In general regenerating sound from frequency spectra takes
considerable time when the amount of frequency spectra is great due to
deal with the convolution calculation. The method used in the spectral
analyzer introduce in the following site:
http://www.geocities.com/square17320...ad/ferftad.htm does not
use the convolution for Fourier transform. So the efficiency of DFT is
improved and it makes the real time Fourier transform and its inverse
transform more portable.