To me, it's about giving instruments the room to breathe. One of the major problems with 2 channel recordings is frequency masking. If two instruments produce sounds at similar frequencies, the louder one wins. This can make instruments sound thin. Decent engineers will compensate for this with "subtractive EQ". In other words, notch frequencies out of a louder instrument, say guitar, that coincide with a different one, like vocals. A little goes a long way. The other way is, of course, to pan the instruments to different speakers. With 5 discrete channels, you can avoid that problem even more. The center isn't a psychoacoustic effect (where equal levels from left and right make it appear that the sound is localized in the center of the listening field). It IS a center channel. I know the guy who mixed the movie Blade (and Dead Presidents, Barb Wire, etc.), and he said that music producers were always impressed by how good their tracks sounded in his movies. His answer was that it was all about localization: each instrument had room to breathe.
The folly is, of course, the ping-pongy thing that quad sound did in the 70's.
