Become a Site Supporter and Never see Ads again!

Author Topic: Relationships between spaced omnis and center mic  (Read 22286 times)

0 Members and 1 Guest are viewing this topic.

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (15)
  • Needs to get out more...
  • *****
  • Posts: 15698
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: Relationships between spaced omnis and center mic
« Reply #15 on: June 22, 2017, 04:43:18 PM »
There are several different reasons one may want to use a center mic (or center pair, or even more) between a pair of spaced omnis.  Here are three reasons which immediately come to mind-

To solidify or otherwise improve the center of the playback image.  Which is to say- To allow a wider spacing of the omni pair than one could otherwise get away with.
To gain some useful control over the balance between direct sound and the ambient sound (ambient sound being primarily room reverberance and audience reaction).
For playback over three front speakers instead of two.

The last won't apply to the majority here at TS, so the first two will be my focus.

I almost added "To more evenly cover a large, distributed sound-source such as a choir or orchestra, or a relatively close stage-lip recording of an ensemble with a wide on-stage setup", but realized that's really just a subset of the first one.
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (15)
  • Needs to get out more...
  • *****
  • Posts: 15698
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: Relationships between spaced omnis and center mic
« Reply #16 on: June 22, 2017, 05:29:06 PM »
To back up just a bit, please note that the reasons mentioned above are quite different than those which would apply to what may seem a very similar question on the surface of things, but really isn't-

"What about adding a pair of omnis to the 2-channel mic setup I am familiar with and like using?"

The reasons for wanting to do that would be quite different.  This is partly why I started with a bit of philosophy of recording.  Adding omnis to a standard 2-channel setup changes the way the standard setup works in complex ways which aren't obvious.  Commonly, that gets ignored.  The understandable temptation is to stick with the known-good two channel setup and add additional mics in an attempt to make things better.  Unfortunately it's not that simple.  Well in some cases it can be - low pass the added omni(s) around where the 2-ch setup naturally rolls off and that keeps everything relatively simple.  It won't significantly change the way the 2-ch setup behaves in the range in which it is sensitive.

The main point is this- typically adding another mic channel (or two, or more) is going to require significant changes to the original setup in order to see a significant benefit.  And for things to work optimally the setup will definitely need to be changed.
« Last Edit: June 22, 2017, 05:30:39 PM by Gutbucket »
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (15)
  • Needs to get out more...
  • *****
  • Posts: 15698
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: Relationships between spaced omnis and center mic
« Reply #17 on: June 22, 2017, 06:02:27 PM »
So let's go back to the two main reasons for using a center mic with a pair of spaced omnis-

1) To solidify or otherwise improve the center of the playback image.  Which is to say- To allow a wider spacing of the omni pair than one could otherwise get away with.
2) To gain some useful control over the balance between direct sound and the ambient sound (ambient sound being primarily room reverberance and audience reaction).

From my own experience, it's clear that unless setting up on stage or at the stagelip, most tapers usually can't achieve a very wide spacing between omnis for simple practical reasons.  It's hard to support a wide spread from a single stand, and it's just too much hassle to set up (and protect) multiple stands.  Sure there are exceptions: balcony clamping with wide splits, great sounding superwide outdoor two channel omni A-B section recordings and the like - yet the fact that those can work at all more than anything says much about the odd nature of PAs as a sound source and how unusual ways of recording them from a distance can work which otherwise would not for most anything else.

So for most tapers, adding a center mic will usually be less about filling a hole in the middle of the soundstage than about pulling in more direct sound and making the simple omni pair sound more present and upfront.  And that's reasonable for relatively narrow omnis spacings like most here will be using.  There is no hole-in-the-middle problem with relatively narrow omni spacings, rather the opposite problem becomes the issue- too much common information picked up by all three microphones.
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (15)
  • Needs to get out more...
  • *****
  • Posts: 15698
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: Relationships between spaced omnis and center mic
« Reply #18 on: June 22, 2017, 07:08:59 PM »
One post on approaches to this from the mixing side before getting into mic arrangement specifics-

So we're probably using a directional mic in the center pointing directly forward mostly to achieve more focus on the direct sound, with less room and ambient pickup in that channel than were're getting from the omni pair on it's own (If the primary goal was filling a hole-in-the-middle between widely-spaced omnis, a 3rd omni would work best, blending in better, with timbre and spatial pickup qualities more similar to the other two omnis). 

