Topic: 4 channels?

[justink]

if you're gonna run 3 mics (one center omni), might as well run four mics (two omnis A-B, two cards ortf) and go Stereo x2 and mix to your liking... 

[Gutbucket]

OK I'll bite, convince me why that's better.

[Todd R]

I don't know that I want to debate with the person that has probably done more multi-mic recording and mixing than anyone, and this sounds like justink's opening, but --

I don't know that I'd say mixing 2 stereo pairs is necessarily a bad thing or can't be done.  For instance, I'm pretty sure using a coincident stereo pair of either XY cards or MS is a variation on the Decca tree arrangement.

I will say I've been attempting to mix a center, near-coincident cardioid pair with a pair of spaced A-B omnis for many years now.  With the exception of multiple mics onstage, this is really the only way I've tried to mix multiple mics.  And I should say, this has always been done at outdoor shows.  And at that, I've only done it for large outdoor shows, say 8-10K people or more.

Overall, I'd agree with a lot of the comments that 2ch is often or even generally better.  I have had a handful of times where I definitely like the 4mic mix better.  For me, the hope was to get a larger, more spacious soundstage from the split omnis and get the fuller low end from the omnis that don't have the proximity effect of the cards, but have the center card pair to ground things and make sure there isn't a hole-in-the-middle effect.  The few times that it "worked", I feel I got this, and I liked the 4ch mix better than either 2ch pair.

For me, this emerged after doing a lot of outdoor, omni recordings, either with a j-disk or with split omnis.  Generally, I didn't feel I got enough stereo separation with the j-disk, and split omnis generally seemed to have a hole in the center.

So, that's the attempt, and what I think I'm buying.  In terms of specifics, a lot of it has to do with the fact that I keep running my generally preferred near-coincident card pair, so I always have that to fall back on as a 2ch recording -- no downside in attempting the 4mic recording.  From a mixing/phasing/whatever standpoint, I realize that I'd be better off with XY cards center, but I don't prefer the sound of that as a 2ch recording, and I don't want to risk screwing up my fallback 2ch recording.  Same thing for why I don't want a single card in the center or a single omni center with my flanking omnis (so something like Decca), since I still want the option of having my best/favorite 2ch card recording as an option.

On your 3mic suggestion with a center omni:  That seems like it could be good for traditional acoustic recording, I'm just wary of it as recording technique for amplified PA music.  Typically, the mix on a PA rock show has very little panned left and right, with most of the music panned pretty close to the center. (At least, I'm thinking here of the large outdoor shows I use 4mic recording at -- individual instruments and guitar cabinets add little in this case to the overall mix, it really is just the PA.)  So there is very little natural soundstage to be found -- putting an omni as a single center channel would seem to me to then just keep the recording really panned to the center, unless you kept that pretty far down in the mix (in which case you lose the benefit of the extended low end of omnis). 

As an aside, I think using a 1.5-3m spread with A-B omnis pretty much does the opposite -- on playback it creates an unnatural separation, really highly exaggerating the stereo spread, driving the sound to either come from the left speaker or the right.  Done to record an actual orchestra, this would probably sound horrible, as the stereophonic zoom would indicate.  With a very center panned PA-driven concert, it helps really make a bigger and wider soundstage than the PA was providing at the event (due to the center panning), but it also makes for a hole in the middle effect as well.  You could correct the hole in the middle effect by only splitting the omnis 0.5-1m, but then you would keep that center-panned soundstage of the PA stacks.  I think when I've got my 4mic attempts to work, it creates a bigger sounding soundstage with the wide split omnis, and then fills in the missing center of the image with the cardioid pair.

All that said, I've just recently come across the Optimum Cardioid Triangle technique.  I've been thinking I might want to try that out, though it violates my policy of not arranging my mics to keep a 2ch "out" that I can always fall back on.  I might try to do a split A-B pair of forward facing cards, with one card as the center on the OCT, and then the side facing supercard pair of the OCT.  That way I can mix the OCT attempt and still have a forward-facing A-B cardioid pair to use as a 2ch recording if I don't like the OCT.

OCT though does seem interesting to me, in that the side-facing supercard pair seems to do what the split omnis do:  make an exaggerated stereo spread, with I hope would help with the typical center-panned PA system.

Gut -- what was your experience with OCT?

[acidjack]

^ Useful post, Todd.  I don't do many big outdoor shows, but I have a few coming up (Mountain Jam being one) and am thinking about setups.

I was thinking A-B 3ft split omnis + either a center card pair as you suggest, OR, similar to what Gutbucket suggests but different, what about a single center cardiod or hyper, facing forward?  Isn't that sort of the method Nak used to recommend with their 3-mic system (except I believe center was a shotgun) - or am I crazy? 

