PPPS: Phases are the ABC of audio engineering, everything that has to do with music, has to do with time. Everything is time related, every EQ, just everything. There's no math that is able to separate channels by phases, assumed the frequencies are all recorded just by audible frequencies and not transformed to higher frequencies, by theory transposing to inaudible and back to audible frequencies could work. Practical, hm?! There already is no perfect backwards compatibility from 2 channels to one channel regarding to timing issues, e.g. the +/- 180° thingy. It's completely idiotic to add a third channel by what ever time shift to 2 channels and then to decode this nonsense. Unbelievable that this is a serious discussion :(. During the decoding there will be overlapping between the clean stereo signal and the third channel. Take a pencil and draw 3 sinuses to a paper by your imagination and than think about adding the third sinus to the two other sinus. Some kind of encoding and decoding, just by an imaginary picture in your mind. I'm quit now.