So we make some trial recordings and play them back to asses them.  We pan the omnis hard left and hard right and pan directional center mic to the middle. We bring up just the omnis and balance them.  Then we slowly bring up the center mic.  The first thing one is likely to hear is likely a simple increase in level.  The second, a difference in timbre or overall frequency balance.  Does it sound better?  Maybe so, maybe not, but let's assume it does.  We have lots of control over those aspects and to my way of thinking we should use that control in a couple ways: First of all to keep ourselves honest - does it sound better because it's louder?, does it sound better because the frequency balance is changed?  The only way to know is to listen both ways and compare.  Adjust the level and EQ of the two omnis alone to sound as close as possible to what you heard when you brought up level of the center mic.  Is that just as good?  If so at least at this point into the investigation of usefulness there is no reason to record using the center mic, unless you are using it as an easy way to avoid having to EQ (nothing wrong with that).  The point is to constantly check changes and adjustments in an iterative way, to to make sure actual improvements are being made without the addition of  unnecessary complications, and to prove to ourselves that the way we are doing it can't be achieved in some other simpler way.

The second way of using that control is to fine tune everything so it all works together more optimally.  Maybe you find you can only use a tiny bit of center channel without things sounding weird (let's say because the omnis aren't really far enough apart for a center channel to work right without complications), but if you EQ the center channel in a certain way that mitigates the problem and allows you to bring up the center channel level to where it's contributing a lot more good stuff.  Again it's a balancing act, how much is the right amount, and of what?  Some of these changes require these kind of adjustments to work right.  EQ, compensation delay, etc.

Some folks will be averse to adjusting all this stuff.  That's fine, if it works for you without adjusting things like EQ and fine-tuning levels, great.  But a huge advantage in having more than two recorded channels is the greatly increased ability to tweak things which doesn't exist with just two channels directly feeding two speakers.  Once you've sort of figured out what works and what needs to be done, you can better asses if it is worthwhile or not, and when.

The level, EQ, and timing of the center may be adjusted (for any number of reasons) without throwing off the left/right balance.  Use that to your advantage.  Just double check to make sure that what you are doing is really an improvement, and that you can't achieve the same in a simpler way.
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (15)
  • Needs to get out more...
  • *****
  • Posts: 15698
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: Relationships between spaced omnis and center mic
« Reply #19 on: June 22, 2017, 07:47:55 PM »
Okay now that that's out of the way, here's some straight dirt on geometry.

A few different aspects are in play.  First time of arrival for the direct sound component-
If all the mics are arranged along a single line (center mic in the same plane as the omnis), arrival of a wave-front originating from directly ahead will be coincident in time as it reaches all mics.  That's generally thought of as a good thing, because at the opposite extreme we certainly don't want a slap echo between the center and omnis, or comb filtering timbre effects at less dramatic timing differences.  But it may be advantageous to use a bit of precedent effect to advantage in drawing attention to the center without needing as much center level to do so.  A center mic position slightly forward means a wave-front from directly ahead pings the center mic a fraction of a second before the omnis- not enough to hear an echo, not enough to change timbre radically, but enough to pull the image to center more strongly via the precedence effect.  A slightly forward center mic can also make the imaginary line between the center mic and left omni perpendicular to the left PA (and vice versa for the right omni, center mic, and right PA) so that the wave-fronts arriving directly from each PA are synchronized in time at the center mic and mic on that side, yet slightly delayed on the opposite side.  Again, if that's a good or bad thing is totally a judgement call.

If the center is far enough out in front (why that might be the case I'll get to later) a compensating delay to it can bring that timing alignment back into sync, but that will only do so for sound arriving from that particular direction- in this case for sounds arriving from in front.  For sound arriving from the rear the compensating delay is added to the time of arrival difference between the omnis and center.  For sound arriving from the sides somewhat less so. That might be used to advantage- or not.

Besides direct-sound wave-front time of arrival aspects, the other main thing with regards to setup geometry is the phase-correlation between channels with regards to diffuse sound pickup.  That deals with how the reverberant/ambient sound which fills the room is handled and is a large part of why spaced omnis sound the way they do.  In a lot of ways, that's the invisible elephant in the room.  I'll try and get to that tomorrow.
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Offline heathen

  • Trade Count: (23)
  • Needs to get out more...
  • *****
  • Posts: 3528
Re: Relationships between spaced omnis and center mic
« Reply #20 on: June 23, 2017, 10:38:19 AM »
So much great information here...I'll be thinking about this, and referring back to this thread, for a while.
Mics: AT4050ST | AT4031 | AT853 (C/SC) | Line Audio CM3 | Sennheiser e614 | Sennheiser MKE2 | DPA 4061 Pre: CA9200 Decks: Zoom F8 | Roland R-05

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (15)
  • Needs to get out more...
  • *****
  • Posts: 15698
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: Relationships between spaced omnis and center mic
« Reply #21 on: June 23, 2017, 11:46:46 AM »
Cool, glad your enjoying all this and I hope it helps.  My aim is to explain what's going so you and others can make educated decisions on what to do, rather than just following convention by rote.  To me that's more helpful than just saying "do this".  It's what I was looking for when I first came around here years ago.