It seems that if you have that "openness" from the omnis all you want to do is "fill the hole" by getting a bunch of direct sound.  Sort of like MS, but I guess lowering the chances of the Fig8 mic picking up too much wind.

For that matter - any thoughts on outdoor M/S?  I know DigiGal has recordings using that technique.... I just feel like it'd end up being 85% M... 

[Todd R]

I think the old Nak 3mic thing was 2 shotguns with a center omni.  So really on par with what Gut was discussing.

For me, until I heard about the OCT arrangement, I never really saw the benefit of mixing 2 stereo pairs of directional mics (say cards and hypers) -- just didn't see what it was trying to accomplish.  Add to that the bigger potential for phase alignment problems while mixing, and I just never wanted to try it.

Which then left mixing in omni mics.  For me, omnis are just an onstage or an outdoor thing mainly.  If I'm indoors and not onstage, that means either:

a) I'm in a small chatty club environment, and I don't want omnis
b) I'm in a larger arena type environment back in an official taping section, and given the acoustics and room reverb in an arena that far back, I don't want omnis
c) I'm in a larger arena, but am taping up front out of a sanctioned taping section, and omnis would need to be head-height or so, and I'm not taking the chances on omnis where people might be talking
d) I could be stack taping, and omnis could work, but I don't do stack taping, so d) isn't happening

Overall, I'm not using omnis indoors (save onstage like I said), so I'm not doing 4mic mixes indoors.

I did actually do a 4mic mix with omnis for a Trey show in Boulder's Fox theater which sounds excellent, and that is one of the rare times I did a 4mic omni mix indoors period, and it turns out I like it better than the 2ch recording.

I guess I have done a few outdoor shows that were more like 2000 people, but generally my outdoor shows are more in the 10K realm -- Red Rocks shows and festivals mainly.

[Todd R]

Oops, ramble, ramble -- forgot I wanted to add:

If you're at a larger festival and doing a 4mic mix with omnis, you might want to spread them 4 feet or more.  Again, idea being exaggerate the stereo separation, then make up for that by adding in the center cardioid or card pair.

If I were to only separate the omnis 1.5 to 3 feet, I'd be more tempted to not then mess with a 4mic mix and just instead use the 2ch A-B pair of omnis.

[bryonsos]

I run a "phased array" indoors at a club all the time. This is omnis spread @3ft with a cardioid pair in the middle. Sometimes I mix the 4 channels if the PA is thin, but mostly it's to hedge my bets and the undying hope that the crowd will behave. If the crowd is good, then I get the sweet omni heat. If it's chatty, then I have the cards to fall back on.

[justink]

OK I'll bite, convince me why that's better.

i'm just thinking it's easier to handle in post as well as accomplishing what one omni would as well...

but as someone else mentioned, i would imagine the omnis would kill your stereo image a bit, right?  (unless you were running dfc with a wide split)

[Gutbucket]

I completely endorse the general idea of retaining a standard two channel setup whenever possible and building upon or adding separately to that to try different things.  Not enough can be said for being able to fall back on a tested and trusted setup if something doesn’t work out or something fails.  Besides redundancy, working that way provides what I think is the greatest tool we have for deciding if whatever new approach we’re trying is actually getting us anywhere- direct comparison of the new setup to your known, tested and trusted one- same date, same location, same material.

I wouldn’t say mixing two stereo pairs is a necessarily bad thing, only that in my experience I’ve found 3 mic mixes easier to get just as good or better sounding results consistently.. and that when a couple stereo pairs are setup for good stereo by themselves, mixing the them together may not bring enough constant benefits to make it worthwhile if both were arranged in close physical proximity on the same stand.  I see that done frequently and if it works for you great, I’m not telling anyone not to do that.  For comparing two separate stereo recordings it’s certainly the best way to go. But I’m not convinced that’s an optimal approach if the goal is a mix of the both which improves on either alone.  Great for redundancy and comparison though.

Why not? And why do I think mixing 3 mics is likely to be a better bet?  Well, I’m not totally sure honestly, it’s a combination of things I think- some practical, some technical (too much correlation between too many mics).  More on that, and other things like thoughts on OCT (and multi omni imaging) later. 

