Oddball microphone technique (OMT) - part 2

[heathen]

Heathen, any possibility of posting short (say minute long or so) samples of the omnis and cards on their own prior to mixing?  No worries if that's too much hassle.
MP3's uploaded directly to the thread would be perfectly fine.  The size limitation on each file in that case is 750KB.

No problem at all...I'll try to remember to do it tonight.

[heathen]

Here's a clip of the raw CA14 and AT4031 files: https://we.tl/GGNxlvRxY6

All I did to these was to make each into a stereo file (you can split them back up into individual channels in you DAW, of course) and trim them down to the short clip size (I tried to find a passage that has loud and quiet parts).  They're both of the exact same passage of music.  Link should be good for seven days.

[kindms]

heathen. grooving the MSMW.

Curious about your thoughts as well as others. rocksuitcase and i were discussing this the other day. You mentioned talkers etc. We are sort of thinking that the OMT has a tendency to mitigate the crowd to some extent. we were thinking perhaps as its more "spread out"  vs. a more direct in a 2 channel scenario.

curious what you guys think about that

Im only about 15 minutes in on the recording and so far the crowd is there but I haven't found it distracting etc

[rocksuitcase]

heathen. grooving the MSMW.

Curious about your thoughts as well as others. rocksuitcase and i were discussing this the other day. You mentioned talkers etc. We are sort of thinking that the OMT has a tendency to mitigate the crowd to some extent. we were thinking perhaps as its more "spread out"  vs. a more direct in a 2 channel scenario.

curious what you guys think about that

Im only about 15 minutes in on the recording and so far the crowd is there but I haven't found it distracting etc

This is one of Greisenger's concepts which informs the OMT theory: decorrellation- the omnis are sufficiently spaced and separated from the center microphone(s), the more "unique" each signal is, time of arrival and pressure difference wise, the more decorrellated the signals are said to be. This is one reason why chatter seems less annoying when doing a 3 or more channel OMT using spaced omnis as the anchor. Correct my cite GB if wrong   

[heathen]

Forgive me for what may be stupid questions, but doesn't OMT still pick up the general din of the room?  Any particular yell or the like may be decorrollated, but that background constant level of noise will remain right?

[rocksuitcase]

Forgive me for what may be stupid questions, but doesn't OMT still pick up the general din of the room?  Any particular yell or the like may be decorrollated, but that background constant level of noise will remain right?

Oh yeah, same as every/any mic. There is always the room chatter these days, I feel for you there, BUT, your recording is very good.
re decorrelation: I think what I hear is when I know a person was coughing or yelling to one side (i.e. I'm sitting next to them), it seems that cough or yell isn't as present on the recording. Of course, anything loud near any mic will be in the mix at one level or another.

[Gutbucket]

You've got it.  The audience sounds are still there and are not reduced in level, but the way in which that is portrayed and perceived during playback is different.  The audience sound pickup is directly related to the how room ambience is picked up and portrayed - both are more-diffuse, and open and "out there", perceptually separated to some degree from the music of interest which is more highly focused in the center front quadrant.  The wide omni spacing does this in an increasing way for direct sound which arrives further off the central axis and especially so for the diffuse sound which is effectively arriving from all directions (in this case that diffuse audience sound component is more murmur and general din, as opposed to any specifically identifiable audience sounds).

Acually, due to the omnis we're most likely picking up more audience and room sound from all directions than we would using a single pair of directional mics, yet that content (at least the room sound and the desirable audience reaction) is less in conflict with the music because of the OMT arrangement because it picks up the direct sound and the room/audience sound in different ways before we combine them.  Of course there remains a lot of overlap between them, and that's a good thing.  But the differentiation is important and useful.

