The Discrete Fourier Transform is widely used in signal processing for the computation of correlation or convolution equations.
The aim of this project has been to study some alternatives that may be more suited for the processing of real signals. The definition of a class of transforms with pseudo-cyclic convolution properties has led to the discovery of a modified Fourier transform that allows efficient computation of aperiodic convolutions. In contrast to conventional Fourier transform techniques, this method does not require zero padding to suppress wraparound or aliasing errors. It requires less memory and less computation time than conventional complex algorithms.
