School me a bit on MID-SIDE stuff. Is it still relevant today?

[rasputin1963]

Can you school me a little bit on the MID-SIDE technique of mixing? After all these years, I don't really understand it.

 

As I understand it, it is a product of earliest stereophony and pseudo-sterephony, when three tracks were all you had, and a track could either be panned L, C or R (Like "Eleanor Rigby" by The Beatles or "California Dreamin'" by The Mamas & The Papas).

 

I think it also has something to do with radio airplay? (Or radio airplay as it used to be)?

 

I do understand that Far Left, Dead Center and Far Right are the important and powerful/interesting places for a sound in a stereo mix to be.

 

As I understand it, MID-SIDE is not strictly necessary to understand in the modern digital domain... Yet it still frequently comes up. BRAINWORX SATURATOR, for example, is a modern VST plugin that adds analog warmth to a stereo track, and its dials refer to MID and SIDE.

 

Huhn? What do I need to know here? What am I not understanding?

 

ras (Don't Hate 'Cause I'm Dumb)

[CTStump]

Here it is in a nutshell. You can process effects to the center or to the stereo fields. One example would be to apply all reverb to the stereo fields to clean the center up.

 

here is an excellent short description from Izotope:http://www.izotope.com/support/help/ozone/pages/mid_side_processing.htm

[Jorge7]

Not sure what your recording application is, but Mid/Side recording is one of Blumlein's original stereo recording innovations, and is also very relevant today. If you are interested in ways to modify the soundstage on your microphone recordings of live performances, then M/S techniques can be very useful. There is an article on M/S processing using Studio One in this month's Sound on Sound but the full text of the past 6 months of articles are only accessible if you subscribe. A longer more general article that puts M/S in context of stereo recording techniques was published in SOS 2 years ago and is publicly accessible:

http://www.soundonsound.com/sos/nov10/articles/stereoprocessing.htm

Although M/S matrix decoding can be done by many widely available DAW software packages and some analog recording consoles, good figure 8 mics and good cardioid condenser mics with smooth off axis pickup patterns are expensive, so it would be good to learn as much as you can about the techniques before you invest the money.

[Anderton]

I can't school you, but Phil O'Keefe's articles in the HC Articles Library will:

 

Basics of Mid-Side Recording

Blumlein Pair Stereo Miking for Better Ambience and Imaging

 

More useful stuff:

 

Stereo Placement

XY Stereo Miking Made Easy

[Philter]

I've tried all the common stereo techniques and I prefer a Blumlein pair in most situations.

 

M/S is commonly described as sounding natural, but it usually sounds weird to me.

 

CTStump's comment is interesting; I hadn't considered that angle.

[Mark L]

I haven't the foggiest idea what you're on about, son

[rasputin1963]

 

I haven't the foggiest idea what you're on about, son

 

 

then you 'n' me laddie had better swot up on the articles mentioned here, hadn't we

[philbo]

While most recommend using a figure-8 mic to pick up the Side signal, when doing early experimenting with it I found you could use 2 cardoid (not hypercardoid!) mics facing 180 deg away from each other. You set up the pair at 90 deg to the sound source, then invert the right one and mix it with the left to get the Side output. You can also angle each of them in toward the source 10 or 15 degrees, which can have a dramatic effect on the stereo imaging.

Very interesting to experiment with!

 

(The Mid mic, of course, would be facing directly at the source).

It's best to arrange them with all 3 capsules aligned on the same vertical line to avoid phase issues.

 

It's much less hassle to use a figure-8, but this lets you try it out before shelling out $ for a fig-8. It'd be preferable if both mics were the same model, but OTOH rules are for other people...

 

If you decide you want to commit to buying a fig-8, here is a very affordable and versatile multi-pattern LDC mic to use for the fig-8: Studio Projects B3

It runs about $150 street price.

I really love mine!

[philbo]

 

I've tried all the common stereo techniques and I prefer a Blumlein pair in most situations. M/S is commonly described as sounding natural, but it usually sounds weird to me. CTStump's comment is interesting; I hadn't considered that angle.

 

 

You do know you need to run it through a MS->Stereo decoder before listening to it, right? If you don't, I guarantee it would sound weird!

[MikeRivers]

It's all done with arithmetic. M-S is the exact equivalent of X-Y with the mid and side mics at equal levels when added andsubtracted. With anything other than a 1:1 ratio you can get anything from mono to big hole in the middle. Two good things to know are that at any ratio, summing the de-matrixed L-R gives you real mono, and that unlike X-Y, the center is always on axix, where most directional mics sound best.

[Philter]

While most recommend using a figure-8 mic to pick up the Side signal, when doing early experimenting with it I found you could use 2 cardoid (not hypercardoid!) mics facing 180 deg away from each other. You set up the pair at 90 deg to the sound source, then invert the right one and mix it with the left to get the Side output. You can also angle each of them in toward the source 10 or 15 degrees, which can have a dramatic effect on the stereo imaging. Very interesting to experiment with! (The Mid mic, of course, would be facing directly at the source). It's best to arrange them with all 3 capsules aligned on the same vertical line to avoid phase issues. It's much less hassle to use a figure-8, but this lets you try it out before shelling out $ for a fig-8. It'd be preferable if both mics were the same model, but OTOH rules are for other people... If you decide you want to commit to buying a fig-8, here is a very affordable and versatile multi-pattern LDC mic to use for the fig-8: Studio Projects B3 It runs about $150 street price. I really love mine!