Continuing with geometry, now getting to what I consider the important stuff about spaced omnis.

What is it about a simple pair of well-spaced omnis?  For now I'm just considering a 3' spaced pair, without any center mic.  They have a few traits which are often valued.  A few of those are simply due to them being pressure omnis: They have deep bass extension; they are resistant to wind and handling noise; the really small ones are not very particular about how you point them; they are generally more weather resistant; and they tend to sound natural.  Some of that natural sound quality is due to their generally flat non-resonant response and open polar pattern.  Of course, the open omni polar pattern is good when the room sounds good, but might not work as well if the room or audience is poor.

Yet some of their other advantageous traits are a mostly a product of the spaced setup geometry- specifically a wide-enough spacing.  Wide-enough spaced omnis sound big and open, lush, wide and ambient, and certainly some of their associated naturalness is due to that.  The wide-enough spacing produces stereo bass information down to a significantly low frequency.  Narrow spaced omnis don't sound as open, airy or as lush, and are essentially remain monophonic up to a higher frequency.  Why is that?  A lot of the open, airy lushness is to do with how they pick up the Diffuse-Field compared to the Direct-Sound.  I'll post about that next.  The stereo bass quality is due to the relationship between the mic spacing distance and the wavelength of the bass frequencies in question - the wider the mic spacing, the lower the cutoff frequency where the mic pair can detect sufficient phase differences between channels to convey low frequency stereo information.  Below that point , the microphone sensitivity retains pressure omni response flatness down to the lowest octave, yet as the frequency decreases and the wavelength grows large in comparison to the mic-spacing distance, the phase difference information picked up between the two mics decreases until the bass becomes mono-ized below a frequency that corresponds with the mic-spacing distance.  The wider the spacing, the lower the stereoized bass quality goes.  The narrower the spacing, the higher the frequency where the bass goes mono.  This low frequency difference information has a large effect on bass externalization and the  spatial qualities which convey a sense of "you are there".  A lot of that depends on having sufficient spacing.

That same relationship also effects a related aspect of the bass quality from spaced omnis- the mic's themselves each have a very flat bass response as pressure omnis, but the interaction of the two mics with each other based on the distance between them creates a partial reinforcement/cancellation curve.  You get a reinforcement peak where the spacing is equivalent to a one wavelength spacing and a partial cancellation dip at half-wavelength spacing, and that shifts with frequency as the spacing is changed.  This can be quite audible with lots of bass information as the spacing between mics is adjusted.

When I first started seriously recording outdoor concerts with just a pair of spaced omnis, I did a series of experiments listening using highly isolated headphones (wearing ear-muffs over isolating in-ear 'phones) while making spacing adjustments to the omnis.  Besides changes in imaging width and diffuse ambient pickup, I could clearly hear the changes in stereo-bass depth as the spacing was varied.  I expected to hear those things. What I had not expected was the relatively strong reinforcement/cancellation curve thing.  It was quite apparent as the spacing was shifted, I could 'tune' the sound of the low bass using the spacing.  Obviously that's not something one would do each time, but at least in that particular ampitheater I settled on an approximate 3' spacing being the best of all worlds with regards to low bass tuning, stereo bass effect, imaging width, and ambient/diffuse qualites, and that remained my standard omni spacing for years until revisiting it again when I started adding center mics.  That the basis upon which I suggest 3' as a good default target spacing for a pair of omnis, and the minimum spacing I recommend if using a center mic.  As always, there can be exceptions.  But I stand by that recommendation and its usually worked very well for me.

Okay that covers stereo bass qualities.  Next I'll touch on to stereo imaging, then diffuse-field correlation.  And finally the implications all these things have when adding center mics.
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (15)
  • Needs to get out more...
  • *****
  • Posts: 15698
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: Relationships between spaced omnis and center mic
« Reply #22 on: June 23, 2017, 01:19:29 PM »
On imaging-
I dig good imaging. It's cool to hear sound sources clearly laid out across the stereo field, even though I consider it a higher-order playback aspect that's not absolutely necessary.  Other things are more important to me to get correct before good imaging.  And often I can't really hear sharp imaging when at a concert.  The sound is big and enveloping, and there is a definite sense of sound coming from the stage contrasted against the sound bouncing around out in the room and the audience sound all around me, but sharpness of image isn't especially apparent.  Mostly I'm relying on vision to accurately pick out where the sound sources are located.  Good imaging is more important in playback, where it serves to reinforce the illusion.

That said, if the direct/reveberant balance is good and once the EQ balance has been dialed in (two of the vitally important things!) I'm relatively happy with the imaging I get from just a spaced omni pair in my live music recordings.  And I do think about the imaging effects of the spacing between omnis along with the other things like bass response and achieving low diffuse field correlation (more on that next).

