MINIMIZATION OF OUTPUT ERRORS OF FIR DIGITAL-FILTERS BY MULTIPLE DECOMPOSITIONS OF SIGNAL WORD

FIR digital fillers composed of parallel multiple subfilters are propo sed. A binary expression of an input signal is decomposed into multipl e shorter words, which drive the subfilters having different length. T he output error is evaluated by mean squared and maximum spectra. A fa st algorithm is also proposed to determine optimal filler lengths and coefficients of subfilters. Many examples confirm that the proposed fi lters generate smaller output errors than conventional fillers under t he condition of specified number of multiplications and additions in f ilter operations. Further, multiplier and adder structures (MAS) to pe rform the operations of the proposed filters are also presented. The n umber of gates used in the proposed MAS and its critical path are esti mated. The effectiveness of the proposed MAS is confirmed.