Why use fourier transforms?
Fourier transforms will show you what waveforms can be combined to form the the signal that you are analysing. This usually takes the form of a sine wave combined with a number of harmonics. Its use in steganography could very well be that changing the harmonics used in simulating a square wave signal would look like simple interference. However, subtracting the fourier transform of the square wave from the signal would actually reveal the hidden signal. As for morphing audio I am not really sure that fourier transforms are that useful Audio being already a complex waveform. However I suppose they could be used to reveal the frequencies required to replicate the original audio.
