Abstract: A high-throughput filtering method and corresponding apparatus in which each output sample sn is determined by weighting N input samples xi by multiplying the input samples by N coefficients, where the N coefficients are each decomposed into a set of coordinates, each coordinate corresponding to P bits of the coefficient. Essentially, input samples are multiplied by the coordinates of each of the coefficients in a pre-processing operation where the output signals corresponding to the resulting products for a coefficient are added in an adding block. A plurality of adding blocks are provided corresponding respectively to the coefficients such that each block includes the summation result of the products for a corresponding coefficient. The summation result from a previous adding block is added to the summation result of a current block, and the result of this addition is supplied to the next adding block. The output signal is obtained in this manner.