Line arrays and inter-element angles.

[Helge A. Bentsen]

I've noticed some trends in my array designs lately independent of model/manufacturer.

Several auto-design features I've tried in various line-array calculators suggest one should use a very sharp J in the bottom of the array. Sometimes they suggest 0 or 1 degrees for a large part of the array and two or three boxes at the bottom with steep angles, the bottom one often at max. According to the calculators this should give you the a more even SPL distribution across the audience area. But in my experience, using 0 degrees and a sharp J gives you some serious sonic disadvantages, so I don't use this approach.

I tend to prefer arrays with a more gradual curvature, starting with at least 1 degree at the top and finishing at something not to extreme at the bottom. Let's for the sake of this discussion say 2-2-3-4-4-5-6-6. This sounds more pleasing to my ear and those who mix the shows. But sometimes I run into situations where I can't get the array high enough up in the air so the SPL differences gets annoyingly large. I've tried using smaller angles and/or fewer boxes or some gain/eq shading, but this is not always enough.

Is there a third way? How do you guys deploy arrays in small venues with limited ceiling height?

I'm often hired to tech other peoples jobs, so I have limited input on box count and model.

 

[Silas Pradetto]

Re: Line arrays and inter-element angles.

How do you guys deploy arrays in small venues with limited ceiling height?

Might want to consider something other than an array, or use delays.

 

[Tim McCulloch]

Re: Line arrays and inter-element angles.

Great topic, Helge. I've noticed this, too.

I think some of the prediction/aiming calculators have been tweaked to customer desires to have a shorter/lighter array while maintaining near-zone coverage. Tomorrow I'm adding elements to an array already deployed because the BE doesn't want to use front fills.

Clair's implementations are _very_ J shaped. With the i5 and presumably i3, the elements come in 60°, 90°, and 120°-ish horizontal dispersion. I think it's kind of distracting to hear the program from above me, but Clair has a stout front fill package to help pull the image down to the stage.

But back to your observation about 0°, I've seen/heard this and to my ears it makes for a very inconsistent spectral response in the middle of the floor, progressing back. At the back and on bowl of an arena or stadium, the HF then outruns everything <800Hz or so. This can be seen in JBL's LAC-II if one takes the time to look at different virtual room mics at more than 1 or 2 frequency bands. This may or may not be a problem depending on the venue. In a venue with lots of low/mid build up in the reverberant field this could be a good thing, but for a better behaved environment it could be a problem.

I still use JBL's Excel spread sheet for a quick auto-design; it doesn't usually give me any 0° except for the longest throws at the very top of the hang, but I find myself making changes to its suggestions when modeled in LAC-II, usually for better coverage down front (see first paragraph). All the tricks you mention, shading and trim height/bumper angles, are needed and often are not enough, as you point out.

In one hotel ballroom gig I do, we hang the arrays upstage. The horizontal coverage and the width of the stage means most of the stage in not in the direct coverage of the system, but still covers all of the front row except for about 3m in the very middle (a single, small front fill kicks ass here). This is a corporate talking head event with presentations from the DSR lectern as well as panel discussions and audience participation off the stage. Usual lavalier mic fun, too.

It will be interesting to see what others are doing. Thanks for posting this.

Tim Mc

ps. Dick Rees once said "Accept that it's a crap situation and just take the money." I'd add to that to read "Do the best you can with the resources available, accept that it's a crap situation, just take the money and don't feel guilty about it."

 

[Jens Bacher]

Re: Line arrays and inter-element angles.

Interesting topic. I think you must take into account the HF spread of the box. I just made a quick rendering from Soundvision, showing a single K1, a Kara and a Kiva box playing straight into a surface with 8KHz shown:

Notice how the K1 is much more focused than the Kiva, because the K1 waveguide is not curved.
Lets try and insert 5 boxes of each type and angle them 10 degrees (not possible in real life with K1 for obvious reasons):


This should show why 5 large format boxes (K1, I5 etc.) and 5 small-format boxes (Kiva, D&B T etc.) will behave quite different, even with the same number of boxes and angles. Also aiming for same SPL everywhere is ofen dangerous, if you design for a -3db reduction/doubling you will often end with a better sounding system.