 

I own a bunch of Studio Projects mics including a B3 and I agree- they are a great value.

 

For a figure 8 pair right now I am really loving my pair of Cascade Fathead ribbons. It's amazing how many different things I can throw them in front of and get a wonderful, natural sound without any effort. You can get a matched pair for $350... another really incredible value.

[Philter]

You do know you need to run it through a MS->Stereo decoder before listening to it, right? If you don't, I guarantee it would sound weird!

 

My decoder is my DAW. It's easy to configure the channels in protools.

 

I know you know this but for anyone else, here's how you set up the channels:

 

 

I've never gotten it to sound as good to me as a figure 8 pair, capsules one above the other, offset 90 degrees does. The stereo image never sounds wide enough to me with M/S, even though you're supposed to be able to adjust that by changing the balance between the mid and the side pair.

[philbo]

My decoder is my DAW. It's easy to configure the channels in protools. I know you know this but for anyone else, here's how you set up the channels: I've never gotten it to sound as good to me as a figure 8 pair, capsules one above the other, offset 90 degrees does. The stereo image never sounds wide enough to me with M/S, even though you're supposed to be able to adjust that by changing the balance between the mid and the side pair.

 

Well, each person has their tastes and preferences, I guess.

 

Interesting thing: I couldn't see your text below the image in your orig. post.

[JeffLearman]

First, there are two closely related but really very different issues. Both are totally relevant to today, and your belief about it being right/left/center channel is off the track.

 

1) mid-side miking. This is using one mic pointed at the center of the soundstage, and the other at right angles, pointing to one side. You generally want the two capsules as close as possible to avoid phase cancellation. In *theory* it should deliver (after decodeing) the same results as XY miking with two mics pointed at 45 degree angles. In practice, it's close but not the same. However, in both cases, the direction information is entirely from amplitude (loudness, i.e. same as panning) and not from phasing. This is the opposite of spaced-pair technique (like ORTF) , where the direction is mostly encoded as a delay: the stuff coming from the right side is delayed in the left mic, but pretty much just as loud. More on that below.

 

2) mid-side encoding. This is just a way to encode a stereo signal. The usual way is LR (left/right), where left is in one channel and right is in another. However, there are advantages to using a different approach, where one channel ("M") is L + R and the other ("S") is L - R. This is what FM radio uses. The simple advantage for radio is that it's also broadcast in a way where if your signal is weak, you can get M but lose S, so you get a usable mono stream rather than just one side. An advantage for mixing is that you can easily control the stereo depth: just turn the S channel up or down. However, to use it you need to decode it. Philter points out one way. Another is to use a simple mid-side encoder plugin, like mda-vst's "Image" VST (a great little free plugin).

 

The plugin is handy because it'll take any input type (LR or MS) and output any type (LR or MS), but in the middle it's MS and you get the nice balance and width controls.

 

Take a look at Philter's drawing. The balance is just the pan pot on the 1st channel: that's where the middle goes. The width is the faders on the two side channels. Pull those back together, and you have a nice mono signal, wherever you've panned it (usually, the middle, but not necessarily -- say, if you recorded one acoustic guitar this way but don't want it dead center). Push them in, and the stereo effect fades in smooothly. Adjust to taste, then back off a bit (my usual advice for any "effect").

 

A common misconception about MS is that you avoid phasing issues. Not true. However, you *do* avoid "phasing issues when summing to mono". If the program gets summed to mono for any reason, the side channel cancels itself out completely. So, any phase issues between the M and S disappear, voila! However, if there are phase issues (if you had the two mics too far apart), anyone too close to one side or the other will hear the phase issues.

 

Mid-Side is a great technique to use with certain effects, especially back in the days when extra FX units were costly. Want a nice stereo reverb? Just return the reverb as a side channel (add to left, invert on right), and bingo! Surprisingly nice, and disappears when summed to mono (which could be good or bad depending on your intention with the reverb). Works even better with chorus or pitch-shift doubling, especially when used subtly just for stereo imaging rather than for tonality change.

 

Above, I mentioned use of delay rather than amplitude for direction (panning). It really works, and anyone who's serious about mixing should try this little experiment some time, to really let the concept sink in.

 

Using headphones, take a mono signal and feed it into two different delay devices (plugins, whatever). Start out with zero (or minimum, but most importantly, equal) delay on both sides. It should sound perfectly mono, dead center. Now start increasing the delay in tiny amounts on one side. As you do, it'll sound exactly like panning the signal to the opposite side. At the delay that's equivalent to the delay between your ears for a signal coming from one side (say, .7 msec), it sounds like it's entirely in one ear in the headphones. While it's doing that, slowly lift the headphone cup away from the loud side -- you'll hear it come back into your other ear. It's really non-intuitive and quite striking.

 

That's beside the point of MS, though. For one-point MS miking, direction is entirely from amplitude, not phasing. That avoids lots of problems.

[JeffLearman]

Oh -- another nifty little MS technique is to record acoustic guitar with a good mic and also the guitar's pickup. Use the mic for M, and a bit of the pickup for S. Sounds surprisingly good! Might not be pro-quality, but beats the crap out of recording only the pickup or mic and then faking stereo. When doing this, put a delay on the S channel to align it with the M, since the mic is farther from the sound than the pickup. The easiest way in most DAWs is to just drag the S track to the right to line it up with the M, visually.

[rasputin1963]

Now here's the channel configuration I've BEEN using...

 

 

 

(Seriously, guys--- Thank you so much for this excellent, valuable information. Much appreciated.).