It's not really 5.1. What it is is that you can fold 4 channels into 2
and then decode them again to 4. There are a number of formats which
use that principle called matrixing such as the various quadrophonic
formats and Dolby Stereo, called Dolby Surround or ProLogic in the
consumer world. The principle is simple: when you have two channels L
and R and the signal is the same in both, you hear the signal as
coming from the middle, provided you are sitting at equal distance to
both speakers. That's called a "phantom center". You can also extract
the signal components which are the same and route them to a "hard
center", so you get 3 channels out of 2: L, C, R. You can add another
channel by using the same principle - putting that channel into L and
R, but this time, shifting the phase 180 degrees before putting it
into one of the two channels (typically the right). So when the
decoder "sees" that L and R are the same but 180 degrees out of phase,
it subtracts those signal components and routes them to to fourth
channel, typically a mono surround.
So you take 2 channels and compare them and when L and R are the same,
it's C; when L and R are the same but 180 degrees out of phase, it's
surround. When L and R are different, it's L and R. Of course, the
matrix doesn't "switch" between the 3 scenarios, all 4 channels can be
on at the same time as only common signal components are extracted. So
L and R can contain a complex signal waveform which contains some
identical elements (=C) and when those get subtracted, the basic
waveform remains, there are no "holes" in the signal.
So you arrive basically at a 4.0 configuration. You can derive a
subwoofer from the front channels by using a crossover and directing
LF signal components to the subwoofer, but there is no discrete
channel for it.
Some newer technologies then take the derived surround and compare it
to what is happening in the remaining L and R signals, and also what
happens right before and after the signal shifts from L and R to
surround and so derive a kind of "educated guess" stereo surround.
That works pretty well, but it's not quite the same as having 5.1
discrete channels.
That principle can be applied to most stereo sources anyway even if
they are not specifically ProLogic encoded as basic mixing and panning
principles still apply - if an engineer wants to mix for 2-channel
stereo and he wants a signal component to come from the C, he will do
exactly the same that the encoding matrix does - put it into L and R
equally. That's what the pan knob on the mixer is for. So it often
comes out OK when played back through a matrix decoder. The same
applies to the surround channel. When you have diffuse elements in the
signal, such as reverb or applause noise, a lot of it ends up in the
surround channel when matrix decoded.
It is not really known what is allegedly so special about AMSI which
is supposed to be played back through a conventional matrix decoder
such as ProLogic, and what is different about it from just playing the
stereo mix and letting the matrix decode it. My guess is that they
introduce tiny phase shifts in parts of the material in order to steer
those parts more predictably to the surround channel. So they probably
take the 2-channel source, matrix decode it, then add elements of the
left and right channel into the decoded surround, then re-encode it
into 2-channels. Or something similar to that. If it yields better
surround sound results than just playing the un-AMSI-fied stereo mix
through the surround decoder, I don't know. It appears though as if
what they do is so minute that when people compare the "reguler" 2-
channel version to the AMSI version (played back on 2 channels), they
don't hear whatever they do to the tracks to make them "more AMSI".
In any case, none of that has anything to do with the tone colors on
the recording. It's really all about channel steering.