Noniterative real-time design of selective linear phase filters

A novel approach for designing and realizing variable linear phase finite impulse response (FIR) filters is presented. The proposed filter structure is a cascade of a prefilter and corrector. The author concentrates on selective and computationally efficient filter structures exploiting the generalized recursive running sums (GRRSs) as prefilters. The proposed analytical window design method for the selective filter structures is very simple and fast and it can be implemented easily in hardware or software. The resulting linear phase filters exploiting the GRRSs were superior to the corresponding optimum direct from FIR designs having the same specifications in terms of additions and multiplications per sample.<>