Topic: M-S setup question

[DSatz]

jbanyai, in principle there's nothing wrong with using two microphones of different brands or types--even a ribbon figure-8 for S and a condenser for M can work together well. But the stereo imaging can get weird if the two microphones have very different response characteristics especially in the midrange and upper midrange. It's not a great idea to pair a mellow S microphone with a brash M microphone, for example, because then the apparent position of any given sound source (instruments, for example) will vary at different frequencies; that can be a bit bizarre-sounding.

Interestingly (to a geek like me, anyway), this implies that sometimes a stereo microphone isn't the best answer, since most of them--the conventional dual-membrane kind with electrically switched directional patterns--change their high-frequency response when you change their directional pattern. (This is usually quite obvious in the spec-sheet curves, assuming that curves are given for each available pattern.) The best results might be when you choose two specific capsules whose frequency response matches well despite their having (most likely) different directional patterns.

As an example, the Beyer M 160 ribbon supercardioid is sometimes paired for M/S with the Beyer M 130 figure-8; both have very flat, smooth response and musically the combination works quite well (although their low sensitivity can cause problems). "Condenserwise" if you look at the response curves for Schoeps ("Colette" and CCM series), Neumann (KM 100 series) or Sennheiser (MKH series), nearly all the small-diaphragm capsules or microphones of each manufacturer have similar high-frequency response characteristics to the other capsules or microphones in the same series. All other things being equal, their similarity in frequency response is a good thing for M/S.

Large microphones, however, generally don't maintain their directional patterns very well at high frequencies. Ribbon microphones typically roll off at high frequencies anyway, thus dodging this problem to some extent--but large condenser microphones typically have narrower patterns above 6 - 8 kHz, and to my mind this makes them less suitabie for "S" microphones if you're trying for the best possible results. Neumann claims that their SM 2 was the first commercially available stereo microphone, for example, and that model used small-diaphragm capsules (the same ones as in the KM 56).

--best regards

[Patrick]

What do y'all think about using two different types of mics when running M/S.  Say, you would use a ADK TL for the side channel and maybe a 480 w/ck61 for the mid channel?  Would this sound too weird?   Would the difference between a SD mic and an LD mic be too "strange" sounding.

Absolutely nothing strange about it.  In fact most M/s is done this way.  Since not all small-diaphram microphones have fig-8 capsules (or any interchangeable capsules for that matter), it is not abnormal to mix and match different brands of mics.

[JiB97]

Thanks for all the responses guys.  +Ts all around.

[uosdwis]

DSatz, do you not even monitor the decoded signal when recording? I've been doing some fx recording in M/S without the ability to monitor properly, and even though I haven't really run into any problems I sometimes wonder if my recordings would improve if I could. One more nudge towards a 722, I guess. Any thoughts on this?

[DSatz]

uosdwis, I monitor M/S recordings via a very useful little battery-powered headphone amplifier and M/S decoder that AEA used to make (the "MS 38 Lite"), but I always record the M and S signals directly. So what I'm hearing over headphones is always just one of many possible eventual decodings of the recording.

The M-to-S gain ratio that sounds best over headphones isn't usually the same as what sounds best over loudspeakers--but that's OK, since I'll have the full range of options in post. If it sounds good over headphones, that tells me that the information is also there to make it sound good over loudspeakers later on.

--best regards

P.S.: When recording the M and S signals directly, if your two microphones have the same sensitivity, the S signal may well be 5 or 6 dB lower in level than M. If you raise the S channel's record gain so that both channels peak at approximately the same levels (-2 or whatever), you'll increase the signal-to-noise ratio of your ultimate L/R stereo version by a few dB.

This probably wouldn't matter with 24-bit converters, but when I record M/S on a portable DAT recorder it's what I always do; in classical concerts the audiences are occasionally quiet, so the difference may matter.

[uosdwis]

That's a nifty device. If I find one, I'll pick it up. What I end up doing is monitoring the mid signal and occasionally checking in on the side mic (when switching setups for instance). Beyond that I use my ears to find a good spot and I have yet to be unpleasantly surprised by what I get when decoding in post. But, it's not perfect.

