Expand FFT Capabilities to Non-power of Two Length Data Sequences

The fast Fourier transform (FFT) is perhaps the most powerful algorithm in all of digital signal processing. However, the traditional radix-2 FFT algorithm can only be applied to data sequences where the number of samples in the sequence is an integer power of two, i.e., 8, 16, 32, 64, etc. This paper provides a signal processing trick that extends the FFT's applicability to data sequences whose lengths are integer multiples of a power of two, for example, 3x16=48 or perhaps 5x32=160 sample sequences.