a spectrum anayzer isn't what you're trying to make then, it's closer to what i was describing, except you wouldn't be using the difference signals beyond the 2L and 2R signals, so you would be using the 2L, 2L+R, 2L+2R, 2R+L and 2L signals. this could actually be done with a resistor ladder, and some op amps that you could trim the gain on to get all of the VU meters reading the same when you have a center channel (L+R) signal (i.e. both channels driven equally and in phase). later tonight i'll draw a schematic along those lines and you can try it.