[Gutbucket]

Some reasons why mixing 3 main mics instead of 4 down to two channel stereo might be less problematic (the short answer is things get complicated quickly)-

I think the technical aspect of the problem is too much correlation between multiple mics- which translates as potential phasing problems (audible as ‘phasing’ or not) by having 4 mics in relatively close proximity pointing more or less in the same direction, not totally coincident but also not far enough from each other for the extra ones to be effectively decorellated.  The reasoning behind this is related to the concept behind the 3:1 rule for individual stage mics, even though the 3:1 rule doesn’t apply to stereo mic’ing, because stereo is all about managing the correlation.  Four mics mixed together is much more likely to work when the two pairs are NOT positioned close to each other- examples are a near-spaced main pair and wide-enough spaced outrigger pair (note that the outrigger pair is then probably so wide it will be too decorellated to be useful on it’s own) OR two stereo pairs that are not in close proximity on one stand- say one pair on stage and another back in the audience.  That can work really well. It gives you control over balancing the direct sound and the room sound, and there are no problems mixing them because the distance between pairs decorellates them from each other.

Practical aspects of the problem: Even with a single stereo pair it’s more difficult to setup moderately wide mic spacings.  I find tapers often setup mics closer together than what might be ideal simply as a practical matter. I regularly see two channel A-B omni setups which are closer than than they should be simply because of practical setup reasons (mic bar is only so long, heavy mics on long bars are hard to support, sightlines, multiple stands are a huge hassle, we can’t do just anything we’d like, etc).  Same goes for directional mics setup without much angle between them.  Many tapers tend to pay more attention to mic angle than spacing, often adjusting angle depending on the venue but leaving the spacing the same.  Narrow mic angles without enough spacing result in too much correlation, which can be reduced by either increasing the angle between them (for directionals), the spacing between them (any pattern), or both.  We don’t want too little corellation either, it’s a balance.  That’s just two channel stereo.  When two mic channels are played back over two speakers, too much collelation simply sounds overly mono but doesn’t cause other problems. Too little makes the speakers sound like separate sources and leaves a hole in the middle, but doesn’t cause other problems.  With more than two mics mixed together to two channels, potential problems arise.

When more mics are introduced into a two channel mix, more space and/or angle between them is needed.  It becomes increasingly difficult to do that as more mics are added which aren’t exactly coincident or spaced far enough apart.

If you want to mix two pairs which are relatively near-spaced, it’s likely to work better if one set is coincident and the other is spaced/angled more than you would otherwise.  That way you have 3 mic locations even though you are using 4 mics.  If you want to use a near-spaced center pair, space the second pair far enough away or point them far enough away in a different direction.

SBD + AUD works well partly because the AUD is decorellated from the SBD by distance.  That setup favors a wider than normal AUD (which is more L/R decorellated) because the SBD, being mono or predominantly so, has plenty of L/R correlation itself and fills the center.  Wide spaced omnis which would have a huge hole in the middle on their own work great when matixed with a SBD for that reason, each stays out of the other’s way.

[Gutbucket]

Mixing (ignoring all that correlation business for this post).

To me, mixing three mic channels down to two is simpler than mixing four.  And four down to two using a coincident center pair is simpler than mixing four with a non-coincident center pair, although none of those need be complicated.

If all you are doing in mixing 4 mic channels is adjusting the relative levels between two stereo tracks and summing them, that’s pretty simple.  That’s a totally legitimate choice, but there are other options which you could exploit if you wanted to, and that makes it potentially more complex. With three mic channels those potential mix decisions are fewer, not that you have to take advantage of them all, but you can adjust them if you choose.

With three mic channels mixed to two, the panning choices are straight forward: pan the left mic channel hard left, the right mic channel hard right, and the center mic channel to the center.  Adjust levels. That’s it.. no more complicated than simple summing of two stereo channels. Of course you can do more than that, and having three raw tacks to work with gives additional options over two channels such as EQing the center channel and the L/R pair differently if you want in addition to or instead of EQing the 2 channel mix.  Same goes for any other processing such as dynamics or whatever.

With four mic channels you add another level of complexity since you can now adjust the panning width and balance of the center pair instead of simply panning a single mic channel to the middle.  When I was using four omnis across the front of the stage frequently, I’d often pan either side of the center pair towards the center somewhat instead of panning them hard Left/Right, which helped firm up the middle and distribute everything evenly across the playback image.  Sometimes the Left channel got panned more to center than the right, or the other way around- whatever worked and sounded best.  That provided good direct sound stage coverage while recording since any one of the four mics was likely to be close to a sound source no matter its location on stage (except for things all the way in the back) and also gave me the mixing flexibility just described which I didn’t have with two channels alone.  But I eventually decided 3 channels of omnis across the front were enough and did pretty much the same job, with less panning decisions, less gear, and less channels to deal with.  Practicality. 