We talk a lot about Michael William's Stereo Zoom approach to microphone setup here at TS, which mostly deals with the imaging aspects of microphone setups and how to modify them.  Wiliams' empirically derived Stereo Zoom tables are mostly applicable to directional microphones, but it loosely applies to omnis as well.  I'll sometimes refer to the Stereo Zoom charts when thinking about what omni spacing I want to use. The basic relationship is well understood by most here- with regards to imaging, a wider microphone spacing has a narrower Stereo Recording Angle.  That is to say- the region of the stage from the center-line out to some angle to either side of that will be reproduced as points of sound between the two speakers (that's the Stereo Recording Angle or SRA), grows narrower as the mic spacing is made larger.  With narrow spacings, the SRA is very wide and everything in the room is reproduced between the two speakers.  With wider mic spacings more stuff is pushed outside of the region between the two speakers.  At the extreme, we get only the fully centered stuff appearing centrally between the speakers (including mono PA information, which is why super-wide omni splits work at big outdoor concerts without a massive hole) and most everything else is shoved over to one speaker or the other.  That's the proverbial spaced-omni hole-in-the-middle.

That narrowing of SRA with increased spacing can be somewhat confusing to comprehend, because by all other aspects except imaging pickup angle, most everything else sounds bigger and wider and more open as the spacing is increased.  I covered why that is for the bass region, and will get to that next for the mids and highs.  But first, more on imaging..
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (15)
  • Needs to get out more...
  • *****
  • Posts: 15698
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: Relationships between spaced omnis and center mic
« Reply #23 on: June 23, 2017, 01:49:45 PM »
Here's a complication with imaging- when we add another mic or mic pair in the center, we are introducing additional imaging relationships on top of the one which already exists from just the pair of omnis.  The SRA imaging relationship between the omnis doesn't go away, and we now also have additional imaging relationships between the omni on each side and the center mic.  If we use a coincident stereo pair in the center we also have that imaging relationship thrown into the mix.  If we were to put a near-coincident pair in the center we'd introduce even more imaging complications- we'd have the imaging relationship between the two omnis, the relationship between the near-spaced pair, the relationship between each omni and the directional mic on the same side, and the relationship between each omni and the directional mic on the opposite side.  That's 6 different relationships between 4 different microphone positions.  Things get complicated fast, and unless we're careful about minimizing conflicts we can easily end up with a mess of things making for not just confusing imaging but also numerous comb-filtering complications.  This is partly why I don't recommend using a near-spaced pair in the center.  A coincident pair in the center keeps the imaging relationship count at 4, just one more than using a single center mic, and the comb-filtering relationships at 3, the same as with a single center mic.

The easy way to avoid these problems is to space the omnis further.  Make the omni spacing wider so that the SRA imaging relationship between each omni and the center mic works on it's own.  Michael Williams extended the Stereo Zoom a couple decades back to multichannel array recording with more than two channels.  In doing so, he describes techniques used to seamlessly join the SRA between the left and center mic with the SRA between the center and right mic.  That's intended for 3 channel L/C/R playback, most commonly for 5 or 7 channel surround recording, but I find it also applies to 3 channels mixed down to 2-channel stereo as well.  If interested, I can link to his paper where he describes how moving one mic forward of the other shifts the SRA sideways.  That in combination with mic angle and spacing is what allows for seamless SRAs all the way around a surround array.  It has a direct correlation with how far forward the center mic might be positioned in front of the left/right pair.  All of his published stuff so far deals with microphones using identical pickup patterns all the way around the array, so one needs to sort of read between the lines to get the implications for a directional mic between omnis, but the general concept applies.  Supposedly he's working toward arrays using different mic patterns, I'd like to see that.
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (15)
  • Needs to get out more...
  • *****
  • Posts: 15698
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: Relationships between spaced omnis and center mic
« Reply #24 on: June 23, 2017, 06:05:09 PM »
Okay, here's a big one, one of the main reasons sufficiently spaced omnis sound sweet and why I use them whenever I can in the pursuit of recordings which aim to transport the listener back to the time and place of the recording-  they produce recordings with low diffuse-field correlation (DFC), which is highly desirable for the ambient/reverberant content in a recording.  Correlation is one measure of signal 'sameness' between channels. Basically, high correlation means the signals are very similar to each other and will image somewhere within the playback window in a somewhat monophonic-like way.  That's good for the direct sounds which we want to reproduce with sharp imaging between the speakers.  It's bad for reproducing the diffuse sound filling the recording space in a directionless way.  Low correlation means the information is reproduced diffusely, sounding wide, open, airy and much more natural.

