"Listening in Side"

I recently read an article by Dave Moulton, one of my favorite audio geek-gurus, in which he suggested a technique called "listening in side". He said he'd learned it from Bob Ludwig.

The idea is to use Mid/Side encoding and to solo the Side component to gain insight into how a record was mixed and produced. I gave it a go, and it really is quite revealing and educational.

If you have Ozone 4 or 5, you have the ability to do this very easily, as the equalizer has a Mid/Side mode and solo buttons for the Mid and Side components. (The solo button is the blurry red splotch on the right side of this blurry screenshot.)



Just enable the equalizer, click the "S" button and you're listening to your favorite references in Side-only. Which is to say you're hearing only what was panned to the extremes. It's like getting an inside look at the multitrack.

If you don't have Ozone, you can use Voxengo's free MSED plugin. Just insert it on the master bus and turn the "Mid Gain" control all the way down:



[EDIT: I'd originally said here that the Channel Tools plugin would not work for this, but I was wrong. Steve Corey straightened me out. You can in fact use Channel Tools for this exercise if you don't have Ozone.]

I created a SONAR project for this and loaded up 8 of my favorite reference songs, representing a wide variety of styles from different decades.

First up was "Eclipse" from Dark Side of the Moon. There is no kick or snare, but cymbals and toms were surprisingly present. This tells me that the overheads contributed more to the toms than close mics, which I'd think would be atypical today. There is also no organ in the sides, which is surprising. It means the Leslie was recorded mono and put in the center, definitely atypical. (That organ sure sounds good for being mono! So much for my long-held belief that Leslies must be stereo to sound good.)

Next, I listened to a clean, modern, semi-aggressive, typical Nashville-style country-pop song by an artist whose name escapes me at the moment (I know he wears a cowboy hat, though). It features acoustic guitars panned hard L-R, which are back in the mix when heard normally, but in Side mode they're pretty much all you hear until the drums come in. That, and some vocal reverb. Many of the references had that characteristic: wide-panned vocal reverb. 90% of the elements of this record are panned up the center, but it has a very satisfying stereo image nevertheless, which goes to show that width depends on what you choose to pan out there.

In Tom Petty's "Free Fallin'", you hear the trademark double-tracked and wide-panned 12-string guitars as expected, but the surprise is how much snare is in the sides. On closer listen, it became clear that the snare had a subtle delay on it, and the echo was panned partly to one side, making the snare sound bigger and wider. I wouldn't have noticed the technique if not for listening "in side".

"Pull me Under" by Dream Theater was interesting. There's a lot more happening in the sides than in any other reference I listened to. Lots of drums, including hats and snare are panned out (I'm surprised at how common it is to not have the snare down the middle). And of course, lots of distorted rhythm guitar, as is required by the genre. Once again, the vocal reverb is spread way out.

I won't bore you with all my observations, but rather invite you to try out this technique for yourself.

I can't wait to try this.

Bore us?  Please.  I can't speak for everyone, but I suspect I'm not the only one who sees a post from bitflipper (or Jeff Evans or Danny Danzi) and thinks "Okay kids, class is in session."

Hmm, snare not down the middle?  That's just crazy talk!

RabbitSeason


Bore us?  Please.  I can't speak for everyone, but I suspect I'm not the only one who sees a post from bitflipper (or Jeff Evans or Danny Danzi) and thinks "Okay kids, class is in session."

Hmm, snare not down the middle?  That's just crazy talk!

+1   +1    +1    +1 .............................................
 
Would you please consider boring us a little more Mr. Bit....
 
I found this topic very intresting since I know very little about it or how it works........but I do find your insights very helpful!
 
Mr. Clifford
 
:-)
 
P.S. I got the link and now am reading it.......this will take me a few readings to digest and understand.
But thank you for bringing it to OUR attention.... 

Silly question: what is considered Mid and what is considered Side, is Mid from 50% L - 50% R and the rest Side?

It's my understanding that mid is the Sum (L+R) of the stereo channels and side is the Difference (L-R). Flip the polarity of one side L or R gets you the difference. Moulton has had this stuff on his website and used to do CDs for "Home Recording Magazine" back in the late 90's early 2000's.

Been listening to more pop music "in Side". Got a chuckle listening to Kelly Clarkson's "Breakway" - they went easy on the autotune on the lead vocal, but the BGVs are Cher'd up and full of artifacts! You don't actually hear that except when soloing the Side component. 

There are a lot of phasey artifacts heard on this song. I suspect this may be because they're doing some phase-shift tricks to exaggerate width. When they're panned apart it sounds good, but when collapsed to mono - which is what we're doing here - they exhibit comb filtering as if they'd all been treated by V-Vocal! Listen to the hats, which sound particularly awful.

The topic of LCR mixing came up in another thread. One of my references turned out to be a good example of the technique: "Bad Things" by Jace Everett, better known as the theme from the "True Blood" TV show. Bass, bass drum, tambourine and vocal are center, everything else is hard left or right. Hey, now I can learn that uber-cool organ part!

"Handle Me with Care" by the Travelling Wilburys revealed interesting techniques. The song is mostly down-the-middle mono except for two components: acoustic guitars and snare drum. This, I'm discovering, is a surprisingly common technique: pan the snare to one side and add a delay panned opposite. The snare on this song sounds like it's ten feet wide!

 

I guuss I'm still learning :)

Thanks!

Thank you for posting this Flipper, very informative.

Actually, the hats sound pretty bad in the stereo mix, too. Soloing the side component is in fact collapsing to mono - the left and right extremes are literally folded into a mono signal when soloed.

