Topic: Spaced Omnis - two schools of thought?

[Gutbucket]

I found this paragraph from DPA especially interesting, as it is the only explanation I've found for keeping the spacing relatively close (emphasis mine):

Since the stereo width of a recording is frequency-dependent, the deeper the tonal qualities you wish to reproduce in stereo, the wider your microphone spacing should be. Using a recommended microphone spacing of a quarter of the wavelength of the deepest tone, and taking into account the human ear's reduced ability to localize frequencies below 150Hz, leads to an optimal microphone spacing of between 40 and 60 cm. Smaller microphone spacings are often used close to sound-sources to prevent the sound image of a particular musical instrument from becoming "too wide" and unnatural. Spacings down to 17 to 20 cm are detectable by the human ear, as this distance is equivalent to the distance between the two ears themselves.

I'm not sure I'm onboard with the justification based on "taking into account the human ear's reduced ability to localize frequencies below 150Hz", party because I feel an appropriate spacing has as much to do with tone as it does with directional imaging.  That seems an attempt to combine to separate realms, a tonal one based on the quarter wavelength constraint, and an imaging one based on human directional listening acuity.

So another thing about omni spacing, which has perhaps as much to do with the overall sound of the configuration as it does with its "stereoness" is the effect of spacing on what I'll call the "contour" or "tone" of the low frequency response.  Shifting omni spacing while listening at the same time,  a change in bass "contour" is one of the primary things I hear, along with the sort of increasingly open "ambience" and "air" as spacing is increased.  I suppose it's the first ripples of the comb-filter cancellation/reinforcement shifting around, with the mic spacing playing against the frequency wavelengths in question.  The approximate 1 meter / 3ft spacing I use as a baseline starting point for a pair of omnis alone was arrived at primarily by listening to the tone and contour of the bass along with the "stereo openness" while listening on headphones and adjusting the spacing.  Although the spacing adjustment was also affecting SRA width, I consider that 'tuning' to have been more like making a sort of EQ adjustment, which I suspect would carry over to other scenarios more so than if I was simply tuning the SRA to adjust image width.  In that situation, things changed significantly and rapidly when shifting between narrow spacings of up to a couple feet, and seemed to open up and settle down frequency-wise at spacings of more than that but less than 5 feet or so.  There was plenty of bass information playing while doing that (it was in an amphitheater with a couple different band playing), so it may be more difficult to do by ear while adjusting spacing for a marching band.

[Bruce Watson]

A very wide omni spacing can achieve that aspect of "stereoness" and envelopment down low.  A wide spacing also decorrelates the hall ambience and reverberance down to a lower frequency, which makes things sound more "open and airy" and reproduces those sounds diffusely, throughout and even outside of the playback imaging space. That's a different kind of stereo thing than left/right imaging and doesn't relate to Williams SRA curves. It's not about accurate left/right imaging directionality or orchestra angle.

I'm thinking that this is the point in using an array such as the Faulkner / Boojum / JNorman, with outriggers. The former for the upper bass and higher, the latter for the low bass and to decorrelate the hall ambience and reflections. This is why, perhaps, Decca always used outriggers with the many variations of the Decca Tree. Maybe it's part of the reason anyway, IDK.

[voltronic]

Consider that the appropriate SRA for recording a marching band may or may not be equal to the Orchestra Angle into which the outer edges of the band's marching movements will fit.  I suspect a narrower SRA will be better, zooming into the action closer to center field a bit.  The outer limit edge of field stuff will then simply image over at one speaker or the other, with more band movement between speakers.  I suspect an SRA wide enough to cover the entire field would provide a more distant perspective with less dramatic, more centralized.

How much of the field do they actually use?

Do you have a way of recording three or four channels?  Would be interesting to do a 10 or 20ft omni split atop the press box with a single directional mic in the middle pointing at center field.. or a stereo pair in the middle instead of the single if you like (which also hedges the bet).

They generally use the area between the two 20-yard lines, so your suggestion of a narrower SRA makes sense.  Playing around with the Sengpiel tool, if I assume an orchestra angle of 60deg, it looks like somewhere between 60cm and 1m may be the ticket here.

I could do 3 or 4 channels, but at the present I only have 2 windscreens up to the task and 2 superclamps.  It's something I might consider down the road though.

