selig wrote:Wait a sec - the mid signal is not JUST the mid content (if indeed that's what you meant to say) - it's the ENTIRE stereo spectrum reduced to mono, sides AND mids (just to be clear). The Mid and Side signal BOTH contain the sides, FWIW, as it's only the side signal that has anything "removed".
ScuzzyEye wrote:
OK, yeah. What I meant to say, is if you remove the mid signal from a M/S encoded pair, and then decode it back to stereo, you'll be left with only the stereo content, the middle will be gone. Which is what my original solution does. Similarly, if you remove the side information, and then decode, you will be left with only middle. The decode part is important.
Not exactly, as the signal won't decode properly without both Mid AND Side information. What you end up with in that situation (dropping the mid signal) is the mono "sides" signal on one side, and the inverted mono "sides" signal on the other (which will totally disappear in mono). An instrument panned hard left will not appear hard left in this scenario because there's no Mid signal to recreate the original panning. Any change to either the side or mid signal will in some way destroy the original stereo image… Similarly, if you drop the sides you are NOT left with just the middle - you are left with a mono version of the original STEREO signal (sides included).
What I wanted to make clear is that the mid signal is simply a mono version of the entire stereo signal, not "just the mid panned signals". This part often gets overlooked by those thinking this process actually separates the mid signals from the sides. Possibly this effect's name isn't as accurate as intended, mainly because the name comes from the microphone technique where the "mid" signal does indeed represent mostly the middle panned sources in the stereo field. And btw, the "side" signal is also potentially misleading since it is mono - meaning it's
not actually the stereo signal with the middle missing.
(where's the "condescending smiley" when you need one)