MS Stereo Recording walk-through

ScoobySounds

ScoobySounds

Well-known member
I know the theory, and have practiced it quite a bit.. in the past I've experimented with MS Stereo atmos tracks which the dub has always been pleased with etc etc.. but i've got a documentary coming up which is about sound, so no doubt, a lot of people who'll be watching it on TV will be knowledagble about sound. so i want to do a good job.

I'm doing a tour next week with this japanese musician who puts on concerts in acoustically insteresting locations, (i've heard one of them has a 15s prolonged reverb!, i'm thinking that can't be right ;)) so i want to be able to capture some of the essence of what this guy and his music is all about.. which says MS stereo to me.

so i want to know some considerations you would take, walkthrough style, when recording Mid-Side.

In my experiments, I've taken the level of the Mid mic and then just bring up the Side mic to fill up the space until it sounds 'right'. Is there a right way and a wrong way to do this?

The right and left channel should be about equal when matrixed, shouldn't they?

Is there any advantage in recording the Mid and Side mics UNmatrixed, (and just monitor the MS Stereo), and leave it to post to do the matrixing?


My figure of 8 mic is a Rode NT2A, a studio mic. (i'm not going to be buying another one, so go easy on me) and my mid could be a 416 or my Schoeps CCM41. What would you choose out of these two for your mid mic and why?

:)
 
use the ccm as the mid -much better response and much better if they will deside go mono track and eliminate the 8 .

about stereo with- you can mix it right away on the spot or leave it ms .
ask the editing what they prefer .
the most important is to find the right spot in the room /field to get the right image , the stereo image and the levels could be changed /fixed in the editing
 
MS recording benefits from having the Mid and Side elements matched as closely as possible. For that reason I prefer to use an MKH30 with an MKH416. You may want to reconsider using a shotgun mic for your Mid element, since a cardiod will yield a more even stereo image (especially since you're recording music). With the `416 as your mid, you're best staying at least four feet from the sound you're recording.

The two mics (assuming they have roughly the same properties) should be recorded at the same level (ie. exactly equal ammounts of gain), but they should not neccessarily be reading at the same level on a VU meter. If you're riding the levels untill both channels are peaking in the same range on the VU meters, you're going to potentially have a very lopsided stereo image. If the loudest part of the sound is directly ahead, then your Mid channel will be louder than your side channel.

If you're boosting the side channel overmuch, you're stereo image will appear far too wide, which can have unpredictable results when monitored using older home theatres that use ProLogic matrixing. This is the danger of M-S recording for broadcast or home video: if the signal's too wide (ie. out of phase) parts of it can start wanderring onto the rear channel.

For best results, use a mixer that allows you to gang two of the faders and that will decode the M-S for your headphones but still pass it matrixed to the recorder or camera. My PSC M4 mixer will do this, and I believe the SD422 will as well (check the specs). Ganged limiters are also very handy, since otherwise when the limiter kicks in on one channel, you'll be effectively widenning or narrowing the stereo field for the duration of the release time.

It's a good idea to record your matrixed M-S to the tape rather than decoding it first, that way you have greater flexibility in post. Just be sure to explain to the picture and sound editors what's up with the tracks.
 
so the safest thing to do is leave all matrixing to post.
how about this..

for recording the channels separately, check gains are equal, gang ch1+2 into stereo, not ms. (hence the limiters will be ganged also) and monitor the return feed in ms to check what it sounds like.

I can matrix from either the 442 or the 744, but prob won't matrix anything since i will want to leave the stereo mix to the dubbing guy, so he can do it to broadcast stardards etc.

as oleg suggests, i will be using a hyper for the mid I know what your saying about a cardioid.. however these are the tools at my disposal.

and the sound report should tell all whats happening on each channel.

[ponder]
I could use all four channels on the 744, 1+2 for my matrixed effort, and 3+4 for unmatrixed. but then 3+4 would be from direct outs, so woudn't have ganged limiter advantages. Hmm.
[/ponder]
 
ScoobySounds said:
so the safest thing to do is leave all matrixing to post.
how about this..
[ponder]
I could use all four channels on the 744, 1+2 for my matrixed effort, and 3+4 for unmatrixed. but then 3+4 would be from direct outs, so woudn't have ganged limiter advantages. Hmm.
[/ponder]
Click to expand...
waist of spacer and effort, the stereo image could be converted back to ms , adjusted and go back to stereo .
 
triangle, i think.

na, I don't know yet.. i'll let you know when they tell me. its just him though, not a philharmonic or anything.

I would imagine he's doing it so the best listening position will be WELL over 4 feet from him.
 
that depends on your camera angle if they like to see the all **** in front of him :-)

triangle - pray for a bicycle bell .:-)
 
Let me start by saying that I do a lot of guerilla-style "audience taping " (see here), and I'm very experienced in certain stereo techniques, but not so much MS specifically (but I do using everything else from split AB omnis to DIN, DINA, NOS, ORTF, and XY cards, plus matrixes with the soundboard and ambient stereo pairs). That said, you really should be using a card, subcard, or even omni if you can (definately not a shotgun, and if possible not even a hyper IMHO unless you are way in the back maybe). Also, I would never (ever) record the matrixed mix personally, that defeats one of the most awesome features of recording MS stereo, which is the ability to have some control over the stereo image in post. No other stereo technique gives you that type of flexibility, so it seems like bad form to me to remove that flexibility.

