US 5,479,454 A · Digital filter with improved numerical precision

Abstract

A digital filter characterized by a plurality of filter coefficients, positive ones of the coefficients being represented by unsigned binary integers and negative ones of the coefficients being represented by 2's complement binary integers, comprising an input for receiving a digital input signal comprising a first predetermined number of bits; means for truncating or rounding at least one leading "0" or trailing "0" of the positive ones of the coefficients and at least one leading "1" or trailing "0" of the negative ones of the coefficients, such that significance of the integers is removed; means for successively delaying the digital input signal to produce successively delayed versions of the digital input signal; means for multiplying the successively delayed versions of the digital input signal by respective ones of the truncated representations of the respective ones of the plurality of filter coefficients and in response generating a plurality of intermediate product integers; means for restoring significance to the intermediate product integers; and means for summing the intermediate product integers and in response generating a digital output signal.

Claims12

Dependent claims

A digital filter characterized by a plurality of filter coefficients, positive ones of said coefficients being represented by unsigned binary integers and negative ones of said coefficients being represented by 2's complement binary integers, comprising:

a) an input for receiving a digital input signal comprising a first predetermined number of bits;
b) means for providing one of either truncated or rounded coefficients by truncating or rounding at least one leading "0" or trailing "0" of said positive ones of said coefficients and at least one leading "1" or trailing "0" of said negative ones of said coefficients, such that significance of said integers is removed;
c) means for successively delaying said digital input signal to produce successively delayed versions of said digital input signal;
d) means for multiplying said successively delayed versions of said digital input signal by respective ones of said truncated representations of said respective ones of said plurality of filter coefficients and in response generating a plurality of intermediate product integers;
e) means for restoring significance to said intermediate product integers;
and
f) means for summing said intermediate product integers having restored significance and in response generating a digital output signal.

4 more elements

Dependencies are inferred from the claim text.

Digital filter with improved numerical precision

Life of this patent

Digital filter with improved numerical precision

Life of this patent

Abstract

Claims12