Bruun's FFT algorithm

Bruun's algorithm is a fast Fourier transform that takes an unusual route: instead of splitting the input into even and odd parts, it works by recursively factoring a polynomial. G. Bruun proposed it for sizes that are powers of two in 1978, and H. Murakami generalized it to arbitrary even composite sizes in 1996. Because its operations involve only real coefficients until the final stage of computation, it was first suggested as an efficient way to compute the discrete Fourier transform of real-valued data. It has not been widely adopted, since adaptations of the ordinary Cooley-Tukey algorithm handle real data at least as efficiently, and there is evidence it may be intrinsically less accurate under finite numerical precision. It remains of interest as an alternative framework that can express both itself and Cooley-Tukey.