The other factor for me in going to 3 main mic channels instead of four, which doesn’t apply as much around here, was that in addition to 2 channel, I was also playing this stuff back on a multichannel system with three front speakers, so the simplicity of directly routing three main mic channels to individual speakers became attractive.


[edit] I realize that some here use software designed primarily for editing stereo tracks, which may not have simple pan or route to center controls for a mono file.  In that case copying the single center mic channel to make identical Left/Right copies, then mixing those with the Left/Right mic channel stereo file is the same as panning that single center mic channel to center. Not that complicated, but it is an extra step beyond simply summing two stereo tracks, so in that case it is slightly more complicated.

[chinariderstl]

Marking thread.

[Gutbucket]

I run a "phased array" indoors at a club all the time.

Four mics in relatively close proximity can work, it’s just trickier.  The fact that it’s called a “phased array” is insightful in that such a setup exploits the complex phase relationship between multiple near-spaced mics.  Tony Faulkner, the world-class classical recording engineer who developed that setup, certainly knows what he’s doing!  In my thinking, the most important difference between Mr Faulkner’s use of his ‘phased array’ setup and ours is not necessarily that he is recording orchestras in great acoustic halls and we’re recording amplified bands in bars, but that he listens very carefully and adjusts the setup to get it just right before recording, where we more or less throw mics up, record, and really only get to listen critically later.

If I were to only separate the omnis 1.5 to 3 feet, I'd be more tempted to not then mess with a 4mic mix and just instead use the 2ch A-B pair of omnis.

I agree.  IMO, the benefits of thee omnis becomes apparent when spacing the outside pair 3’ apart or more.  If the center omni is included, 4’- 6’ between the outside pair is better in my experience, which means a distance from center omni to each outside one of 2’- 3’ or so and that’s the distance I feel like I can split two omnis safely without the luxury of being able to listen closely before pushing record.

I was thinking A-B 3ft split omnis + either a center card pair as you suggest, OR, similar to what Gutbucket suggests but different, what about a single center cardiod or hyper, facing forward? [snip..]

It seems that if you have that "openness" from the omnis all you want to do is "fill the hole" by getting a bunch of direct sound.  Sort of like MS, but I guess lowering the chances of the Fig8 mic picking up too much wind.

For that matter - any thoughts on outdoor M/S?  I know DigiGal has recordings using that technique.... I just feel like it'd end up being 85% M...

A directional center mic can work well between wide spaced omnis by helping to focus on the direct sound like you say.  In that sense it is conceptually similar to matrixing a spaced omni AUD with SBD to reinforce the center with clear, direct sound.  I’ve used a supercardioid before and also a Blumlein pair in a M/S arrangement as a center mic between wide spaced omnis, which gave me loads of options using only 4 mic channels in 3 mic locations.  The few times I did this I only spaced the omnis about 3', partly because that was my 'standard config' which I'd run there before and gotten good results with (Spirit of the Suwannee amphitheater)and I wanted to keep that unchanged as a basis for comparison.  Doing it there again I'd split them 6' if possible.  Here's some of the options which that gave me to play around with using the resulting 4 mic channels:

Two channel L/R stereo-
● L/R= Blumlein stereo
● L/R= Blumlein stereo mixed with the wide spaced L/R omnis.
● L/R= omnis mixed with forward facing figure-8 panned center.

Three channel L/C/R stereo-
● L/R= omnis, Center= forward facing figure-8.
● L/R= Blumlein (M/S adjusted wider than normal), C= forward facing figure-8
..etc

5/6 channel surround-
● L/R= Blumlein (M/S adjusted wider than normal), C= forward facing figure-8, Ls/Rs= wide spaced omnis. Optional Center-Back surround= side facing figure-8 (delayed)
● L/R= omnis, Center= forward figure-8, Ls=side facing figure-8, Rs= side facing figure-8 with reversed polarity.
● etc..

The multiple channels available on the DR-680 gives me lots of options to play around without messing up my standard setup.  This year at Springfest instead of running the wide omnis with Blumlein at the center, I essentially ran two rigs into one recorder: The first was 3 spaced omnis in a Decca tree-like arrangement with a 6' wide spacing on the L/R mics and the center omni about 1-1/2' forward (actually 4 omnis, the extra one extending out to the rear of the triangle just for use as a surround channel), the second setup was a pair of Gefell cardioids in ORTF at the center of the Decca tree triangle.  I've listened to both setups seperately, but haven't had the time to play around with combining them and that wasn't my primary intent anyway. 

That setup sounds complicated, but everything was pre-rigged on a single small stand which I could walk in with and attach to a large stand I had staked down FOB and all six mics went into the one recorder, so it wasn't difficult to run.

