Topic: Oddball microphone technique (OMT) - part 4

[mrfender]

"...young man furiously playing polyrhythms on a triangle off to the left..."   

[capnhook]

Ok, insignificant then




Carry on, make great tapes

[rocksuitcase]

The answer is still no.  It can be simulated, but it's not the same.  Whether that matters to you is up to you.

IMO- It is quite obvious that capnhook is trolling in this thread. I am aware there are a few on here who think that some people's knowledge is worth less than others. Apparently THOSE PEOPLE go by how many recordings one has posted or shared as part of their knowledge of physics and acoustics. Silly, bullying games - IMO.   

[Gutbucket]

Grawk nails it, with admirable brevity.

It's not insignificant. It's a complex relationship, and a usefully comprehensive answer can't be reduced to a simple yes/no answer.. except for the fact that the two things aren't equivalent.. in most cases.  Actually, there is one very-narrow, specific case where they are fully equivalent, which equates to recording a monophonic source, centered directly in front of the recording arrangement in an anechoic chamber.  My assumption is Unclehoolio isn't doing that!

[rocksuitcase]

Grawk nails it, with admirable brevity.

It's not insignificant. It's a complex relationship, and a usefully comprehensive answer can't be reduced to a simple yes/no answer.. except for the fact that the two things aren't equivalent.. in most cases.  Actually, there is one very-narrow, specific case where they are fully equivalent, which equates to recording a monophonic source, centered directly in front of the recording arrangement in an anechoic chamber.  My assumption is Unclehoolio isn't doing that!

This is how we tested loudspeakers for the Carrier Dome installation at SU during the construction of the Dome. We used the anechoic chamber usually used by the jet engine/turbine guys to measure them and decide which ones we preferred for the various arrays distributed through the building. B&K measurement mic 1 meter from the source into a B&K dual channel FFT analyzer (which compared the noise source to the speaker output).

[grawk]

This is a forum for hobbyists. Lots of things matter, and people pick and choose based on their goals and desires. Every choice is a compromise of some sort.

[rocksuitcase]

This is a forum for hobbyists. Lots of things matter, and people pick and choose based on their goals and desires. Every choice is a compromise of some sort.

correctumundo. My "compromise" here is that I chose to call out some BS rather than ignore it as I typically do. Boo hoo.     

[grawk]

I didn’t get that you were criticizing mic placement when you said you tested loud speakers that way. I thought you were just dropping an interesting fact. What’s the argument that says that test applies to real world 3d microphone techniques?

Conflated a couple of posts and missed where you pointed out the obvious trolling. My bad.

[Gutbucket]

I take it as a cool factoid in that the special case described parallels standard anechoic testing methodology. 

Also, I share an SU connection with rock in having attended SU after he was there.  Carrier Dome had been in operation for a few years at that point.  I never looked into the rumored to be well-paying student job of shoveling snow off the roof (I doubt that was actually a student job thing), but did get paid to usher concerts in the Dome and found I could make additional cash by sticking around to help pack up roadshow gear afterward. Fond memories of helping to pack the inflatable Floyd pig in it's case and disassemble the airplane.

[capnhook]

I have tried the technique of moving an XY pair a few inches ahead of a split omni pair


Not aware of which OMT configuration that one is


I heard no difference or improvement in the stereo image

Actually it sounds "smeared" to me when sources aren't time-aligned



I have heard no examples of Lee's for me to change my mind

I think most of this theory resides in peoples' heads, and not in reality


YMMV

[checht]

^
Same empirical experience as the cap'n. Moving the x/y center pair forward seemed to add phase smear, but could have been an artifact of my expectations.

Now I run OMT4 with the mics all in a line.

[Gutbucket]

I don't hear much difference, particularly when the wider pair are omnis.. but I do hear it somewhat when the pair on either side of the center position are directional mics forming something more akin to a near-spaced stereo triplet.  I also hear it more when playing back using a center speaker, routing the center channel mic directly to that  in an L/C/R speaker arrangement).  Sometimes shifting the center forward just a bit seems to solidify and locks down the center of the stereo image a bit more, maybe because a wave front from directly ahead pings that mic a fraction of a second first, but that's speculation and could just be my imagination.

I'm sure I've written about this previously, but after wondering about and exploring center mic position a bit more last year, here's how I ended up with the current OMT suggestion of placing the center position around 15-20cm / ~6-8" forward in most situations.  It's mostly a best compromise solution juggling a few different things, a way of doing minimal harm in the absence of complete data about which thing is most important, a suggestion that I've found won't cause more issues than it attempts to solve.

I see two primary variables that we are playing with here.  One involves attempting to optimize for the coherent arrival of a planar wavefront pinging at all mics at the same instant.  The other involves optimizing the smooth handoff of imaging sectors across adjacent microphones in the array.  Ideally we want both.  If either is sufficiently wrong we are more likely to get "smearing" of some sort or another.

The 15-20cm spacing suggestion is something of a compromise between achieving the most coherent arrival across multiple microphones for a wavefront arriving from directly ahead (best achieved with all mics in a horizontal line), verses wavefronts arriving from a slight angle, such as from each PA speaker.   Can't have both at the same time.  For a while I schemed about possibly spacing the center mic forward in such a way that a line between the center mic and left mic would be perpendicular to wavefront arrival from the left PA speaker.  And a line between the center mic and right mic would be perpendicular to wavefront arrival from the right PA speaker.  That angle is going to change with recording position, just like PAS angle does, but 15-20cm is reasonably close to achieving something like that in many cases.

