# Do I understand Mid/Side processing correctly?

Recently I've been reading about the mid/side processing and I'ce got this image in my mind:

Let's assume I have 3 guitars in my recording. Guitar one "1" panned 100% left, second "2" in center and third "3" in 100% right. Now when I sum them all together I can process them with Mid/Side processing so I will either deal with guitars "1" and "3" together (Because they are on sides) or with signal "2" because it's in centre. Am I right?

The second question is about the perceived loudness of it. If I have a hi hat panned 70% to the right, that means that 30% of the sound still goes to the left speaker. When I listen to this in Mid/Side processing what amount of dB will come out in either (Side and Mid) of them from my speakers? Will in side processing it be 70% from right and 0% from left, or 40% from left? And what about the Mid side?

In this case scenario, what if my sounds are panned only a bit, a tiny 1% to the side? How Mid/Side processing will affect what I hear in headphones then?

I think you’ve got a few things off. First, let’s look at what you’ve got with your three guitars.

Your left channel has 100% of 1, 50% of 2, and 0% of 3.

Your right channel has 0% of 1, 50% of 2, and 100% of 3.

Mid = L+R which means Mid will have 100% of 1, 100% of 2 and 100% of 3.

Side = L-R which means side will have 100% of 1, 0% of 2, and -100% of 3. In this case, the negative sign means that the polarity of the signal is reversed.

Using Mid/Side means that you can drop the side channel out and still have a mono mix, which is why the mid channel has all three guitars in it. What the side channel does is contain the “stereo width” in a way. If you mid-side encode, then the fader that controls the level of the side channel just widens the image as you bring it up. On the other hand, if you drop out the mid channel you lose guitar 2 entirely.

So what mid/side processing does is let you separately control the overall level and the stereo width. If you compress the side channel more than the mid channel, for example, then you limit the width of the image. If you compress the mid channel more than the sides, then you limit the overall level while still having a dynamic stereo width. It’s a bit weird to work out in your head, so I’d suggest playing around with it to really get it.

Your second question I don’t get. If you have something panned 30% left, then it’s still 30% left. If you mid/side process it then by raising and lowering the side channel you’re basically increasing or reducing how far left it is. 0 side means it’s panned center. 0 mid and 100% side means it’s panned 100% left.

One thing to understand about “mid/side processing” is that it’s not about taking away the normal left/right panning. What you do is you take the left/right and encode it mid/side (as above), then you process the mid and the side differently, then you decode it back to left/right. That last step Is pretty much always done because we don’t have mid/side speakers or mid/side ears. Sending mid to the left speaker and side to the right would around very strange.

As I mentioned above, I recommend playing around with it and using your ears to understand best how it all works.

For reference, decoding mid-side looks like this:

Left = mid + side

Right = mid - side

• Good explanation. I'd throw in a ref back to the 'origin' of mid-side, which is a physical mic setup which generates the 'math' you have already explained. - uaudio.com/blog/mid-side-mic-recording looks a good example. Jul 15, 2019 at 14:17