To take best advantage of that differentiation, it helps to keep the central mic pair as focused on the direct sound pickup from stage and PA as possible, with the omnis less intentionally focused on that and more oriented on the diffuse sound arriving from all directions.  So more-directional PAS center pairs are helpful for maximizing direct sound pickup from stage and PA, and a coincident arrangement of the center pair is helpful in producing a tighter and highly-correlated type of stereo for that direct sound component (contrasting against the more-decorrelated and diffuse omni pickup of room/audience).  Omnis which don't especially favor the forward direction are helpful, achieved by using miniature omnis which have less directionality due to their small size, or omnis with diffuse grids like the AKGs rocksuitcase and kindms are using, or pointing more directional omnis sideways or backwards.  And spacing the omnis as far as practical helps decorrelate the diffuse pickup to a lower frequency, narrows the SRA angle of the omnis, meaning more of the off-axis audience direct sound pickup by them will image far-left or far-right leaving room for the center pair stereo stuff in the middle.

The other thing which spacing the omnis sufficiently helps with is audience sounds close by the recording position.  Close sources will tend to be considerably louder in one omni than the other and that will cause them to image further to one side or the other, leaving more perceptual separation between that and the music which is imaging more tightly in the center, allowing it to be more easily perceptually ignored even though its still there (cocktail party effect).  Again, the setup is actually more sensitive to close audience sounds due to the omnis, but the arrangement helps to make that seem less in conflict with the music.

[Gutbucket]

Heathen, thanks for the MSMW.  Sounds good (dig those guys).  And very cool that you were able to manage a 7' spacing on the omnis.

Here's a clip of the raw CA14 and AT4031 files: https://we.tl/GGNxlvRxY6

Thanks for that.  It's helpful to hear each part in isolation.  Here's something which I think will make things even better.

Are you one who does any EQ?  No problem if not.  This is a big topic in general, and even more so with regards to the potential application to OMT where each part can be EQ'd separately and/or the resulting combination can be EQ'd as well.  In general, its a good approach to EQ the center pair and omni pair separately to each sound as good as possible on their own prior to combining them.

But there is an important exception to that, which can be useful to do even if you don't really care to make the effort to EQ the two parts separately.   It's this- Often it helps to clean the sound up greatly by reducing the low frequency content of the center pair.  Let the omnis provide the majority of the low and upper bass content.  Its okay, even good, if the center pair is more-focused on the midrange (especially) and high frequency content.  Directional mics with weak bass response (like those more intended for speech pickup) can actually work well as center pair because of this, since their response is sort of pre-EQ'd that way already and fits the roll.  Otherwise it helps to use a gentle roll off from the below the upper mids to reduce any excess "thickness" and boom in the center pair.  And although I've not tried it, a high pass filter could work for this, especially if you can set the slope to first-order (-6dB per octave).

This represents another aspect of the "division of labor" sort of thing described in the previous post, letting the omnis do their thing down low without conflict.

And a midrange focus in the center pair plays nicely against the tendency of omnis to be less forward in that range.  Again, the mic patterns used sort of naturally do this for you, but you can emphasize the trend somewhat with EQ if you want to play with it, just keep the difference moderate.

At higher frequencies I find the interaction between the omnis and center pair to be productive.  No need to reduce one to let the other "breathe" sufficiently.  Probably partly because the signals are sufficiently decorrelated at high frequencies (short-wavelengths) so the content is different enough, and partly due to how we perceive phase differently at high frequencies, where  randomized phase tends to sound "airy" and "open".



TLDR- reduce any tendency toward "mud" or excess "thickness" in the center pair via mic selection or EQ, even erring toward what would be an "over-thin" bass-weak sound when the center pair is listened to in isolation.

[heathen]

I'm not sure if you just made a typo, but my omni  spacing was 70 inches.

As for the EQ, I have EQ'd some recordings in the past but I still don't feel very comfortable with it.  While I'm not opposed to messing with it more and trying to get better, I'm leaning more towards mics that sound the way I like without need for EQ (hence my love of the AT4031...maybe not for everyone, but to me they sound damn near perfect).

