Non-recursive digital filter
First Claim
1. A non-recursive digital filter having predetermined N filter coefficients hi (i=0 to N-1) to perform ##EQU11## in response to a sequence x(n-i) of N consecutive binary coded words, where h*j represent the coefficients of the absolute values of hi arranged in the order of magnitude;
- x*(n-j) represent x(n-i) corresponding to h*j ;
sign (hi) represent the polarities of h;
Δ
j represent differences between adjoining coefficients h*j ;
x*(n-j) represent the products of sign (hi) and x*(n-j); and
##EQU12## said filter comprising;
arranging means for arranging said sequence x(n-i) corresponding to the coefficients h*j to obtain x*(n-j);
first means for obtaining the product of each of the outputs x*(n-j) given from the arranging means and the corresponding polarity of each of the coefficients hi ;
second means for accumulating the outputs x*(n-j) given from said first means to obtain u(n-j);
coefficient generating means for producing said differences Δ
j ;
third means for multiplying the outputs u(n-j) supplied from said second means by said differences Δ
j ; and
fourth means for accumulating outputs Δ
j u(n-j) given from said third means.
1 Assignment
0 Petitions
Accused Products
Abstract
A non-recursive digital filter has N predetermined filter coefficients to perform a specific computation in response to a sequence of consecutive binary codewords. The absolute values of the computation coefficients are arranged in an order of magnitude. The filter arranges the sequence to obtain a sequence corresponding to the arranged coefficients. A product is obtained for each sequence, which is given by the arrangement and by the corresponding polarity of each of the coefficients. The outputs given by these products are accumulated and are multiplied by certain differences.
14 Citations
4 Claims
-
1. A non-recursive digital filter having predetermined N filter coefficients hi (i=0 to N-1) to perform ##EQU11## in response to a sequence x(n-i) of N consecutive binary coded words, where h*j represent the coefficients of the absolute values of hi arranged in the order of magnitude;
- x*(n-j) represent x(n-i) corresponding to h*j ;
sign (hi) represent the polarities of h;
Δ
j represent differences between adjoining coefficients h*j ;
x*(n-j) represent the products of sign (hi) and x*(n-j); and
##EQU12## said filter comprising;
arranging means for arranging said sequence x(n-i) corresponding to the coefficients h*j to obtain x*(n-j);
first means for obtaining the product of each of the outputs x*(n-j) given from the arranging means and the corresponding polarity of each of the coefficients hi ;
second means for accumulating the outputs x*(n-j) given from said first means to obtain u(n-j);
coefficient generating means for producing said differences Δ
j ;
third means for multiplying the outputs u(n-j) supplied from said second means by said differences Δ
j ; and
fourth means for accumulating outputs Δ
j u(n-j) given from said third means.
- x*(n-j) represent x(n-i) corresponding to h*j ;
-
2. A non-recursive digital filter having predetermined N filter coefficients hi (i=0 to N-1) to perform ##EQU13## in response to a sequence x(n-i) of N consecutive binary coded words, where h*j represent the coefficients of the absolute values of hi arranged in the order of magnitude;
- x*(n-j) represent x(n-i) corresponding to h*j;
sign(hi) represent the polarities of hi ;
Δ
j represent differences between adjoining coefficients h*j ;
x*(n-j) represent the products of sign(hi) and x*(n-j);
##EQU14## Cl represent differences between adjoining coefficients Δ
l * obtained by arranging Δ
j in the order of magnitude;
##EQU15## and u*(n-l) represent u(n-j) corresponding to Δ
*l, said filter comprising;first arranged means for arranging said sequence x(n-i) corresponding to the coefficients h*j to obtain x*(n-j);
first arithmetic means for obtaining the product each of outputs x*(n-j) given from the first arranging means and the corresponding polarity of each of the coefficients hi ;
first adding means for accumulating the outputs x*(n-j) given from the first arithmetic means to obtain u(n-j);
second arranging means for arranging the outputs u(n-j) supplied from said first adding means corresponding to the coefficients Δ
*l to produce u*(n-l);
second adding means for accumulating the outputs fed from the second arranging means to obtain u(n-l);
coefficient generating means for producing said coefficients Cl ;
second arithmetic means for multiplying the outputs u(n-l) supplied from the second adding means by said coefficients Cl ; and
third adding means for accumulating the outputs Cl u(n-l) given from the second arithmetic means.
- x*(n-j) represent x(n-i) corresponding to h*j;
-
3. A non-recursive digital filter comprising a pair of random access memory means, each of said random access memory means including means for processing digital signal code words having different coefficients by arranging said digital signal code words according to the coefficients of said digital signal code words and the differences between said digital signal code words, read only memory means for driving each of said random access memory means through a sequence of operational steps, selector means responsive to said read only memory means for alternately operating said random access memory means, and addressing means for successively addressing said random access memory means for successively scanning the contents of said random access memory means, means for multiplying an output of an operating one of the random access memory means selected by said selector means in order to establish a digital sign, means responsive to said multiplying means for accumulating a codeword from said products of the multiplications, and means for multiplying the output of said accumulating means by differences between absolute values of said coefficients, as new coefficients.
-
4. A method of filtering digital signals comprising the steps of:
-
a. converting analog signals into digital signals to provide binary coded words, said digital signals having different coefficients; b. arranging said digital signals in a sequence corresponding to said coefficients; c. multiplying said arranged signals by a signal representing a sign of said digital signals; d. accumulating products of the multiplication of step c; e. multiplying said accumulated digital signals by differences between absolute values of said coefficients, as new coefficients; and f. accumulating the products of the multiplication of step e.
-
Specification