Spaced omnis are excellent at diffuse-field ambience upon reproduction because of their wide spacing.  Near spaced omnis aren't.  Similarly near spaced or coinciding directional mics are not very good at at this unless the mics are significantly more directional than cardioid and pointed far enough apart, due to the omni component of the cardioid pattern combined with the limited or non-existent spacing.  Blumlien crossed bidirectionals has excellent low diffuse field correlation, and widely-angled hypercards are pretty good too.  Coincident cardioids are terrible at this and partly why many don't care much for X/Y cards at a distance.  Spacing cardioids apart a bit helps some, but most near-spaced configs are still to close together to achieve significantly low DFC. Back in 2008 I posted a thread here at TS investigating why recordings made with spaced omnis and coincident crossed bi-directional mics (Blumlein) in many ways sound quite similar to me, even though they are total polar opposites in terms of mic spacing, polar pattern, and imaging.  http://taperssection.com/index.php?topic=110861.0 Now I have a much clearer understanding as to why.

With Blumlein coincident crossed 8's the mics are coincident, yet have no omni component, and have minimal shared in-phase pickup pattern overlap over the entire sphere, even directly above and below, except for the narrow quadrant directly in front and behind where the sound is correlated (in front is the only place we we want it to be correlated).  With omnis, somewhat similarly the direct sound arriving at the microphone pair from the median plane produces a correlated signal by arriving at both microphones simultaneously, but from all other directions the sound is increasingly uncorrelated as arrival moves off the median plain, corresponding to the source's increasing off-center angle, the distance between microphones, and the wavelength of the frequency in question.  The diffuse sound in the room - that is, the directionless sound which effectively arrives from all directions more or less equally, is picked up with very low correlation overall as long as the spacing between mics is sufficient.  The greater the mic spacing, the lower the frequency down to which the signals will have a randomized phase response producing low diffuse field correlation.

Okay, that gets pretty technical.  But it's a huge deal (the elephant in the room if you will) and why I had quotes about the importance of low DFC from acousticians, audio researchers, and famous recording engineers in my TS signature for so many years.  Achieving low DFC is the magic sauce which makes the very highest-order stuff from in my list of what is important possible - that is achieving a sense of envelopment and immersion.
« Last Edit: June 23, 2017, 06:09:41 PM by Gutbucket »
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (15)
  • Needs to get out more...
  • *****
  • Posts: 15698
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: Relationships between spaced omnis and center mic
« Reply #25 on: June 23, 2017, 06:56:10 PM »
So following the above reasoning, I now ideally have quite widely spaced omnis which achieve low DFC, and since they are widely spaced (lets say 5'+ or so) they not only do that better, they also provide good stereo bass to a lower frequency, and their wide spacing produces an imaging hole-in-the-middle which sort of clears that area for the sharp imaging contribution of a coincident pair in the center.  Also the wider A-B omni spacing forces any nearby localized audience chatter to the outer edges of the playback image because those close-proximity noise sources tend to localize more strongly at one mic or the other rather than in both simply due to the difference in relative proximity to those nearby sources.  That doesn't make them go away, but it does make them far less distracting and helps to get them out of the way of the music.  Everything falls nicely into place, without conflicts.  This is why I ideally suggest turning the single center mic into a M/S pair or a PAS X/Y pair between more widely-spaced omnis as an ideal solution when and where one can get away with doing so.

Rocksuitcase- the optional rear-facing directional mic for gaining control over additional ambient content which you and KindMS have been using, spaced slightly rearward from the center mic, keeps DFC between channels low because it is as distant from the omnis as is practical, and is spaced apart from the forward facing center mic and pointing the opposite direction.  Yet at the same time it is not far enough away from the plane in which all the mics are arranged to produce a significant delay to the direct sound component from the PA or stage. 


Okay, enough for today.  Since most of what I've posted consists of arguments for wider omni spacings than most can easily achieve, next week I'll post more about about what one might do when unfortunately the omnis can't be spread wide enough.
 Thanks for the soapbox, I hope all this doesn't come across as pedantic.  I very much enjoy sharing what I've learned and am really thrilled to find some of these ideas being taken up by the larger taping community over the past couple years.  I never really expected others to try much of this stuff- it's just plain weird compared to traditional 2-microphone recording setups at first glance, even though the conceptual basis is sound.  But it's fun and provides powerful tools unavailable to traditional recording, which apply uniquely to the oddball application of recording live music from a distance.  We're really the only folks doing that in the recording world, and unique solutions can provide unique and welcome advantages.