[2manyrocks]

You could gaft tape your omnis to the rail and use some fuzzy fabric from the fabric store as a windscreen freeing up your clamps for a center card.  Or you can buy a $2.00 clamp from walmart or TSC if you have one and drill a 3/8 hole in it for a 3/8 bolt to make a useable mic clamp.  Or get the school welding shop to weld a 3/8 bolt on the clamp. 

I'd have reservations about using only omnis, but of course will be interested in your results.  How much boundary effect you get might depend on the distance from the railing to the wall of the press box. 

 

 

[voltronic]

You could gaft tape your omnis to the rail and use some fuzzy fabric from the fabric store as a windscreen freeing up your clamps for a center card.  Or you can buy a $2.00 clamp from walmart or TSC if you have one and drill a 3/8 hole in it for a 3/8 bolt to make a useable mic clamp.  Or get the school welding shop to weld a 3/8 bolt on the clamp. 

I'd have reservations about using only omnis, but of course will be interested in your results.  How much boundary effect you get might depend on the distance from the railing to the wall of the press box. 

 

 

Before I go into more channels, I really do want to try the omnis on their own at field level and also up on top of the press box.  My goal is also to keep this setup small and quickly deployed from my bag.  4 mics makes that more difficult.  I have two 6ft, rugged cables Darktrain made me for this purpose, and all my other cables are 25' snakes pairs which is just silly for this application.  I know that I can get good results from my CM3s up close; less way back on top of the box.  I want to experiment to see what I can get out of just the pair of omnis alone at different distances and widths before I decide to add anything else in.  These are all good ideas though, and I appreciate it.

The railing is on the roof of the press box, so no worries about boundary effect - there's nothing behind me but empty space.  In a couple weeks when I get up there, I'll try to take some pictures, and if I get a good capture I'll post it here on the marching band thread I started last year.

[MIQ]

Hi Voltronic,

I have been following this thread but haven't had a chance to chime in yet.  I wanted to add a little more info on the combing due to spaced mics.  I was looking at this last month and made a spreadsheet to study it.  Attached are some screen captures of the sheet I created that will illustrate the basic trade-offs. 

The fundamental idea is that the delay between to the two mic signals determines the freqs where the notches appear.  Two factors determine this delay, the distance between the mics and the angle of the sound source with respect to the microphones.  At any spacing, if the sound source is directly in front of the two mics (Theta = 90 degrees on my sheet), the arrival of the sound source at each of the mics will be the same and there will be no delay.  As you move further and further to one side or the other (off axis) the delay will get bigger and bigger since the sound will be arriving at one mic sooner than the other.  Obviously, the worst case angle is directly off to one side (Theta = 0 or 180 on my sheet).  As this delay gets bigger and bigger, the freq where the first notch occurs gets lower and lower in freq.  It is important to realize that this combing notch is at a different freq for all angles in front of the mics.  As Gutbucket mentioned, depending on where the low freq sounds are in relationship to the mics, there will be differences in the low freq contour created by the first notch.  Playing with the spacing allows you to change this bass contour, with larger spacing pushing this first notch lower and lower in freq.  Please take a look at the attached pics that illustrate these effects. 

When I created the sheet I limited the number of freq points where I graph the notches to keep the file size and calculations to a reasonable size.  Since I am not including every possibe freq in the chart, and the notches themselves are so narrow, the graph doesn't always "catch" the true minima of each notch.  The graph shows the basic location of the notches but may not give an accurate "depth" to each notch.  All the nothces should theoretically have the same amount of cut but my chart will not always show that.  It is an artifact of limiting the number of different freqs.

The attached show that as the angle of the sound source moves more off axis the freq of the first notch goes down in freq.

[MIQ]

In this set of pics I kept the sound source angle at 30 degrees and changed the spacing between mics.  Here you can see that the larger mic spacing will increase the delay between mics and again push the first notch down in freq. 

I started at 17cm (ear spacing) and went up to 70cm (beginning of "hole in the middle" imaging in many set ups). 

As others have noted, very wide spacings (3-5 ft) are often used not so much for accurate imaging location but to create very uncorrelated L/R signals even at very low freqs.  These wide spacing will push these notches even lower in freq.  Not that that is a huge deal but you may need to dial in the spacing so the notches do not lead to problems with the angles that the bass sound sources arrive from.