If you tinker with that clip to the point you get something that sounds good, I'd love to hear what you come up with (and perhaps more importantly, how you got there).

I'll admit that for this show I was thinking of the AT pair as my primary focus, and OMT was a secondary thought.  That's why the ATs were DIN...if I were primarily focused on OMT then it seems from the recent comments that having the ATs PAS or XY would have been better.

[rocksuitcase]

^^^^^^^^^^^^^^^^^^^^^^^^^^
Heathen- I find that unless conditions are ideal, often I like to EQ the omni channel usually taking out high end above 3kHz gradually to 20kHz (always depends on source material)
Then import the middle channels, and typically do not need to EQ our AKG ck61ULS or ck3's. but if I do I do it separately from the Omnis as GB mentions. THEN
I level the two stereo sources against each other, typically boosting the center 2-4 dB over the omnis, but it can go the other way as well. (often coming up with different rations of Omni to center even between sets)
THEN export to two tracks THEN re-level or do a light overall EQ if something seems it needs shaping. ( I like to NOT do EQ at this stage if it all sounds right)

IMO, the complete approach to OMT involves knowing there WILL be post processing involved and much more than with 2 channel recordings. For me, after all the years of running co-incident patterns this last 3 years has been much more FUN doing the various things involved to go multi-channel both equipment wise and in my interactions with FOH and musicians. I have had several musicians, producer types, ask about the mic technique and in fact have had a few of them reach out after the event to ask for their raw files. (which I eagerly provide)

I will post a wetransfer just for you of one of my best takes like this using 4 channel OMT plus 2 ch SBD. The band isn't "taper friendly" so it will be on the DL via PM to you.

I'll admit that for this show I was thinking of the AT pair as my primary focus, and OMT was a secondary thought.  That's why the ATs were DIN...if I were primarily focused on OMT then it seems from the recent comments that having the ATs PAS or XY would have been better.

If this is the case, then your recording is all that much "better". I hear a nice wide soundstage plus a relatively centered overall image.

[heathen]

Thanks for the input, and for sharing the file!

[Gutbucket]

My mistake, I meant to say 70 inches not 7 feet, yet the point stands - that's more spacing than most tapers are able manage from a single stand, and is in the enough for "good OMT spacing" IME.

No worries in considering your DIN pair primary, and that's the way I'd approach it if I was just getting into OMT.  Best to start from what you are comfortable with and what has worked well for you in the past, using that as a baseline of comparison for whatever changes you end up making, before changing things up too much.

Conceptually I work the opposite way, starting with the omnis alone and getting them balanced and sounding good by themselves, then embellishing on that with the other channels I have available.  I like starting from a well balanced, natural ambient sound, then adding presence, focus, imaging and depth through the addition of the center mic(s).  Either approach can work, this just works best for me, partly because I began taping with spaced omnis rather than near-spaced pairs so this approach comes naturally to me.  But I also think it's helpful to build the mix this way from the bottom up (from the ambience inwards).  And when the omnis are considered the primary pair, that sort of opens things up to more non-mainstream options for what one might do with the "additional" center mics in trying various things over time.

The omnis sound pretty good alone, don't they?  I bet a lot of tapers may be surprised by that.  Especially since they are facing to the sides and not forward.

[Gutbucket]

Don't worry about all this EQ stuff.  But do listen to each pair on its own, and I encourage you to listen carefully to identify the frequency balance thing I'm talking about in the center channel or center pair.  Not just with respect to this recording, but in general when using this technique.  Listen for any excessive energy or resonance in the lower mids and bass in the center pair which may tend to cloud the clarity of the mids in any way.  It also helps to compare what you hear in that below-mid frequency region with what the omnis are providing.