Apologies for commandeering the thread thus far..
TL:DR!
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (15)
  • Needs to get out more...
  • *****
  • Posts: 15698
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: Relationships between spaced omnis and center mic
« Reply #26 on: July 07, 2017, 04:12:05 PM »
Back for more-

Following what I outlined above, when adding a center microphone or pair of mics between a pair of omnidirectional microphones,  the omnis ideally need to be spaced far enough apart - farther than they normally would be if used as a pair on their own. So, what to do when you can't or otherwise space a pair of omnis widely enough? Use some method to generate additional level difference between the two omnis:

1) Angle the omnis further apart.  This works at high frequencies only, down to a frequency determined by the size of the microphone capsule housing.  Since both the limited omni spacing and the introduction of a third microphone or pair in the center do much to eliminate any "hole-in-the-middle" problem, go ahead and maximize the limited effect by pointing the omnis directly to either side, 180-degrees apart from each other rather than directly ahead or partly angled outwards.  This is easy to do and effective, even if the frequency range in which it is effective is quite limited, making it the first and most obvious thing to do.  In addition, the center mic or pair is handling forward detail and presence, so don't worry about pointing the omnis so far away from the "main source of sound up front". Microphone attachments like spheres, disks, and angled plates for boundary mounted microphones can lower that frequency somewhat into the upper-midrange/low-treble region, significantly enough that their use is usually audible, but don't typically lower the frequency at which the microphones become directional a whole lot further so as to strongly effect stereo imaging.  They mostly affect frequency balance and apparent spatial spectral balance, which may or may not be desirable for other reasons. 

2) Place a baffle of some sort between the omnis.  This is the dummy-head/Jecklin-disk thing. The baffle shadows the microphones from sounds arriving from the opposite side, increasing level differences.  Like microphone attachments this only works above a certain frequency point, which is determined by the size of the baffle.  Jorg Jecklin experimented with a three microphone version of his system using a somewhat wider omnis spacing and a bigger diameter disk with a notch cut into it to fit a forward facing center cardioid.  Problem is that baffles which are large enough to be optimal are unwieldy and not conducive to audience taping.  But if you can't get enough spacing between your omnis, have a Jecklin disk or something similar which will work as a baffle on hand, and can rig it without problems, a baffle can help significantly.

3) Use of a Blumlein Shuffler circuit to filter the output of the omni pair.  This is a circuit which converts low frequency phase differences to level differences.  Alan Blumlein used this with a pair of close spaced baffled omnidirectional microphones for his initial stereo experiments in the early 1930s before quality bi-directional microphones became available to his EMI development team, and it's use is described in his original patent.   Jecklin's later ideas about using a baffle between omnis is a direct decedent of this technique, and he also experimented with Blumlein Shuffling.  I've never tried it, but am quite intrigued by it as it could potentially be a very valuable tool for audience taping using omnis which can't otherwise be spaced far enough apart, and may also apply to recordings made using near-spaced directional mics, allowing them to sound more spatially enveloping at low frequencies than they otherwise do.   In terms of the direct sound (not the ambient sound), by choosing a particular spacing distance between mics we are selecting a particular frequency range in which we want to optimize stereo imaging, sacrificing imaging performance at the other end of the range.  Wide spacing is more optimized for low frequency stereo imaging.  Narrow spacing is more optimized for higher frequency range stereo qualities.  One way of looking at it is that Blumlein Shuffling is a way of correcting for insufficient spacing at low frequencies.  This is related but not quite the same as converting from Left/Right to Mid/Side stereo and boosting the low frequency Side information with a corresponding cut to the low frequency Mid information- a technique I do use frequently.  For more information on this search out Blumlein's original work and Micheal Gerzon's extension of it 40 years later.  Also here is an outfit building hardware boxes which do several types of Shuffling including the original Blumlein variety intended for narrow spaced omnis- http://www.phaedrus-audio.com/shuphler.htm.  Partway down that page are explanations of the different shuffling circuits and what they are intended for.  I've not heard them, but that site posts listening examples here- http://www.phaedrus-audio.com/intro_to_shuphlers.htm.   Lest costly and probably more applicable to processing our recordings is the "in-the-box" version of those same filters, which are available for Apple platforms here- http://pspatialaudio.com/index.htm. With a further description of Blumlein difference techique here- http://pspatialaudio.com/blumlein_delta.htm.  In addition to being more affordable (I presume) and eliminating the need for a dedicated hardware box incorporating analog circuitry, the digital filters take advantage of non-causal time-symmetric filtering which supposedly correct for some inherent deficits of analog filtering which can only be realized via digital processing.  Unfortunately it doesn't seem to be available for Windows platforms.  If anyone has experience with Blumlein Shuffling or use of other shuffling filters other than simply boosting the LF of the Side channel, please let me know.
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Offline heathen

