Digital signal processor architecture

Claim 2

A linear-phase finite impulse response digital filter comprising means for storing a plurality of filter tap coefficients each having a predetermined magnitude, means for receiving a sequence of digital samples of a signal, each of said samples being associated with a respective one of said coefficients, means for combining said samples into pairs of samples such that the coefficients associated with the samples in a pair of samples are of equal magnitude, and means for generating and outputting a digitally filtered representation of said samples as a function of each of said pairs of samples and the magnitude of one of their respective coefficients, said filtered representation being symmetrical when the samples in each of said pairs of samples are summed and being antisymmetrical when one sample is subtracted from the other.

Claim 3

The filter set forth in claim 1 wherein said generating and outputting means further includes means for generating said representation as a function of both of said symmetrical and antisymmetrical representations when said filtered representation is to be nonsymmetrical.

Claim 4

A linear-phase finite impulse response digital filter comprising, means for storing a plurality of filter tap coefficients each having a predetermined magnitude, means for recieving a sequence of m digital samples of a signal, where m≧1, and means for generating and outputting a digitally filtered representation of said received samples by (a) combining said samples into pairs of samples such that the most recent sample is paired with the m th previous sample of said sequence, the next most recent sample is paired with the (m-1) st previous sample of said sequence, etc., (b) multiplying each of said pairs of samples by a respective one of said coefficients, and (c) accumulating the product of each such multiplication, said means for generating and outputting being arranged to sum said samples in each of said pairs of samples when said filtered representation is to be symmetrical and to subtract one sample from the other in each of said pairs of samples when said filtered representation is to be antisymmetrical.

Claim 5

A linear-phase finite impulse response digital filter comprising, means for storing a plurality of filter tap coefficients each having a predetermined magnitude, means for recieving a sequence of m digital samples of a signal, where m≧1, and means for generating and outputting a digitally filtered representation of said received samples by (a) combining said samples into pairs of samples such that the most recent sample is paired with the m th previous sample of said sequence, the next most recent sample is paired with the (m-1) st previous sample of said sequence, etc., (b) multiplying each of said pairs of samples by a respective one of said coefficients, and (c) accumulating the product of each such multiplication, said means for generating and outputting being arranged to sum said samples in each of said pairs of samples when said filtered representation is to be symmetrical and to subtract one sample from the other in each of said pairs of samples when said filtered representation is to be antisymmetrical.

Claim 6

The filter set forth in claim 3 wherein said filtered representation is nonsymmetrical when said multiplications include the pairs of samples that are summed and the pairs of samples that are subtracted.

Claim 7

A linear-phase finite impulse response digital filter comprising means for storing a plurality of filter tap coefficients each having a predetermined magnitude, means for receiving a sequence of m samples of a signal, where m≧1, means for generating a plurality of signal vector elements from predetermined pairs of said samples, said pairs of samples being formed such that the most recent sample is paired with the m th previous sample, the next most recent sample is paired with the (m-1) st previous sample, etc., and means for generating and outputting a filtered representation of said received samples by multiplying each of said signal vector elements by a respective one of said coefficients and accumulating the product of each such multiplication.

Claim 8

A linear-phase finite impulse response digital filter comprising means for storing a plurality of filter tap coefficients each having a predetermined magnitude, means for receiving a sequence of m samples of a signal, where m≧1, means for generating a plurality of signal vector elements from predetermined pairs of said samples, said pairs of samples being formed such that the most recent sample is paired with the m th previous sample, the next most recent sample is paired with the (m-1) st previous sample, etc., and means for generating and outputting a filtered representation of said received samples by multiplying each of said signal vector elements by a respective one of said coefficients and accumulating the product of each such multiplication.

Claim 9

The filter set forth in claim 5 wherein said plurality of signal vector elements includes a first group and a second group of signal vector elements, said first group being generated by adding the samples in each of said pairs of samples and the second group being generated by subtracting one sample from the other in each of said pairs of samples.

Claim 10

The filter set forth in claim 6 wherein said filtered representation is symmetrical when said signal vector elements are said first group of elements and is antisymmetrical when they are said second group elements.

Claim 11

The filter set forth in claim 6 wherein said filtered representation is nonsymmetrical when said signal vector elements include said first and second groups of elements.