 

[Brandon G Romanowski]

Re: Line arrays and inter-element angles.

The problem with J arrays like you describe in the OP is you end up with a line array and a point source hanging underneath the line array. It is a dramatic thing to go from 0 to even 1 degree, and its just not a single array anymore when you jump from 0 to 6 degrees. It becomes 2 very separate arrays like main PA vs Front fill. "We are trying to mix together two shapes that cannot possibly be married together over distance" -606 McCarthy
........ BTW if you point all your speakers at the ceiling the coverage really evens out on those calculators

 

[TJ Cornish]

Re: Line arrays and inter-element angles.

I've wrestled with this a lot over the last 2 years, trying to understand what going from a constant-curve system to a variable angle system would mean in the rooms I get to work in, in my case ceiling heights of 16' or so, and throws up to about 80' or so. I have no first hand system tech experience of anything more than 5 boxes/side of Vertec 4886, but have many hours invested in various calculators.

Some observations, not presented as gospel truth:

- I'm convinced that certain calculators lie. RCF and DBTech come immediately to mind. The answers are overly simplistic, and they give no indication of how the answer is arrived at. The DBTech calculator does as Helge mentioned - all boxes at 0 degrees except the last one at full angle. I simply do not believe the results. The Nexo calculator also seems to do a lot of "smoothing" in frequency plots, though I know from experience that the results do actually sound very good.

The other calculator I've spent significant time in - JBL LAC, is the only one I have experience with (I'm sure other calculators are good - I just don't have time on them) that I believe the results - simply because the results aren't overly glossy. Interestingly, I have found with my admittedly limited experience, that setting the array to the angles that LAC suggests produces a good result, however the overall tonal shape doesn't seem to match what I measure - LAC suggests a significant HF boost as being required, though in practice the system is almost bang-on with V5 presets - no HF boost required.

- The shapes that LAC suggests are of the typical form that Helge referenced - increasing increments between boxes done gradually - what RCF calls a "spiral" array. This also makes intuitive sense to me - more elements firing at the back of the room, and gradually fewer of them - accomplished by the increasing angles between boxes as you go down the array - as you go forward. The "everything at 0 degrees except the bottom box" method makes no sense.

- Two possible end goals - constant volume, or constant tonality. When I first started playing around, I spread the boxes out so that volume across the depth of the room was the same. This only works down to where the array loses pattern control, and depending on the array length, the array may lose most of its pattern control north of 500Hz. After starting a thread on the forum here and asking the local pro I bought the system from, it became clear the real goal is to make tonality as similar as possible over the whole room. That's great, but then the array behaves similar to what we all "know" of a point source - 6dB drop in level for every doubling of distance, and not the mythical 3dB that line arrays are supposed to be able to do.

Extending this logic is very interesting to me. Most folks say that a real "line" starts at about 2 meters, or 6 1/2'. That holds pattern control down to around 400Hz, according to LAC. Extending this to 4 meters means that control extends down to 200Hz or so. This is all well and good, but even with a monster 6 meter array, there comes a frequency where the array loses pattern control and behaves like a point source, meaning that tonality changes over distance.

It seems like throws up to 80 - 100' aren't too big of an issue in a room with 16' ceilings and a modest array length. Much more than that seems to be untenable without either a pretty severe volume gradient or delay speakers.

 

[TJ Cornish]

Re: Line arrays and inter-element angles.

........ BTW if you point all your speakers at the ceiling the coverage really evens out on those calculators

I know you're kidding, but is interesting to me that I always would tend to aim the array too low in my initial models. It seems that the top box(s) of the array need to point over everyone's heads to pull the pattern up enough to cover the back of the room.

