All that I heard so far about phase spectrum - that it is very hard to process. Are there any methods to do that, and is it possible to extract information about the noise from phase spectrum?
you can look at
the short answer is that for noise, phase is random. There shouldn't be any structure to exploit, while for signal, there often is.