I would like to average several FFT's of pink noise to get a better idea of the response of my system. The individual FFT's vary greatly from one to the next. It would be easier to correct the system for a flat frequency response if the FFT did not change so much from one try to the next on a long file.