I meant to post this one here ---- I found this good article / site for those looking for good commercial reference songs:
 
http://www.digido.com/honor-roll-of-dynamic-recordings.html

bitflipper


Exactly. If, for example, the left and right signals are the same but phase-shifted, then they are part of the Side component. And they'll sound nasty when combined. They may sound great in headphones, but that comb filtering will be audible when listening to speakers off-axis, or in mono. 

No.

What I am saying is that some comb filtering that maybe apparent when the side component is heard in isolation may not be apparent when you are listening in stereo or true mono (both mid and side components included).  IOW listening to the side component alone can remove elements that CAN cause a comb filter effect.  I'm not saying that's true in your example but it does occur in some examples when both components are recombined the problem isn't apparent.

If you go through enough references you'll notice quite a few examples where isolation causes problems that are not there when everything is put back together either in full stereo or summed as mono.

I'll have to find a good example, I don't have one to hand at the moment but I've certainly come across it before.

I've only been using Ubermod for a short while, mostly for its chorus and ensemble effects, and can't comment on how well it maintains mono compatibility. Of course, any chorus-type effect involves recombining delayed signals, and the resulting comb filtering is what makes it sound like it does. I should really spend some time exploring Ubermod more.

ValhallRoom, OTOH, has become an old friend, being my go-to reverb that I end up using 80% of the time. VRoom's "diffusion" parameter is not equivalent to the Sonitus Reverb's "width" setting nor to the "diffusion" settings on other reverbs. The help text says it "controls initial echo density", but TBH I'm not sure what it really does. I know of no way to narrow VRoom's stereo dispersion using only its own controls.

Yes, you're right,  a 0.25ms delay is equivalent to a 90-degree phase shift at 1KHz.

There is no kick or snare, but cymbals and toms were surprisingly present. This tells me that the overheads contributed more to the toms than close mics, which I'd think would be atypical today.

Why would it tell you that if the close mics are panned off-centre which they clearly are in this instance?

Unless somebody is keeping a secret from me it's not atypical for the overheads to be the prominent basis of the drums still.  In fact on most commercial music these days you'd be lucky to hear anything other than the overheads and maybe the room from the original recording.  The rest these days is commonly used to trigger replacement sounds and has been for years.

Another variation of this listening technique is to split the mid and side signals and 'send' them prefade to seperate buses put another instance of channel tools or whatever you are using to mid/side encode and decode the components back to the individual stereo components but on seperate busses.

You can then switch between mid, the full width stereo component in isolation as well as combining them back into the full mix really quickly.  You can even choose send some of the side component back into the mono mid bus which can be a definite aid to mono compatibility in some circumstances.  Try adding some extra air to the sides and cutting some LMF mud with a separate eq there, try different comp settings on the mid and sides and such.

It's by taking all this stuff apart and putting all back together that helps make sense of it all.  You don't need an invite via Bob Ludwig to find out about this stuff.  Hey, you'd even find out how completely cool Cakewalk's Channel Tools is and what it's capable of...  Not only that purely playing around is fun.

bitflipper


My, we're feeling a little cranky today, aren't we, JB?

Why do you say that?

I'm not as it goes.

Some of the things you were saying here just didn't make any sense in relation to the subject that you raised.  I felt some clarification was in order in case people interested in the subject of Mid/Side were left with the wrong impression.

Indeed it's a useful, and well-used staple for any mixer to understand the concepts of, not only for listening for mixing tips from old recordings but for gaining complete control over your wide and centered elements.

In fact there is a whole range of specialist plug-ins that one needn't purchase once you understand how mid/side encoding works and you are able to treat both components seperately by bussing them independently.  Especially as stereo source tracks are not uncommon these days with soft-synth patches and stuff that of course will require some management, Channel Tools is awesome for this task it covers every aspect you need whether it's encoding/decoding to/from mid/side or just being able to limit the width of these sources and to give you the ability put them exactly where you want just like a mono source.  I'd recommend an instance of it on every stereo track you have on a mix.  That fact that you were unaware of it's capability in this area rang some alarm bells for me, but there you go.

Colour me cranky if you like but I felt some clarification was in order here.  If however I have said anything inaccurate here feel free to correct me.

bitflipper


I should clarify: when you solo the sides in Ozone, the output is folded to mono. This is what I was referring to when I said that phase differences become apparent when soloing the Side component. This would not be the case if you're using Channel Tools or MSED to mute the Mid component.

I'm still not sure of what is happening here that is specific to Ozone then.

If you listen to either the side or mid component on any MS decoded signal then each component is 'mono' (or rather a single track) whatever tool you use.

I still don't think you've understood that phase issues maybe inherent when listening to either the Mid or Side in isolation merely because the side has lost it's relationship with the mid component and therefore some phase issues are likely to be apparent because you have singled them out and may not (or indeed may) exist when the signals are recombined.

If you are in Mid/Side mode (which you are as soon as you simply click the input mode to M/S on Channel Tools without needing to touch anything else) the left channel is usually the mid and the right channel the side component.  So if you pan a send hard right to one bus and pan another one hard left then you have seperated out the two components.  Convert them back to stereo on the seperate buses with another instance of your encoding tool and you can hear (and treat) either in isolation. 
 
Every and any Mid/Side decoding/encoding matrix works the same way it's just that Channel Tools comes free in the box already and performs every task that you'd want for dealing with stereo whether L/R or M/S or switching between the two.  If anything else is happening then it ain't truly Mid/Side as defined by Jimbo's equation earlier.  In the digital domain tt's a simple logical switch on the signal to find what is Mid and what is Side and to convert back, it's always one thing or the other there is no interpretation or complexity involved.