Interleaved method for parallel implementation of the Fast Fourier Transform

摘要

The present invention generally relates to a method for computing a Fast Fourier Transform (FFT). In one embodiment, the present invention relates to an interleaved method for computing a Fast Fourier Transform (FFT).

主权项

1. A method for processing a vector having at least one dimension comprising the steps of:
obtaining a vector having at least one dimension, the vector having one or more vector values; and computing a Fast Fourier Transform (FFT) of the vector, comprising the steps of: providing a first processor and a second processor; inputting at least one set of vector values of the vector into either of the first or second processors; computing at least one sum and at least one difference from the input set of vector values from the vector; calculating from the results of the computing step at least one set of halves of an even vector, such that the even vector has a first half and a second half; calculating from the results of the computing step at least one set of halves of an odd vector, such that the odd vector has a first half and a second half; computing in parallel simultaneously, at least one discrete cosine transform (DCT) of either of the first half or the second half of the even vector by the first processor, and at least one discrete sine transform (DST) of either of the first half or the second half of the odd vector by the second processor; processing the output of the discrete cosine transform (DCT) and the output of the discrete sine transform (DST) from the second computing step so that they are made twice as large; combining the results of the processing step; outputting at least one value from the combining step with a correct magnitude at one or more desired locations; and interleaving the at least one value to compute the Fast Fourier Transform of the vector.