Digital filter

Claim 1

A digital filter which performs a filter mathematical operation on input 2N sample values Zi expressed by binary codes in which an output sample y(n) of said filter is expressed by the following equation ##EQU20## where the magnitude of ##EQU21## represents a product of a sign of Ai and Zi, aij represents a partial coefficient obtained by dviding Ai with k, k is the number of groups of partial bits each including M bits, l is an integer between O and 2 M -1, S jl represents a total sum of Zi* corresponding to ##EQU22## said digital filter comprising: first memory means responsive to a plurality of input sample values Zi for storing continuous 2N sample values; second memory means for storing partial coefficients aij (j<k) which are obtained by dividing each coefficient Ai with K and sign of Ai; first operational circuit means responsive to said first and second memory means for determining a total sum S jl of sample values Zi stored in said first memory means and relating to l=aij of a partial coefficient aij read out from said second memory means; shifter means responsive to said first operational circuit means for effectively multiplying by shifting an output S jl of the first operational circuit means by 2 Mo-jM where Mo and M are positive integers; first adder means responsive to an output from the shifter means for cumulatively adding k times an output 2 Mo-jM ·S Jl of the shifter means to obtain ##EQU23## second adder means responsive to the results of said first adder means for cumulatively adding (2 M -1m) times the output of the first adder means to obtain ##EQU24## and third adder means responsive to the results of said second adder means for cumulatively adding (2 M -2) times an output of said second adder means to obtain an output y(n) from said filter.

Claim 10

A method of digitally filtering an input sample signal having 2N sample values Zi exptessed by binary codes in a digital filter employing memories, shifters and adders wherein an output sample signal y(n) obtained from said filtering is expressed by the equation: ##EQU25## where the magnitude of ##EQU26## represents a product of a sign of Ai and Zi, aij represents a partial coefficient obtained by dividing Ai with k, k is the number of groups of partial bits each including M bits, l is an integer between O and 2 M -1, S jl represents a total sum of Zi* corresponding to E1 ? ##STR1## said method comprising: storing in a first memory continuous 2N sample values Zi; storing in a second memory partial coefficients aij (j<k) which are obtained by dividing each coefficient Ai with k and the sign of Ai; combining the sample value Zi from the first memory with the partial coefficients aij from the second memory to derive a total sum signal S jl of the sample values to l=aij; shifting the output sum signal S jl by 2 Mo-jM where Mo and M are positive integers; cumulatively adding k times the output 2 Mo-jM ·S jl from the shifting operation to derive a first interim output given by ##EQU27## cumulatively adding (2 M -1-m) times the interim output S l to obtain ##EQU28## and cumulatively adding (2 M -2) times the second interim output to derive the desired output y(n).

Claim 11

A method of digitally filtering an input sample signal having 2N sample values Zi exptessed by binary codes in a digital filter employing memories, shifters and adders wherein an output sample signal y(n) obtained from said filtering is expressed by the equation: ##EQU25## where the magnitude of ##EQU26## represents a product of a sign of Ai and Zi, aij represents a partial coefficient obtained by dividing Ai with k, k is the number of groups of partial bits each including M bits, l is an integer between O and 2 M -1, S jl represents a total sum of Zi* corresponding to E1 ? ##STR1## said method comprising: storing in a first memory continuous 2N sample values Zi; storing in a second memory partial coefficients aij (j<k) which are obtained by dividing each coefficient Ai with k and the sign of Ai; combining the sample value Zi from the first memory with the partial coefficients aij from the second memory to derive a total sum signal S jl of the sample values to l=aij; shifting the output sum signal S jl by 2 Mo-jM where Mo and M are positive integers; cumulatively adding k times the output 2 Mo-jM ·S jl from the shifting operation to derive a first interim output given by ##EQU27## cumulatively adding (2 M -1-m) times the interim output S l to obtain ##EQU28## and cumulatively adding (2 M -2) times the second interim output to derive the desired output y(n).