As far as the SRA and the visualizers (including the Stereo Mic Tools app I helped desing), these are all relying on the human psychoacoustic perception of sound localization.  This localization is based on the time and level differences that occur at your ears.  Since the signals recorded by both Omni mics will be the same level regardless of the sound source angle, all the imaging is created by the time differences.  Depending on who's psychoacoustic data you choose to work with (Williams/Simonsen, Sengpiel, Wittek/Theile, etc.) you will arrive at different SRAs based on the mic spacing.  Stereo Mic tools uses the Williams/Simonsen data for spaced mics to calculate the SRA but uses the Wittek/Theile data for calculating the RA_75%.  As has already been discussed, these SRA calclations will work for small mic spacing where the delays created fall within the narrow range used by our brain for determining source location but will fall apart at very large mic spacings.  Again, not a big deal since very wide spacings are used to create very uncorrelated signals even at low freq giving a sense of a very large ambient sound space, not neccessarily an accurate sound location.   

Hope this helps give a little clearer pic of signal delay created comb filtering. 

MIQ

[Gutbucket]

A couple important notes-

The comb-filtering seen above is what happens when the two signals are combined electrically.  In other words, mixed together to form a single monophonic signal.  This is the basis upon which 2-channel stereo recorded using spaced omnis is considered to be less than optimally mono-compatible.  Of course, in reality a recording made using spaced omnis is perfectly mono compatible by simply eliminating one channel and only using only the other-  "wa-la" we then have a single monophonic omnidirectional channel with zero comb-filtering, from a single microphone placed just slightly off the original center-line.

That clear and distinct comb filtering does not manifest in the same way when the recording is played back normally in 2-channel stereo.  It will not manifest at all with headphone listening since the signals don't mix at all until they are in the listener's brain.  With loudspeaker playback, the two signals mix only upon reaching the listeners ear(s), the sharp nulls go away, and the comb-interaction will be different at each place in the 3-dimensional space of the listening room, as well as being different for each frequency in question.  Regardless of the specific location of the listener's ears in the room relative to the speakers, there will be no sharply defined comb cancellation notches as seen in the graphs above, that only happens when the two signals are mixed electrically.  At anything above the first notch (which shifts around by both frequency and position as noted) the combing interactions become so complex they essentially become random.


There are going to be level differences in most widely spaced omni setups.  Any source which is close enough to the mics and not on the centerline is going to produce level differences as well as arrival time differences in the two channels. 

For spaced omnis out in the audience, the level differences will be mostly nearby audience sounds, and one reason I like wide-spaced omnis of 5' or 6' is that the sounds from nearby audience members will not be recorded with the same level in each channel unless they are directly in front or in back of the mics.  Because of that, those "unfortunately too close and unwanted" sounds image off to one side or the other, "getting out of the way" of the music imaging across the center between the two speakers.  It's still there, but becomes less distracting and offensive, easier to mentally separate from the music and ignore.

For wide-spaced omnis placed at the stage-lip or on-stage, the increased proximity of the various sound sources on stage to one or the other of the two microphones will produce level differences in addition to time of arrival differences.


Even quite wide-space omnis isn't going to decorrelate the lowest frequency information.  The spacing may be wide enough at those frequencies to produce some stereo difference phase information, but to be decorrelated, the signals need to have a difference of more than one full wavelength.  The first cancellation notch in the graphs appears at a frequency where the phase difference between the two signals is exactly half a wavelength apart (the phase difference = 180 degrees).

[MIQ]

Thanks Gut, those are all important notes.

While the graphs above do not apply directly to stereo playback at all spots in the room, the fundamentals are similar. The acoustic summing will cause combing based on the delay between the two channels arriving at the listening position in the room.  For listeners sitting in the center line of the playback system this delay between acoustic signals will be dominated by the delay between the recorded electrical signals however room reflections will also complicate the scene. There will be low freqs where these acoustic signals will combine destructively. Of course the combination will never be as destructive as the theoretical worst case (like the electrical combination).  Also, as you noted there will be so many reflections and complex interactions at high freqs that it becomes random and not clearly defined.   You can see how bunched up the combs get at high freqs. Due to the abundance of reflections at high freq this region of the room response becomes more "statistical" and is hard to predict or simulate.  You get a ton of reflections combining acoustically.

Miq

[voltronic]

More food for thought.  This has turned into a very interesting discussion.