[china_rider]

Hey now... Just wanted to give all a big thanks!  Got my mod TLs back and ran M-S stage lip at on the one tonight.  Had to take a quick listen before bed now that I know I am setting things up correctly.  Still need to do the final mix and am just doing the preview now but wow this M-S sounds good with just the default audition settings.  So again thanks for all the great advice in this thread. T+

[china_rider]

In CEP/Audition:  Effects | Amplitude | Channel Mixer | Mid-Side to L-R preset.  Adjust the width by changing the relationship between the L and R values under New Left and New Right Channel.  The greater the difference between the L and R values, the wider the stereo field.  The smaller the difference between the L and R values, the narrower the stereo field.

Ok... back to my stupidity about this other than default cable hook ups.

Defaults are:
New Left channel has L/R at 100 invert not checked.
New Right channel has L at 100 and right at -100 with invert checked.

Whats the proper way to play with the mix here?  I assume the right chan should always be inverted and the left not (unless I mess up in setting up the mics.)

Should you adjust new left channel L/R evenly and not invert.

So about new right?  Find some happy value for L/R between -100 and 100 and keep it inverted?

Does anyone have a way they go about this to get the best mix?  Right now I'm just sort of randomly changing numbers in preview mode and not saving things.

Sorry for the 3am rambling but I'm so excited about the mix I can't go to bed until I play with it.

Stay Kind,
Dana

[DSatz]

china_rider, sorry I wasn't on the board at 3 a.m. your time, but as I recall in CEP/Audition there is a percentage value which you choose, and it determines the amount of S that goes into the (software) matrix.

The thing is, the optimal value depends heavily on the geometry of your listening setup among other things. If you have your computer connected to your main playback system, you can try a few different percentage settings and converge on a result that seems right. Otherwise you have to try a few plausible settings, write down which one is which, burn a CD from them, and run to your main system to try them out and choose. This can eventually wear tracks in the carpet between your computer and your stereo system. I sometimes prefer one setting on one day and a different setting on another day, but I'm an old '60s person and I find freedom intoxicating. If it were up to me, most musical recording projects would never get finished.

What you're listening for is the best match of stereo width and room reverberation; adding more S increases them both. With 0% S you have pure mono audio and the driest possible mix; with 100% S you have something that's very odd and "phasey" sounding, since it contains only incidental amounts of direct sound, plus you're hearing the same signal in both loudspeakers, but with the two speakers driven in opposite polarity from each other. Somewhere in between is nirvana (if not Nirvana any more).

The true nature of nirvana is that the spatial arrangement is plausible and the music has the mix of clarity plus room sound, all of which gives you the most convincing and pleasant possible feeling, somewhere on the spectrum between "this could really have happened like that" and "I wish this could really have happened like that."

[Brian Skalinder]

Alright, DSatz summed it up far more clearly and eloquently than I ever could have, but I'm going to post, anyway, since I spent the time typing it all up.  If it just confuses things more, I'll probably delete it, but hopefully it will help a bit.

No stupidity here at all, Dana...we all have aspects of this hobby about which we're not fully knowledgeable, and for me, anyway, part of the fun is learning new things.  I learned about Mid-Side a while ago when I had my 414s, others learned about it long long ago, and you're just learning about it now.  Nothing stupid about it, especially since Audition's interface is a bit confusing.  There are LOTS of things about this hobby, especially post-production, about which I'm woefully uninformed and inexperienced.  Just part of the hobby game.

Whats the proper way to play with the mix here?  I assume the right chan should always be inverted and the left not (unless I mess up in setting up the mics.)

There's no real proper way to play with the mix, it's basically trial and error until you find something you like.  You're correct that the R channel should always be inverted, and the left not (unless you mess up setting up the mics).

Does anyone have a way they go about this to get the best mix?

It's been a while since I've used Audition, but, IIRC...

To control the amount of Mid channel in either New channel, adjust the L value.  You probably want the L value the same in both New channels.  In general, the more L (Mid) relative to R (Side), and the more mono the mix will sound.  The more R (Side) relative to L (Mid), and the wider the stereo field will sound in the mix.

The only way to go about finding the "best mix" is to try different values for every recording, since every recording is different.  I've found it's best to not judge the final mix using headphones (though for initial learning, headphones work just fine), but to get it out on my main playback system for critical listening, instead.  At any rate, I usually pick a single track and generate at least a handful of options to see which sounds best, and then refine from there.