  • Trade Count: (23)
  • Needs to get out more...
  • *****
  • Posts: 3528
Re: Relationships between spaced omnis and center mic
« Reply #27 on: July 07, 2017, 04:53:59 PM »
Wide spacing is more optimized for low frequency stereo imaging.  Narrow spacing is more optimized for higher frequency range stereo qualities. 
Does this only apply to the Blumlein Shuffler you discussed, or to any spaced stereo pair of omnis?  Also, if forced to choose, wouldn't one generally prefer to have the stereo imaging in the higher frequencies, since we can't really hear direction of the really low frequencies?
Mics: AT4050ST | AT4031 | AT853 (C/SC) | Line Audio CM3 | Sennheiser e614 | Sennheiser MKE2 | DPA 4061 Pre: CA9200 Decks: Zoom F8 | Roland R-05

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (15)
  • Needs to get out more...
  • *****
  • Posts: 15698
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: Relationships between spaced omnis and center mic
« Reply #28 on: July 07, 2017, 07:02:08 PM »
^ It applies to some problems inherent with stereo reproduction in general.   Blumlein recognized that early on with regards to stereo recording using omnis.  EMI recognized it when stereo came of age in the late 50's and built somewhat different shuffler circuits into their first stereo mixing desks.  The  Landmark 1957 article by Clark, Dutton and Vanderlyn which marked the start of the stereo era is an outstanding reference on the basis of two channel stereo- how it works, hearing, recording and reproduction.  Pretty amazing that it more or less comprehensively covers many of the things we still discuss over half a century later.  I rank it up there with the original Blumlein patent for both historical and academic value as well as breadth of coverage.  The discussion in the appendix is worth checking out as well.  You can find it here- http://www.phaedrus-audio.com/stereosonic.pdf

The basic issue is that low frequency and high frequency information image differently in stereo for numerous reasons.  If you were to listen while panning a monophonic low-passed source and a monophonic high-passed source to what sounds like the same mid-left or mid-right position, and then look down at the pan pots, the two would be set differently.  Usually you'll find that you'll pan the low-passed source further to the side than the high-passed source to achieve the same apparent position.  That's actually an argument for using microphone configurations or mixer panning laws which slightly over-emphasize low frequency width, rather than an argument for less stereo width in the bass than the highs.  Much of that imaging problem correction stuff ended up ignored and forgotten by the time stereo really took hold, partly due to the tools available to do so at the time were not as advanced as now (the analog vs digital filtering thing) and apparently caused coloration problems which were more problematic than the imaging problems they were designed to correct.  And that's the correct value judgement to my way of thinking. 

But remember that this is about a lot more than just imaging.  And although stereo imaging isn't my top priority, I still want as good imaging as I can get.  If you forced me to choose a single frequency range for good imaging, I'd pick the 700Hz to 3kHz range where our hearing is most attuned to imaging aspects.  If forced to choose between monophonic bass up to 500Hz or monophonic highs above 5kHz I'd have to do a listening test to confirm, but I suspect I'd choose stereo bass and monophonic highs.  Good stereo bass information is a big part of conveying the "you are there" enveloping experience of live music.  But I don't want to have to choose, I want good imaging across as wide a frequency range as possible.  And these multi-microphone techniques are ways I can achieve that.  Engineering is all about juggling and optimizing tradeoffs to maximize what you want while minimizing what you don't.
« Last Edit: July 07, 2017, 07:21:35 PM by Gutbucket »
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (15)
  • Needs to get out more...
  • *****
  • Posts: 15698
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: Relationships between spaced omnis and center mic
« Reply #29 on: July 07, 2017, 07:04:32 PM »
Okay, one more post and I'll call it quits as I'm sure most of this is way deeper than most here care to go.  But happy to discuss if anyone wants to do so. 

My previous post covers things you can do if you can't space a pair of omnis as far as you'd like, and relates to either straight 2-omni recording or using an additional microphone or pair in the center between omnis.  But what about a situation where you have the omnis spaced exactly as far apart as you want them for recording using the omni pair alone, but then want to add a center microphone to that? 

As mentioned previously, in most cases we want to greatly increase the omni spacing when introducing a third mic in the center- often doubling the spacing in general simplified terms.  There are several reasons for that- reducing  combfiltering from three mics in close proximity, improving the imaging across three mics instead of two, etc.  But there is another way of getting the spacing we need without moving the omnis.  We can space the center mic sufficiently far away from the omni pair.  In other words, we can move the center mic away from the omnis instead of increasing the spacing between the Left and Right omnis, in order to get the omnis far enough away from the center mic.