[Gutbucket]

[snip..]
Generally, I didn't feel I got enough stereo separation with the j-disk, and split omnis generally seemed to have a hole in the center.

I felt the same way with the J-disk, I liked the baffle effect, but always wanted more spacing.  With that idea in mind I built one big enough to work with a wider mic spacing, but never had the gall to fly it.  I’ve thought about using two normal sized baffles, one for each mic with the typical mic to J-disk spacing distance on each, which could then be spaced farther apart and even angled, but never perused that farther than the idea stage.  Just too unwieldy.

With either wide spaced or baffled omnis, the best answer for me has been a third omni in the center.  The difference in application (not sound) is that three non-baffled omnis need to be spaced more, like the 3’ between each, 6’ total I've mentioned above.  Baffling allows me to get away with spacing the mics less, allowing me to still use three placed closer than I could otherwise by making them more directional. 

An unused stage monitor at the center-front of the stage can be used as a 3 omni baffle with small omnis by simply  taping them to each face of the monitor, but usually I just space the omnis for stage-lip or on-stage use without using a baffle.  Most of the time when I use the near-spaced baffled 3 omni technique (actually in my case 4 omnis, with the extra mic baffled and facing away from the stage to eliminate as much direct stage sound as possible for use as a surround channel, also sometimes useful for adding additional room ambience or crowd reaction in the 2 channel mix) I am the baffle. In that case the L/R omnis are spaced ~20”.  All the mics are baffled, the center and back ones more so, the center facing forwards and the L/R facing the sides.  From the best location, in a good room I prefer this even over wider spaced unbaffled omnis as the imaging is more precise, though the envelopment and bigness is somewhat less than wider omnis, which may be somewhat unnatural anyway, but sounds good to me.  But that technique doesn’t work well for much of what is recorded around here in a standing crowd.

On your 3mic suggestion with a center omni:  That seems like it could be good for traditional acoustic recording, I'm just wary of it as recording technique for amplified PA music.  Typically, the mix on a PA rock show has very little panned left and right, with most of the music panned pretty close to the center. (At least, I'm thinking here of the large outdoor shows I use 4mic recording at -- individual instruments and guitar cabinets add little in this case to the overall mix, it really is just the PA.)  So there is very little natural soundstage to be found..

To me that’s an argument for what may be an unnaturally wide (if very good sounding) stage which three non-baffled omnis produce.  It also gets back to the thing about why record a mono PA or solo acoustic performer in stereo at all.  Even a single point source on stage or a mono PA fills the room and produces a complex soundfield.  And that soundfield where your mics are (or head is) is very 3-dimentional. 

As an aside, I think using a 1.5-3m spread with A-B omnis pretty much does the opposite -- on playback it creates an unnatural separation, really highly exaggerating the stereo spread, driving the sound to either come from the left speaker or the right.  Done to record an actual orchestra, this would probably sound horrible, as the stereophonic zoom would indicate.  With a very center panned PA-driven concert, it helps really make a bigger and wider soundstage than the PA was providing at the event (due to the center panning), but it also makes for a hole in the middle effect as well.  You could correct the hole in the middle effect by only splitting the omnis 0.5-1m, but then you would keep that center-panned soundstage of the PA stacks.  I think when I've got my 4mic attempts to work, it creates a bigger sounding soundstage with the wide split omnis, and then fills in the missing center of the image with the cardioid pair.

The middle mic fills the wide spacing for orchestra recording, which is the basis of the classic wide A-B 3-mic Mercury label technique from the ‘50s which had all three mics quite widely spaced in a line across the front of the orchestra.  Like the not as widelly spaced Decca tree technique, they mixed those three channels down to 2 for LP, though I understand the original 3 channel recordings were made available as SACD or DVDA multichannel releases in the early ‘00s.  I've never heard them myself.

The mostly mono two channel PA with stacks on each side of the stage is an odd animal in the acoustic world.  I think wide 2 channel omni splits can work better for concert recording (if tricky to predict) than for acoustic material or anywhere else specifically because of that.

[Brian Skalinder]

Generally, I didn't feel I got enough stereo separation with the j-disk, and split omnis generally seemed to have a hole in the center.

I felt the same way with the J-disk, I liked the baffle effect, but always wanted more spacing.  With that idea in mind I built one big enough to work with a wider mic spacing, but never had the gall to fly it.

Did you guys ever try the J-disk with tighter mic spacing, i.e. mics closer to the disk?  Using a j-disk, in my experience: the tighter the spacing, the greater the stereo separation.  I've gotten the spacing too tight before, producing too much stereo separation -- though it was fairly up close for an unamplified vocal performance.