 

[Brandon G Romanowski]

Re: Line arrays and inter-element angles.

I usually aim too low because the prediction plane is on the ground and I like to forget that the listener's ears are 5' above that.

 

[mackerr]

Re: Line arrays and inter-element angles.

I know you're kidding, but is interesting to me that I always would tend to aim the array too low in my initial models. It seems that the top box(s) of the array need to point over everyone's heads to pull the pattern up enough to cover the back of the room.

The actual coverage pattern is usually smaller than the included angles of the array. If you use a laser on the top speaker it should be well above the intended top coverage. On the bottom the natural off axis rolloff works in your favor.

Mac

 

[Evan Kirkendall]

Re: Line arrays and inter-element angles.

Clair's implementations are _very_ J shaped. With the i5 and presumably i3, the elements come in 60°, 90°, and 120°-ish horizontal dispersion. I think it's kind of distracting to hear the program from above me, but Clair has a stout front fill package to help pull the image down to the stage.

Close, but no cigar.

FWIW- The i3 is a 140x10 degree box. No other flavors available there. The i5 comes in 4 flavors- 90x2.5, 90x5, 90x10 or 120x10. The iDL is the only 60 degree line array element.

The secret to hanging the i5 is to not pin the angles at all(just the travel pin), and just pull them back to their native box angles- hence why you usually see 4 2.5's, 4 5's and 4 10's in an array. Gives you nice, consistent coverage.


Back on topic-

I am not a fan of doing 0 0 0 0 0 0 0 0 0 max max max on arrays. Or for that matter, I hate "downfall" boxes in general. The worst I usually see is the L'acoustics stuff with the main hangs pinned 90% flat, and then 3 or so downfall boxes covering the whole floor. Why people like to send 90% of the sound right into the rear seats and then rely on a few 8" drivers to cover the whole floor is beyond me... I personally like a nice even spread. I take the main hang and hit everything from the top row, all the way to the barricade. I'll use gain shading as needed, but it usually works out pretty evenly when I can fly high enough.

As far as low trim heights- I usually end up with less boxes, pinned at lesser angles. I'll still try to stay away from 0, but lesser numbers are common. Gain shade the bottom boxes down, and EQ zones as needed to get things consistent. There's always trade offs, but you make the best out of what you can. There's been a few situations where flat and a few degrees down is the only option that works, and you as the SE make it work.



Evan

 

[Sean Thomas]

Re: Line arrays and inter-element angles.

Just currious - What is the issue with 0-0-0 hangs if you need 150' to 200' of coverage, expecially if the software specs it? Are you suggesting that LA boxes don't sound good at 0?

My last hang out doors was 20' to 150' with (4) T12's and (2) T8's was 3 0 0 3 3 3 (per software spec). So my first 3 boxes were all 0 and straight.

 

[Timo Beckman]

Re: Line arrays and inter-element angles.

This is a nice way of looking at things :
SIM3 Optimization & Design Seminar at UC Irvine

Specially the bits about tuning the parts of the array to the far field "Golden Trace" is something i do not see a lot (mostly due to not getting enough time)

 

[Richard Mitton]

Re: Line arrays and inter-element angles.

The D&B Line Arrary theory workshop (Line array workshop | d&b audiotechnik) is highly recommended if you're interested in this stuff. They do an excellent demonstration with their flexible "klein array" showing how the J shape works and what happens as the angles change.

 

[Tim McCulloch]

Re: Line arrays and inter-element angles.

The spiral array is discussed by Mark Ureda here: http://www.jblpro.com/products/Tour...c_manual/MSU AES Fall 2001 Monday Morning.pdf

Mark Engebretson has some other thoughts: http://www.jblpro.com/products/Tour/Vertec/support/vertec_manual/AES_MarkE_DirRadiation.pdf

Your observation about the DBTech and RCF calculators is spot on. EASE Focus 2 produces an audibly superior result with very different angles and different element EQ settings.