One way to do that is to mount the center mic directly between the omnis as before, but put it much higher or much lower than the omnis.   Imagine a tall 15' mic stand with the forward facing center mic at the very top and the spaced omnis on a bar clamped to the same stand something like 7' or 8' up.  Together, the three microphone locations form an upward pointing triangle.  That arrangement keeps the timing relationship more or less unchanged for direct sounds arriving from the horizontal plane.  Sound from directly in front reaches all three mics with about the same timing relationships as if all three were in a single line.  But the diffuse reverberant sound arriving from all directions has a significant time of arrival difference between the center mic and omnis because the arrival time difference serves to decorrelate the signal between the center and omnis, reducing comb-filtering issues and providing some of that wide-spaced omni openness and ambient sheen.  I recently suggested this as one option which avoids having to do any post filtering in Edtyre's thread- Suggestions for a center channel omni to run into my MixPre3. Basically it's the same idea applied there, even though we were mostly talking about adding a single omni to a standard near-spaced directional mic configuration rather than a center directional mic to a spaced omni pair.  Alternately one can simply mount all three mics relatively close together and low pass the omni channel so that it only provides low bass below where the cardioids begin to roll off, extending the low bass response without comb filtering and imaging problems since it will not interact with the near-spaced cardioids at higher frequencies.

Another way to physically space the center mic away from the omnis is to leave all the mics in the horizontal plane, but move the center microphone far enough forward of the omni pair to sufficiently decorrelate it from them.  To keep the imaging correct we need to introduce a compensating delay the center mic so that it's signal will be time-aligned with the omnis for sounds arriving from in front (approximately 1.1ms for each foot of distance).  Sounds from other directions will arrive with different timing relationships between the three mics.  Sound from in back will arrive with a delay equal to that of twice the forward spacing of the center mic (corresponding to it's distance in front + the delay compensating for that same distance for sound arriving from the opposite direction).  The obvious problem with this is the setup- you probably need a separate mic stand for the forward center mic.

That can actually be done in such a way that the SRA imaging between each omni and the center mic is seamless and can be made to match that of the two omnis alone, eliminating the inherent imaging conflicts which otherwise would occur between three mics all pointed in the same direction.  Michael Williams extends the Stereo Zoom (2channel) to Mulitchannel Microphone Array Design (3 to 7 channels) to special multi-channel arrays which can be used to record both simultaneously which he dubs "Magic Arrays".  His AES papers are available at his website- http://www.mmad.info/.  Below is a photo of an 8-channel Magic Array from Jerry Bruck's Posthorn site.  Obviously that's a specialty multi-channel recording array, but if you were to eliminate all but the 3 forward-most microphones, notice that the center mic pair are more or less in a NOS or DIN configuration with the center microphone moved forward by a significant distance.

Edit to insert this note of clarification on the photo below- These may appear to be wide-Left/Right-spaced omnis plus six other mics but are not.   A 3-mic main output like what we are discussing as produced from this array would consist of just the forward most extended mic and the two angled mics closest to it.  In the photo those microphones are the three closest to the top of the frame (center mic top-most + the Left/Right pair immediately beneath it).  Please ignore the other five microphones for the sake of this discussion.  The diagrams below are more clear on this- the center microphone being well forward of the left/right pair (closer to the top of the diagram). With respect to the diagrams, simply ignore the two rear facing (bottom most) microphones.



Below are William's setup examples for a few different 5-channel Magic Array using omnis, cardioids, and supercarioids.  Each array can be used to provide mono (center forward facing mic only), 2-channel stereo (90degrees/24.5cm forward facing pair only), quad-stereo (all four central mics in an IRT-cross-like configuration), or 5-channel audio at the same time.  Although not intended this way I also find that of the MMAD arrays I've used, they all seem to fold down to 2-channel stereo really well. I extracted the setup diagram images below from this paper on his site- Magic Arrays - Multichannel Microphone Array Design applied to Multi-format Compatibility 

As you look at these examples, simply ignore the two rear-facing microphones and you have a 3-microphone array.  The first example below is then a spaced omni pair arranged to achieve a 90 degree SRA, with a completely optional center omni added to it which in imaging terms at least isn't going to mess up the main 2-channel omni pair whether you choose to use much of the center mic or not.  Substitute a cardioid for the center omni and you have very taper optimized spaced omni pair with a directional center mic.  Notice that in the examples using cardioids and supercardioids, the left/right pairs are pretty much in basic microphone configurations everyone here is familiar with- close to NOS and ORTF for the cardioids, and 110 degree X/Y for the supercards.

This sort of brings my thread-jack full circle, with reasoning for why one might want to put a center mic far forward of a pair of omnis- It allows you to use standard stereo setups for the left/right pair.  Another way of thinking about a far forward center mic is as a sort of as a spot mic, or like matrixing a SBD feed with your AUD pair.  In booth of those cases you often need to delay the spot or SBD to align it with the AUD source, and after doing that everything meshes nicely without comb filtering conflicts because the spot or SBD are placed far enough away from the main AUD pair.
« Last Edit: July 13, 2017, 12:54:09 PM by Gutbucket »
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

 

RSS | Mobile
Page created in 0.102 seconds with 44 queries.
© 2002-2024 Taperssection.com
Powered by SMF