Secondarily, OMT is strongly influenced by Michael William's arrays (Stereo Zoom and MMAD or Multi Microphone Array Design) and Gunter Thiele's OCT arrays.  Those arrays are designed to achieving very precise image placement and smooth image hand-off between adjacent mic-pair sectors of in the array.  OMT is not as seriously concerned with "distortion free" imaging.  Clear, distinct, believable and engaging imaging is its goal.  It's more specifically attuned to PA recording, and values achieving better impulse coherence for a wavefront arriving from directly ahead, in addition to and even more so than, sharp and clear if not totally accurate imaging.  Still, sharp and clear imaging is a nice to have thing and one of the goals.

But with imaging concerns in mind, it quickly becomes apparent that arranging all mics in a line in unable to achieve accurate image linking between adjacent mic pairs along that line.  Any pair of microphones regardless of the spacing between the two will have some Stereo Recording Angle that is greater than 0-degrees.  Yes the wider apart they are the narrower that SRA becomes, but it is always "splayed out" at some angle wider than 0-degrees.  So the imaging from adjacent linear segments is always going to overlap each other with some image smear.  The only way to correct for that is to angle the segments somewhat relative to each other.  Pushing the center mic forward a bit does that.  The edges of the two adjacent segments that each have an SRA of some angle come closer to handing off from one to the other without overlap.  If you play around with the Schoeps Image Assistant microphone-array visualizer, you will note that to achieve good image linking with any 3 mic array requires pushing the center mic forward.

So I compromised and made it a bit of a Goldilocks thing - seeking the middle way - not going fully one way or the other in the attempt to maximize collective benefit and minimize the potential for harm.  15-20cm seems to do that. 

If you can't achieve that, or choose not to for whatever reason, no problem.  Do your own thing.  These are only guidelines that are deeply considered and have worked well for me.  I hope they work well for others as well and very much enjoy our discussions about all this stuff here.  Thanks for the bandwith.

Respect to all you folks, including those who feel this is just a bunch of hooha.

[Gutbucket]

Taking all this full circle, another reason for pushing the center mic forward a bit is that if afterward one decides they want to emulate having all the mics arranged in a line so as to get the tightest possible alignment for a coherent wave front arriving from directly in front, you can delay the center channel by a fraction of a ms to achieve that.

That may sound at first like I'm contradicting what I said earlier, but I'm not.  Delay does not achieve the same thing as physical spacing.  However, in this case we are flipping it around and going the opposite way.  That allows the way in which the delay effects arrivals from other directions to work more in our favor rather than against us.  Delay applied to a forward positioned center channel brings the wavefronts from the on-stage drummer and accordion player directly in front of the recording position into full alignment across all three mic channels.. and at the same time it further delays the sound of the penny-whistle players behind the recording position to the center mic channel, rather than shifting them forward in time in that channel and confusing things as it did in the previous example.  It helps to keep the sounds arriving from behind perceptually in the back, and shift them out from the center toward the sides.  The application of delay in this case works toward what we want to achieve rather than against it.

In other words, we can compensate for a forward positioned center microphone by applying delay to that channel in a way that works much better than attempting the opposite of trying to emulate a forward positioned mic position by delaying all the other channels of an array that had all mics arranged in a horizontal line.


In the more detail than you probably want category, here are two examples of arrays that require the combination of positioning AND delay in combination to achieve the desired result-

Gunter Thiele's OCT arrangement later incorporated a special optional modification intended to best optimize it for two channel stereo playback rather than of L/C/R speaker playback.  It's called OCT2 and the change is that it pushes the center microphone position much farther forward than the 8cm / 3" standard center forward positioning of OCT, and then requires the application of delay to the center channel to bring it back into alignment again (with regard to a wavefront arriving directly from the front).  That change achieves improved diffuse field decorrelation that makes for a better, more open and diffuse-sounding impression of the room impression and reverberant decay, and reportedly improves phantom center stereo imaging a bit.

Similarly Michael Williams special 8-channel Magic arrays use an even farther forward positioned center channels (something like 2 meters or so foreword I think) along with a required application of delay to that channel to bring the imaging sectors into alignment.  Interestingly, that extreme forward spacing of the center mic position allows the flanking near-spaced mic pair to remain essentially unchanged from what would be optimal for 2-channel recording.  That pair no longer needs to be spaced something like twice as wide horizontally as it otherwise would when incorporating the center channel mic.  Ironically this would imply it would be the best option for tapers who wish to try adding a center mic to their preferred near-spaced two-channel stereo configuration, since the two channel config remains the same as before, however the center mic position needs to be positioned WAY foreword and delayed. And unfortunately that's totally impractical for tapers.  It's far more challenging to space one mic position forward by something like 2 meters than it is to space a line of mics out horizontally buy a couple meters, either with or without a little forward spacing of the center mic position.

Part of the game is juggling compromises to find a comfortable limit of what can be practically achieved, and we all have to draw the line somewhere.  The ongoing evolution of OMT reflects the pursuit of optimal solutions within the framework of what has been practically possible for me, along with some projection of where it might be able to go further.  Take from it whatever is useful for you.

[capnhook]

Thanks for the bandwidth, also



I'm a tech guy used to fixing engineering mistakes, and I look for practical solutions




Still waiting to hear some of Lee's magic

[rocksuitcase]

Thanks for the bandwidth, also



I'm a tech guy used to fixing engineering mistakes, and I look for practical solutions




Still waiting to hear some of Lee's magic

I am a tech guy with a degree in acoustics who has earned US Patents on audio signal processors.
I do not need to hear anything from Lee to know these theories have merit.
In fact, since I own microphones and record live shows all the time I test out these things for myself.
I know what FFT analysis is and can reasonably deduce the multi channels of signals with my ears and then use analysis tools. Therefore I am certain of both the benefits and limitations of these applications.

It truly isn't rocket science although there is math involved. bandwidth used