

a second plurality of registers operative to store a number of groups of filter coefficients, which are automatically adjustable during operation of the filter, such that each group contains one filter coefficient for each of the plurality of taps;

a third plurality of multipliers, coupled to multiply the samples in the taps by the corresponding filter coefficients so as to generate respective tap outputs for each one of the groups;

one or more summers, coupled to sum the tap outputs for each one of the groups, thus generating respective group partial sums; and

at least one switch, operative to select among the groups at a selection rate that is substantially equal to the product of the input sample rate times the number of groups divided by a selected integer, so as to generate an output signal comprising a sequence of the partial sums having an output sample rate that is a non-integer rational fraction of the input sample rate; and

an error signal input line, coupled to provide to the filter an error signal derived from the output signal for use in automatic adjustment of the filter coefficients,

wherein the one or more summers comprise a respective summer for each of the groups, and the at least one switch is operative to select among the respective group partial sums.

32. An adaptive digital filter for processing an input signal having an input sample rate, comprising:

a first plurality of taps, arranged in series to receive a succession of samples of the input signal at the input sample rate;

a second plurality of registers operative to store a number of groups of filter coefficients, which are automatically adjustable during operation of the filter, such that each group contains one filter coefficient for each of the plurality of taps;

a third plurality of multipliers, coupled to multiply the samples in the taps by the corresponding filter coefficients so as to generate respective tap outputs for each one of the groups;

one or more summers, coupled to sum the tap outputs for each one of the groups, thus generating respective group partial sums; and

at least one switch, operative to select among the groups at a selection rate that is substantially equal to the product of the input sample rate times the number of groups divided by a selected integer, so as to generate an output signal comprising a sequence of the partial sums having an output sample rate that is a non-integer rational fraction of the input sample rate; and

an error signal input line, coupled to provide to the filter an error signal derived from the output signal for use in automatic adjustment of the filter coefficients,

wherein the third plurality of multipliers comprises a number of multipliers for each of the taps that is equal to the number of groups, such that one of the multipliers is assigned respectively to generate each of the tap outputs, and

wherein the multipliers operate at a rate that is substantially equal to the input sample rate divided by the selected integer.

33. An adaptive digital filter for processing an input signal having an input sample rate, comprising:

a first plurality of taps, arranged in series to receive a succession of samples of the input signal at the input sample rate;

a second plurality of registers operative to store a number of groups of filter coefficients, which are automatically adjustable during operation of the filter, such that each group contains one filter coefficient for each of the plurality of taps;

a third plurality of multipliers, coupled to multiply the samples in the taps by the corresponding filter coefficients so as to generate respective tap outputs for each one of the groups;

one or more summers, coupled to sum the tap outputs for each one of the groups, thus generating respective group partial sums; and

at least one switch, operative to select among the groups at a selection rate that is substantially equal to the product of the input sample rate times the number of groups divided by a selected integer, so as to generate an output signal comprising a sequence of the partial sums having an output sample rate that is a non-integer rational fraction of the input sample rate; and

an error signal input line, coupled to provide to the filter an error signal derived from the output signal for use in automatic adjustment of the filter coefficients,

wherein each filter coefficient is automatically adjusted at an adjustment rate that is substantially equal to the input sample rate divided by the selected integer.

34. An adaptive digital filter for processing an input signal having an input sample rate, comprising:

a first plurality of taps, arranged in series to receive a succession of samples of the input signal at the input sample rate;

a second plurality of registers operative to store a number of groups of filter coefficients, which are automatically adjustable during operation of the filter, such that each group contains one filter coefficient for each of the plurality of taps;

a third plurality of multipliers, coupled to multiply the samples in the taps by the corresponding filter coefficients so as to generate respective tap outputs for each one of the groups;

one or more summers, coupled to sum the tap outputs for each one of the groups, thus generating respective group partial sums; and

at least one switch, operative to select among the groups at a selection rate that is substantially equal to the product of the input sample rate times the number of groups divided by a selected integer, so as to generate an output signal comprising a sequence of the partial sums having an output sample rate that is a non-integer rational fraction of the input sample rate; and

an error signal input line, coupled to provide to the filter an error signal derived from the output signal for use in automatic adjustment of the filter coefficients,

wherein the filter coefficients are automatically adjusted using a least-mean-square method.

35. An adaptive digital filter for processing an input signal having an input sample rate, comprising:

a first plurality of taps, arranged in series to receive a succession of samples of the input signal at the input sample rate;

a second plurality of registers operative to store a number of groups of filter coefficients, which are automatically adjustable during operation of the filter, such that each group contains one filter coefficient for each of the plurality of taps;

a third plurality of multipliers, coupled to multiply the samples in the taps by the corresponding filter coefficients so as to generate respective tap outputs for each one of the groups;

one or more summers, coupled to sum the tap outputs for each one of the groups, thus generating respective group partial sums; and

at least one switch, operative to select among the groups at a selection rate-that is substantially equal to the product of the input sample rate times the number of groups divided by a selected integer, so as to generate an output signal comprising a sequence of the partial sums having an output sample rate that is a non-integer rational fraction of the input sample rate;

an error signal input line, coupled to provide to the filter an error signal derived from the output signal for use in automatic adjustment of the filter coefficients; and

an echo cancellation filter,

wherein the output signal comprises echo estimation values, and wherein the error signal provides an estimate of a residual echo remaining following subtraction of the echo estimation values from a received signal.

36. A method for adaptive digital filtering of an input signal having an input sample rate, comprising:

receiving a plurality of successive samples of the input signal at the input sample rate;

automatically computing a number of groups of filter coefficients responsive to an error signal, each group containing one filter coefficient for each of the plurality of samples;

multiplying the samples by the corresponding filter coefficients so as to generate for each of the samples respective sample outputs for all of the groups;

summing the sample outputs belonging to each one of the groups, thus generating respective group partial sums for all of the groups;

selecting among the groups at a selection rate that is substantially equal to the product of the input sample rate times the number of groups divided by a selected integer, so as to generate an output signal comprising a sequence of the partial sums having an output sample rate that is a non-integer rational fraction of the input sample rate; and

deriving the error signal from the output signal.

37. A method according to claim 36, wherein the number of groups is chosen to be one of the values two and three, and wherein the selected integer is equal to the other of the values.

38. A method for adaptive digital filtering of an input signal having an input sample rate, comprising:

receiving a plurality of successive samples of the input signal at the input sample rate;

automatically computing a number of groups of filter coefficients responsive to an error signal, each group containing one filter coefficient for each of the plurality of samples;

multiplying the samples by the corresponding filter coefficients so as to generate for each of the samples respective sample outputs for all of the groups;

summing the sample outputs belonging to each one of the groups, thus generating respective group partial sums for all of the groups;

selecting among the groups at a selection rate that is substantially equal to the product of the input sample rate times the number of groups divided by a selected integer, so as to generate an output signal comprising a sequence of the partial sums having an output sample rate that is a non-integer rational fraction of the input sample rate; and

deriving the error signal from the output signal,

wherein automatically adjusting the filter coefficients comprises adjusting the filter coefficients using a least-mean-square method.