MIQ, would you consider developing an Andriod version of your app?  I have been waiting for one to be released for quite some time, and it doesn't appear to be on the horizon.  I use a combination of these apps to get similar results, but yours is a much more efficient solution:

AR Bearing + Baseplate Compass
Protractor
Neumann Recording Tools

This Friday night, I plan to run my omnis for the first marching band show of the season, and I'll be on top of the press box as I mentioned earlier (estimated included angle of 60 deg).  My plan is to go with a 1-meter spacing for this first run, unless you all here think that's going to be too narrow a spacing / too wide an SRA.  If I get the opportunity to get up there again, (which may or may not happen) I'll try 2 meters for comparison.

For competitions through the season I'll be close, either on field level or in the first row of stands.  My starting point there will be 60cm (estimated included angle of 110-120 deg).  The ensemble balance is going to be somewhat worse because of the low height, but that will be balanced by the more optimal placement distance to the band.  A tall stand isn't possible for these events, nor would I risk it with the winds we tend to get.

[Gutbucket]

My plan is to go with a 1-meter spacing for this first run, unless you all here think that's going to be too narrow a spacing / too wide an SRA.  If I get the opportunity to get up there again, (which may or may not happen) I'll try 2 meters for comparison.

Sounds good to me.

A tall stand isn't possible for these events, nor would I risk it with the winds we tend to get.

You'll get some welcome bass reinforcement from having them lower to the ground instead of up high.  Just watch out for spectators in close proximally.

[MIQ]

More food for thought.  This has turned into a very interesting discussion.

MIQ, would you consider developing an Andriod version of your app?  I have been waiting for one to be released for quite some time, and it doesn't appear to be on the horizon.

Hi Voltronic

I get that question a lot. Unfortunately, the other half of the Stereo Mic Tools development team (Rob) is not set up to code for Android. He is strictly iOS for the forseable future. I was piecing together iPhone apps similar to you before we put together Stereo Mic Tools. 

I'm sorry we can't help you. How about buying a used iPod touch for cheap so you can run our app? 

MIQ

[Colin Liston]

i'm a big fan of spaced anything.

we've run "split stack" with cards before.  one mic on the left stack and one on the right stack.  came out amazing.  really picked up the stereo mix well and you hear one fan in one ear and not in the other too.  interesting listen.

if runnig ominis back by the board...  why not space them out as far as you can?  one mic/stand on the right side of the board, one on the right, both pointed at the respective stacks?  same for cards, hypers, whatever...

it's the best stereo config you could do, if you have the gear....

my split cards:  https://archive.org/details/dbt2005-03-02-dpa4023.flac16

[Gutbucket]

Colin, you did a similar straight split cards at the Suwannee a year or two back which I remember sounding quite nice.  Sounded very similar to my 6' split omnis from the same location there.  How big was that split? like 8' or so?

[DSatz]

A late entry to this thread: If the two schools of thought are ~3 feet apart vs. ~1 foot apart, times have really changed since I was first learning to record--and very much for the better as far as I'm concerned. If you look in any American-published books or magazine articles from the 1960s and 1970s about stereo recording, the "cookbook" formulas were exactly two: One was to take the width of the stage, divide it by three, and place one omni at each of the lines of demarcation (i.e. 1/3 of the way across and 2/3 of the way across respectively). The other was to take a pair of cardioids and aim them with the heads pointing together at a 90 degree angle.

Both are perfectly dreadful ways of recording music in stereo, and are especially regrettable because with the same microphones, it's so easy to do so, so much better.

But as far as omnis are concerned, the more enlightened texts used to distinguish between "small" A/B, which was the European style, and "large" A/B, which was the American style as described above. Back then, the two encampments were pretty much mutually exclusive. I remember being shocked that small A/B could work at all, since the 1/3 and 2/3 rule was the only approach I ever saw used; with anything less than that wide spacing, I couldn't believe that you would "get enough separation" for stereo until I tried it myself.

By the way, this included the fellow who used to record our high school band back in those years, as well as the Honors and State bands in Western Pennsylvania that I played in: a Mr. Renner who ran a company called "Century Recording Service" just on the other side of the Ohio state line. He later became better known as the chief engineer for Telarc Records. While I'm happy for the Grammy awards he received (and the fact that he mainly used Schoeps microphones), most of the famous Telarc orchestral recordings have a huge hole-in-the-middle problem; once it's pointed out, you can't not hear it.