Finally, regarding levels, my take on that is run each channel as hot as you can without clipping and worry about everything else (mixing) in post. Who really cares what the differences are between channels if you are going to mix it in post anyway? The most important thing is to get the sound way above the noise floor. Also, regarding monitoring, that's always nice to monitor the matrix, so if you can, great, but if you can't, you can still make a great recoding. Use your ears to find the sweet spot, and just make sure you are getting good solid and clean audio into both channels during the recording. And regarding limiting, yuk! I wouldn't do it during aquisition because after that it's done and there's no going back, and if you decide to use more compression in post for the edit, then you're layering compression on top of compression. I'd try to get the cleanest "pure" recording with the full dynamic range during acquisition, then deal with compression in post if you want to (although this means you need some good experience with your gear and the music to know where to set levels to begin with). Anyway, those are my two cents at least.
 
your post makes a lot of sense to me mate.. simply put and I agree with you.

especially about the use of limiters. They just sound, well.. like limiters. If I have the allowance, i'm gong to hear the peice before recording to get them just right, as a sound check - so i don't have to ruin a perfectly good recording with limiters kicking in.
 
Tainted said:
so it seems like bad form to me to remove that flexibility.
Click to expand...

Correct me if Im wrong, but you can theoretically run the signal through encoders and decoders a billion times and it will be the same signal?
so if you recorded it encoded to MS, you could easily decode it (perhaps in realtime) during post to adjust the relative levels of M & S.


@ SMuir

I found your post interesting because im practising/learning MS rcording at the moment and I always try to get the best signal to noise ratio for both channels, which usually means increasing the gain of the side channel.
I do find that my stereo image can sometimes be VERY wide, I thought this could be partly because I use a shotgun mic as my Mid component.
Next time I'm recording I will follow your advice to see if it makes a difference.
 
Haydn said:
Correct me if Im wrong, but you can theoretically run the signal through encoders and decoders a billion times and it will be the same signal?
so if you recorded it encoded to MS, you could easily decode it (perhaps in realtime) during post to adjust the relative levels of M & S.


@ SMuir

I found your post interesting because im practising/learning MS rcording at the moment and I always try to get the best signal to noise ratio for both channels, which usually means increasing the gain of the side channel.
I do find that my stereo image can sometimes be VERY wide, I thought this could be partly because I use a shotgun mic as my Mid component.
Next time I'm recording I will follow your advice to see if it makes a difference.
Click to expand...

You've got some stuff not quite right. Once you've matrixed the channels, there's no going back unless you have the originals still. All the matrix does is add and subtract one channel from the other. But once you do that addition/subtraction there's no going back. It's a destructive process just like adding compression or anything like that. Now as long as you have the originals, you can of course go back and try again, but the idea that you can encode/decode at will is not true.

Regarding getting the highest levels without clipping, I'm talking about when you are NOT matrixing it on-the-fly. I only recommend that if you are recording both mics seperately so you can completely control the matrix in post. If you are mixing/matrixing the mics into the MS on-the-fly, then you HAVE to get the matrix right as you record because you are literally recording the stereo image as you hear/monitor it, and you CAN'T go back later and change it (unless you have the seperate recordings of each mic still).
 
a+b left a-b right
dematrix
2a left - 2b right
there is no problem to go from ms to stereo or versa
in analog world that was cost you generation ,with high sample digital recording the lost is minor .
you can dematrix and change the levels and so the stereo image spread
than matrix it again
 
now i'm just thinking about this from a mathematical point of view, i've never done it.

If you have an MS signal and you add the left to the right channel, you will cancel out the side channel and be left with the mid.

now if you take the the left channel away from the right channel, you obtain the Difference between them, hence the side.

thats right isn't it? and presumably you could do that all day long, non-destructively.
 
This is how i'm going to use it.. not as good as matched mics, or single point system, but should still give good results in, like a hall. have to be careful if i'm outside though.


IMG_0537-1.jpg
 
ExileOnTheMainStreet said:
a+b left a-b right
dematrix
2a left - 2b right
there is no problem to go from ms to stereo or versa
in analog world that was cost you generation ,with high sample digital recording the lost is minor .
you can dematrix and change the levels and so the stereo image spread
than matrix it again
Click to expand...
Looks like I stand corrected, but I'm still not sure I'd be into the back and forth unless I was at 24-bit. I wouldn't be surprised if quantization noise might start becoming audible at 16-bit after all that back and forth and back again. But point taken, thanks for clarifying that. I still think you should record it not matrixed and then mix it in post, and of course, being able to monitor the matrixed MS real-time would be nice, but I'm not sure it's that crucial. But like I said before, I'm a cards XY, DINA, DIN, NOS, ORTF, and Split Omnis kind of guy at heart. :)
 
Hold on a sec. The reversing of the MS discussed above assumes a full width MS doesn't it? But sometimes you don't run full width like that, in fact a lot of the time you don't right? If you were monitoring, you'd adjust the MS until the stereo spread sounded "right," so the image was what you wanted. If you were to try to reverse that, you'd have to nail exactly how you mixed it in reverse wouldn't you? And would that even be possible if you didn't go full A+B / A-B in the first place, which chances are you wouldn't would you? Am I missing something here? (Probably.)
 
NO PROBLEM WITH 16 /24 / DID IT ALLOT OF TIMES AND CORRECTED OR PULLED OUT THE M ONLY TO MONO MIX .
not a problem .
i dont like to do stero mix in afirst place , but sometimes you know that the edotor woldnt be able to correct or dont even know what is ms , so you do stereo .
 
tainted, i'm sure the stereo spread is simply determined by how much side is mixed, ie. the amplitude. so this isn't affected by the summation / subtraction process.
 
You must log in or register to reply here.
Back
Top