Brandon's comment about aiming the PA waaaaay up is not without validity. If one looks at LAC-II (I keep referencing it because I use it most), you can see that the HF, MF & LF lobes have different shapes. The MF & LF typically collapse toward the middle of the line (experiment with 4, 8, 16 & 24 element hangs to see what happens) so tipping the array UP will make sense if you really want the MF to match the HF... some folks do, others take the "hey, it's better than what we did 10 years ago" approach and let it go.... but it makes covering the front of the room a genuine challenge because of the weight, length, and cost of the array needed. This is one of the compromises I alluded to in my first post. This is also what TJ is talking about when he says

- Two possible end goals - constant volume, or constant tonality. When I first started playing around, I spread the boxes out so that volume across the depth of the room was the same. This only works down to where the array loses pattern control, and depending on the array length, the array may lose most of its pattern control north of 500Hz. After starting a thread on the forum here and asking the local pro I bought the system from, it became clear the real goal is to make tonality as similar as possible over the whole room. That's great, but then the array behaves similar to what we all "know" of a point source - 6dB drop in level for every doubling of distance, and not the mythical 3dB that line arrays are supposed to be able to do.

It can be done, but you need a lot more gear, trim height and load support to accomplish it.

As the adults in the room, we have to decide what compromises to make.

 

[Chris Johnson]

Re: Line arrays and inter-element angles.

I would certainly agree that in general, excessive and sudden variation in array element angles doesn't sound good.

However, I'm not sure this is a failure of the calculators, but rather a failure of the user to interpret the data.

I bit like with smaart, if you apply enough averaging, you can make anything look flat

If I take an array and I use d&b arraycalc, it will highlight the fact that while at 1k, extreme array element angles makes little difference to the smoothness of coverage (particularly with small arrays and their high lower-limits for line-array like pattern control), switching the trace to 8k will highlight these problems. As will using the 3D plot view.

So I think the information is there, but I think people just misuse it. I also think people worry too much about level variation from front to back, and not enough about tonal variation. In a lot of the design work I do, it actually suits the event that there is a natural increase in volume up front, and a smooth volume rolloff through the venue. So why fight the physics? instead, focus on minimizing the volume gradient without compromising tonality. Its no good having it be exactly 90dB everywhere if at the back its all >1k, and at the front all <1k!

Sometimes extreme downfill is a necessary evil (particularly in the corporate and TV world), but in general I prefer to have more gentle variation.

 

[Ivan Beaver]

Re: Line arrays and inter-element angles.

- Two possible end goals - constant volume, or constant tonality. When I first started playing around, I spread the boxes out so that volume across the depth of the room was the same. This only works down to where the array loses pattern control, and depending on the array length, the array may lose most of its pattern control north of 500Hz. After starting a thread on the forum here and asking the local pro I bought the system from, it became clear the real goal is to make tonality as similar as possible over the whole room. That's great, but then the array behaves similar to what we all "know" of a point source - 6dB drop in level for every doubling of distance, and not the mythical 3dB that line arrays are supposed to be able to do.

.

And there are limitations to those "theories. For example people assume that because with a point source the level is actually dropping off at a 6dB rate-but that ONLY applies if everybody is on axis of the loudspeaker. In sound design 101-you point the loudest part of the speaker (the middle) towards the furthest listener. So that the people that are closer are in a part of the coverage that the level is lower. This can result in ZERO (give or take a little) difference in front to back-NOT the 6dB people just automatically "throw out".

And unless the line is long enough (most aren't-due to budget-weight etc) the 3dB rule does not apply at all freq. So you have a different freq response at seats from front to back. Yes some some freq will follow the 3dB rule-but other will follow the 6dB rule.

Not exactly what people would like to "believe".

 

[TJ Cornish]

Re: Line arrays and inter-element angles.

