4.3. Using FFT to De-Skew

When real-world data consists of strongly correlated bits, it may still contain useful amounts of entropy. This entropy can be extracted through various transforms, the most powerful of which are described in section 5.2 below.

Using the Fourier transform of the data or its optimized variant, the FFT, is interesting primarily for theoretical reasons. It can be shown that this technique will discard strong correlations. If adequate data is processed and if remaining correlations decay, spectral lines that approach statistical independence and normally distributed randomness can be produced [BRILLINGER].