To first get an idea of how different values affect the mix, try two extremes and a middle ground:

• Mono stereo field:  For both New channels, set the L value to 100 and the R value to 0.  The net result here is that both New channels have content from the Mid mic only, creating a mono mix.
• Super-wide stereo field:  For both New channels, set the L value to 0.  For the New Left channel, set the R value to 100.  For the New Right channel, set the R value to -100.  The net result here is that both New channels have content from the Side mic only, as if you'd only run the figure-8.  The New Left channel has only the positive (+) lobe of the figure-8, while the New Right channel has only the negative (-) lobe of the figure-8.
• Medium stereo field:  For both New channels, set the L value to 100.  For the New Left channel, set the R value to 100.  For the New Right channel, set the R value to -100.  The net result here is a 50/50 mix between Mid and Side for each channel.  The New Left channel is 1/2 Mid, and 1/2 the positive (+) lobe of the figure-8, while the New Right channel is 1/2 Mid, and 1/2 the negative (-) lobe of the figure-8.

Now, it's very unlikely you'll use either extreme as your final mix.  But it should give you some idea about how changing the values impacts the mix.

I usually started off with a bunch of mixes (L value first, R value second):  100/±100 (an even mix of Mid and Side), 100/±80, 100/±60, 00/±40, 100/±20.  Then I'd listen to each track to see which sounded best.  If even 100/±100 didn't provide a wide enough stereo field, then I'd widen the stereo field by changing the L value to 80, 60, 40, with a lower number resulting in a wider stereo field:  80/±100, 60/±100, 40/±100.  Alternatively, you could lave the L value alone and increase the R value:  100/±120, 100/±140, etc.  The end result is that there's more side in the mix, relative to mid, and so the stereo field is wider.

Anyway...start off with 100/±140, 100/±120, 100/±100, 100/±80, 100/±60, and you'll hear the stereo field getting narrower as the New Right Channel numbers shrink.  Determine which of these files sounds best to your ears.  Then go back and try again, using the track that sounds best as your "center point" (create a few mixes both wider and narrower than the one you like best), and use smaller increments.  So let's say 100/±80 sounds best to your ears.  Create a new set of mixes, say 100/±90, 100/±85, 100/±80 (you already have this one, no need to generate it again, and it's your "center point"), 100/±75, 100/±70, and listen again to see if you prefer one of the more incremental changes better than 100/±80.  If even 100/±60 sounds too wide (probably not, but it could), then keep ratcheting the R value down until you get close to something you like, then refine as noted above.

Over time, you'll learn what general mix sounds better in which environments, locations, etc., and you'll get faster at finding the final mix that sounds best to your ears.  But there's always going to be trial and error involved.

Or...and get this...

You could just install the free Voxengo MSED plugin and use it's very simple interface, instead:



Set the Operation Mode to Decoder, then adjust the Mid and Side gain until you find a point you like.  Same general process as above, but you're twisting the dials (which have values represented in dB), instead of dorking around with entering numeric values in Audition's interface.  In either case, the concept, and trial and error, is the same.

Hopefully I got all that correct...

[china_rider]

Thanks for all the info.  I think I have a good handle on the channel mixer now (never used it before) and I think I'm finally dialing it in.  I've got a mix that sounds great on my headphones and another on my playback system.  Now just to find a happy medium.

One last question, whats the purpose of inverting the new right channel?

[ianstone]

ive been wanting to run M/s now for a while, but i have been unsure that i have the correct mounts


i have the mk6 caps, which are lateral pickup. i don't really know how the fig 8 needs to be pointed although i know i want the cardioid mic pointed straight at the source.
can i do this with a vark bar?

i am thinking that the two caps should almost look like a capital "T"

with the cardioid mic being the vertical line in the T and the fig 8 being the horizontal line in the T -

or do i want them on top of each other, a la "xy" when done with the 4v's on a vert bar (with the tops of the caps almost touching)

?

[Tim]

on top of each other

card faces forward to the source and the the  fig 8 lobes facing out

crude illustration of how the fig 8 should be facing
sound source ====> 8

[ianstone]

on top of each other

card faces forward to the source and the the  fig 8 lobes facing out

crude illustration of how the fig 8 should be facing
sound source ====> 8

awesome, i'm pretty sure i can do it with a vark bar then
thanks +T

[Tim]

+T backatcha

have fun Ian

those mk6 caps look like fun