Can someone explain m/s processing to me?

[Zak Kazanski]

Hey everyone,

Have been reading alot on this site about M/S processing, but I don't really get how it works exactly, and what I would use it for? Anything that would clear it up would be greatly appreciated.

[always in record]

processing the sides different then the middle


edit- thought you were talking about m/s recording

[Ben F]

It's pretty much only used in mastering to try and fix a bad mix.

[complex]

well M/S is Mid/Side which means Mono only signal and Stereo only signal, i use it on some mixes if i want stuff to poke out more in stereo or mono

[matucha]

MS is a very powerful thing. The 1st thing to understand MS is to hear M and S separately on a lot of different things. M is obvious Mono. S is everything else.

Say you have mono track. If it's centered it is just in M. When you start to pan, it starts to appear in S raising in volume and when you have panned fully to the side your M and S have the same loudness (well that depends on what your pan law is, but let's not go there).

When you copy your mono track to 2nd track and turn the polarity on that one. Then you get just S and no M.

Stereo recordings have different amounts of M and S with different spectral curves. Very diffuse recordings have M and S of the same spectra and level. Things miced closer to the source (in less reverby conditions) have less low freqs in the S signal because they get more in phase than the high ones. But if there is delay between L and R your M and S will get more similar.


Knowing what MS can do and where the dangers lie help quite a bit with mixing and mastering.

With electronic/pop music, it helps you maximize the impact because you can do much wilder things than with acoustic music. You can add out of phase samples to your perfectly sounding mono ones to get extra width without loosing mono compatibility. Mixes with little S signal can't compete with the ones that had addresed S in the process. Of course sometimes you want almost mono for some artistic reasons...
In acoustic music, you can do a lot of harm with MS, esp when it's well recorded. But there are stereo recordings that sound harsh and dull at the same time and MS EQ can be the solution, you get mud out on M and tame some himid/highs on S and suddenly it works. You have to be carefull though, this will affect the localization.


When EQing mixes /let's call it mastering... You can focus your eq on the element you need and leave the rest as untouched as possible. For example (lead) vocals are mostly just in M. L R panned guitars in S. Body of snare and kick in M.
There are mixes with almost everything in the M and just some reverb in the S. So you can pull out some mud off the reverb if needed... etc.


So this is what I've found about MS. I'm still learning and won't stop until I die .

[worm]

[side[mid]side]

[Ksp]

Originally Posted by Ben F

It's pretty much only used in mastering to try and fix a bad mix.

i use it all the time to widen parts and also to create a more holophonic mix.

I have not a clue what technically happens accept that it allows you to first seperate the mono signal ? or centre signal from the side or stereo parts of a sound / track or mix and then you can treat just the stereo parts without effecting the mono signal or balance them up with the right software.

I use it creatively and just to add width and also some psychoacoustic edge + at a track level for creating space between parts in a mix.

[StringBean]

Originally Posted by matucha

Say you have mono track. If it's centered it is just in M. When you start to pan, it starts to appear in S raising in volume and when you have panned fully to the side your M and S have the same loudness (well that depends on what your pan law is, but let's not go there).

Does "S" in this scenario refer to stereo, or side? Conceptually I understand that panning a mono track takes it out of "mid" or "mono" (or what I perceive as phantom center), but if I pan a mono track fully left it doesn't sound to me like it is the same loudness on the left speaker than it is in the middle. I hear significantly more loudness coming from the far left. Is it due to the pan law I'm using cause I've never had it work that way?

[RaboonTheBaboon]

Hey,

The Waves S1 manual has a pretty decent explanation on the processing side of things:

http://www.waves.com/manuals/plugins/S1.pdf

Page 11 "What is M/S? How is M/S used?" etc.

It's more than just L C R.

[TRANQUILO]

It was used for TV alot in the era of TV boom and hifi when TV's where supplied with only one speaker. A more expensive TV would be in stereo. Since the consumer television market was saturated with mono and stereo TV's mide/side mic technique was implemented...so that no matter if you were watching a TV program in stereo or mono it would sound the same to an extent...remember the phrase during intro's to TV programs;"in stereo where available" meaning it would be in stereo if your tv was stereo otherwise its just mono. The mid/side technique allows the audio prgram material to sound consistant by either mono or stereo systems. It is still used in video and TV alot for this reason. As far as studio use not so sure, I once used this technique because I was limited on track count on my 8 track tape recorder. I have a 16 track now SO I dont do that as much.

[asynchro_nous]

M-S processing

Here's a rather old thread but one with some excellent information.

[Zak Kazanski]

Thank you all for the very prompt & informative info.

I have been experimenting with the S1 from waves a lot and now I have a much better understanding of how it works conceptually. I have been using it for a while now on parts of my mixes I want to have more "Width." I am now wondering how you would Eq (for example) Just the side signal as opposed to the Middle, and how why I would do something like that? Do you need a special plug-in or piece of hardware, or could I do it with the S1? I know there alot more experienced engineers on here and if one of you seasoned engineers is willing to share more in depth how you do it and why you do it, It would be greatly appreciated.

[Ben F]

Originally Posted by Ksp

i use it all the time to widen parts and also to create a more holophonic mix.

I have not a clue what technically happens accept that it allows you to first seperate the mono signal ? or centre signal from the side or stereo parts of a sound / track or mix and then you can treat just the stereo parts without effecting the mono signal or balance them up with the right software.

I use it creatively and just to add width and also some psychoacoustic edge + at a track level for creating space between parts in a mix.

Lets call it sum and difference processing, which was generally the term used for cutting vinyl.