One way to test out this center pair frequency shaping aspect for yourself without using any EQ afterwards is to try switching in the low-cut on the mic bodies (if they have them) or on the recorder for the center pair, although that does sort of commit you to mixing with the omnis instead of keeping a fully independent center pair.  Besides making everything clearer in general and improving imaging, this can also allow you to use somewhat more center-pair level in the mix if appropriate, without cluttering things up.

[heathen]

The omnis sound pretty good alone, don't they?  I bet a lot of tapers may be surprised by that.  Especially since they are facing to the sides and not forward.

Yes, I think I've commented before that I was surprised at how good the omnis sound by themselves since this was indoors.  I think I also made an off-hand comment wondering if Chris Church tunes his omnis for being used farther from a sound source, since most tapers are doing that (would that be diffuse field?  I can't keep it straight).  If that's the case, I would guess that's part of why they sound so good.

If nothing else, this is definitely going to make me more inclined to try this sort of thing in the future, even indoors.  I've got some indoor shows coming up in the next few months where I might be able to try more OMT stuff (of course, I've also got some indoor shows were the spacing of my mics will be limited to the dimensions of my stupid head  ).  I'm already trying to think of ways I could mount these omnis at Red Rocks this summer.  (Summer?  Sheeeeeeit...I'm going to start seeing show there in a month!)

I think I'll also play with this recording a bit more, taking your approach of starting with the omnis.  That seems to be consistent with rocksuitcase's approach of EQ'ing his omnis first.

[Gutbucket]

I find that unless conditions are ideal, often I like to EQ the omni channel usually taking out high end above 3kHz gradually to 20kHz (always depends on source material)

I don't mean to turn this into too much of a discussion on tweaking EQ (keeping Heathen in mind at this point, but have been meaning to explore OMT EQ in depth, so definitely later).  And some of the difference in approach may be a result of the different omnis we are using.  But as a general trend, I don't like throwing away that high frequency omni ambience, for me much magic lies there.  It can open up the recording, and also provide sufficient high frequency energy in that region when that's needed without becoming strident, as can sometimes be the case when trying to add sufficient energy there by boosting that region in the center pair.  Again in general, I find a sort of loudness-curve shaped response tends to work well for the omnis (sort of the classic lopsided smile-shaped EQ, mostly emphasising the bottom, less mids, yet rising at the very top providing sufficient high frequency "air" so it doesn't sound dull or muffled, maybe somewhat distant but still retaining a nice live "sparkle" and naturalness.   And that loudness-curve trend thing sort of makes sense when you consider the lower level of the ambient sounds compared to the direct sound.

As for a generalization on EQing the center pair, it's mostly that smooth reduction in low frequency content as mentioned for less competition with the omnis down there.  If I want more clarity or presence (SBD like) I'll work the mid range and push up between 700 and 3kHz in the center mic or pair, which acts somewhat like a zoom control for vocals.  The part I find curious but seems to work well for blending the center seamlessly and allowing more center energy without the center over dominating the mix is a cut centered somewhere around 7-10kHz or there about. Keeps the center from getting "shouty" and calling attention to itself.  Perceptual HTRF stuff going on there for sure.*  I'm doing this monitoring through three front speakers (dedicated center speaker), but it seems to work the same (as far as I can tell, at least similarly) mixing to two channel stereo.



*BTW, I contacted David Greisinger recently and he kindly sent me his Personal Headphone Equalization App, which helps the user find a personal headphone equalization curve which matches the response of a speaker located directly in front of the listener.  The user adjusts frequency bands and balance controls to find the same apparent centered loudness level in each band.  The resulting curves can then be used for any stereo source listening by that user through those headphones.  This corrects for both the specific headphone's response as well as the listeners personal HRTF, producing much more natural headphone listening both in terms of frequency response and "out of head" spatial imaging.  I've not had a chance to do anything with it yet or go through the process yet, but am excited to do so.  I've asked his permission to discuss it on this board, which I plan to do in a separate thread at some point.  I suspect this is related to that lower treble center EQ cut I describe above.