You can't really process one without affecting the other when encoded back into stereo. You have to watch how messing with the difference signal affects the phase coherence in the low end, and listen to how it affects panning on instruments in the mix. Used with care it can be powerful, but on a decent mix stereo is preferable. Using a linear phase EQ is preferable for MS.

[3phase]

i guess its a german thing, at least there are traces of this technology back to the 1930´s here..but.. i dont know who invented it..the handwriting in the concept fits the german audio stuff.. but whatever.. dont worry.. ..

its a brillant concept but better learn to achive what you need to achive the conventional street ways.. and leave it to the mastering to change you stereo width,,

only practical aplikation in the electronic music production domain is when you get hold of a nice vintage mono hardware and you want to use it on a stereo bus.. something like a cool mono compressor or eq.. however a difficult task to adjust the side bands than digitally.. you need to compensate for the AD/DA delay and so on..

what it is? in studio technology we use symetric lines and all in and outputs are x-former isolated floating balanced signals..(everything else is not studio technology in the classical sense really ) therfore we easily can reverse the polarity of a signals and combine them thru transformators.. .. that was the base.. and the idea emerged ..instead of a L and R signal we have a mono and a side band signal..

2 channels each way.. but the the M/S technic allows to to adjust the center and the width seperatly while the the standard L/R technology only allows to adjust L and R seperatly.. so only for drastical panning fx..while M/S allows to adjust the stereo fiels itself..
As louder the side bands get the wider the stereo base.. as louder the mono/mid band gets as more everything comes out of the midlle.. boostin the lows in the mid and the highs in the side band gives a fat punch in the middle and a very wide transparent width and stereo..

but thats a mastering fx.. except rare circumstances its better to design your stereo file without the aid of such tricks.

M/S was initially mainly a microphon technology that allowed the engineer to adjust the stereo width of a recording remotly just by adjusting the levels between m and S band.. a genius idea..actually..

the mono band is L+R channel..the mono sum..and the side chanell is L-R ..

so the left chanel gets mixed with a phase reversed r channel.. all components that are in booth channles cancel out..only the difference betwen the left and the right channel remains..therfore the S band is usually much lower in level than the M band..

The genial thing about it.. a M/S encoder and decoder in hardware is just the same circuit.. so in a hardware domain you need 2 preamps with 2 output transformers outputs/windings each and you use them to create the MS signal to be recorded on tape..and you playback the recording just thru the same preamps to decode the signal to a standard L/R signal again..

great idea..and easily to implement in software..therefor many plugs theese days offer the option because it dont costs a penny. Opposite to the hardware world where especially the old way to do it electric with transfomer induction is really expensive. . as usual.. And you get what you pay for.

the nerds havent got what makes the mastering magic in hardware..and therfore the digital simulation is rather a joke.. can be usefull when a mix is ****ed and you need to enhance or reduce the stereo in the mastering..but no real magic as in a hardware real 20K.-$ Fairchild will happen.. so

dont be fooled.. just by using M/S you are not achiving was the mastering gurus achive with the aid of dedicated vintage hardware.

in the old hardware the M/S decoding is done within transformators..so you have a dynamic crosstalk between the M and S band in the -30 to -40db db range.. in discrete circuits thats around -100db and in the digital domain no crosstalk at all, especially nothing dynamical. ( => natural )

In the old ways with the transformators you can give the mid and side band very different processings that give and extrawide field and the dynamic crosstalk glues that all together again and makes it a natural sounding soundstage again.. and thats indeed a sexy sound fx..

doing the same discrete or even digital gives a much more arteficial and unnatural sounding result.. except you do it very carefully right...

In no way its the super mastering trick that will make everything better... Without transformators doing it its just a special fx.. use with caution.. danger

with the old gear it had a bit that better box fx.. but we talking about 2 total different technologys here and cant compare a M/S mastering with vintage analoge with one in the digital domain..

In the digital domain ist rather a surgical tool that can damage more than it brings..

so:
experiment with it..but as allways less is more and dont believe the hype..
You dont need to use mastering tricks exept you do the mastering.

Within syth sound design it can be intersting thou.. just like 5000 other things ;-P

[Zak Kazanski]

Amazing reply 3phase. You gave me far more than I expected. Its people like you who inspire people like me, so Thank you.

Although, I was following you up until you got to "Dynamic Crosstalk," could you elaborate a little more on what this means?

[Beany]

'Cross-talk' is where some of one signal gets through to the other. In this case a faint amount of the mid was leaking into the side channel and vice-versa, ultimately helping to glue it back more naturally.

Great reply 3phase! You are giving me ideas - Slates VCC incorporates cross-talk in group mode, I'm wondering if I can use it with DDMFs Metaplugin & Brainworxs BX_Solo to similar effect...

[bluegreengold]

It originated as a way of making binaural recordings using either a omni or cardiod microphone for the center, and a figure 8 patterned microphone (typically a ribbon mic) for the sides. You take the figure 8 response, which will naturally have a +/- polarity for each lobe, and send the signal to two channels, and then flip the polarity of one of them. Then you mix the 2 side channels on a bus with the mid and you get a stereo signal. The width of the stereo signal can be manipulated from very narrow (mono) to supernormal wide.

This has advantages over other stereo micing techniques that use spaced pairs. Because you set up the M/S microphones to be as coincident as possible, you minimize phase interference, making it more compatible with mono speaker systems, and also it sounds 'natural.'

The process of doubling the sides, and flipping the polarity of one channel is called 'm/s decoding. It's also possible to do the process in reverse in order to separate out the mid and the side components so that you can an use eq or effects only on the stereo difference portion of the signal.