And there are limitations to those "theories. For example people assume that because with a point source the level is actually dropping off at a 6dB rate-but that ONLY applies if everybody is on axis of the loudspeaker. In sound design 101-you point the loudest part of the speaker (the middle) towards the furthest listener. So that the people that are closer are in a part of the coverage that the level is lower. This can result in ZERO (give or take a little) difference in front to back-NOT the 6dB people just automatically "throw out".

And unless the line is long enough (most aren't-due to budget-weight etc) the 3dB rule does not apply at all freq. So you have a different freq response at seats from front to back. Yes some some freq will follow the 3dB rule-but other will follow the 6dB rule.

Not exactly what people would like to "believe".

True, but the pattern control limitations work for point source speakers as well - you can point the speaker at the back row and gain back some of the level gradient, but only for frequencies the speaker has pattern control over.

 

[Art Welter]

Re: Line arrays and inter-element angles.

FWIW- The i3 is a 140x10 degree box. No other flavors available there. The i5 comes in 4 flavors- 90x2.5, 90x5, 90x10 or 120x10. The iDL is the only 60 degree line array element.

The secret to hanging the i5 is to not pin the angles at all(just the travel pin), and just pull them back to their native box angles- hence why you usually see 4 2.5's, 4 5's and 4 10's in an array. Gives you nice, consistent coverage.

Back on topic-

I am not a fan of doing 0 0 0 0 0 0 0 0 0 max max max on arrays. Or for that matter, I hate "downfall" boxes in general.
Evan

Evan has pointed out a very important difference between the i5 and most other systems- it is designed for proper divergence shading, using narrow HF vertical dispersion for the elements that cover longer distance, and wider for shorter.

Most line arrays offer one choice- a 5 degree, 10 degree or whatever. Some offer a wider downfill, or other smaller boxes that can be rigged at the bottom.
Problem is, any HF vertical dispersion angle other than 0 is not correct for a 0 degree hang, so the compromises begin.
All the people discussing here know that horizontally overlapping patterns, like the obvious comb filter mess of two flat fronted 90 degree cabinets side by side is problematic, but seem to be ignoring the same problem in the vertical dispersion in line arrays, where the high frequency dispersion is determined largely by horn throat design, not line length.

What is best as far as inter-element angles is completely dependent on the HF design angle- flattening out a 2.5 degree HF dispersion element is quite different than flattening out a 10 degree element.

When discussing the compromised situation of one vertical dispersion angle used for all the boxes in an array, that angle is very important, generalizations can't be made that because 2-2-3-4-4-5-6-6 works on one system it will on a system with a different vertical dispersion angle.

 

[Ivan Beaver]

Re: Line arrays and inter-element angles.

True, but the pattern control limitations work for point source speakers as well - you can point the speaker at the back row and gain back some of the level gradient, but only for frequencies the speaker has pattern control over.

Even if there is no pattern control-the 6dB rule is still not what people like to refer to.

Lets say a venue is 100' deep. That is a 30dB loss from front to back (assuming 3' is the closest listener
But if the speakers are flow at say 25', then the closest listener is 14dB down and the furthest listener is still 30dB down. So the difference is right at half of what the 6dB rules says-so it "acts" as if it were 3dB- like the line.

So by proper DESIGN there are ways around the "common thinking" and misconceptions.

Understanding the basic principals goes a LONG way-much further further than "wives tales".

 

[TJ Cornish]

Re: Line arrays and inter-element angles.

Even if there is no pattern control-the 6dB rule is still not what people like to refer to.

Lets say a venue is 100' deep. That is a 30dB loss from front to back (assuming 3' is the closest listener
But if the speakers are flow at say 25', then the closest listener is 14dB down and the furthest listener is still 30dB down. So the difference is right at half of what the 6dB rules says-so it "acts" as if it were 3dB- like the line.

So by proper DESIGN there are ways around the "common thinking" and misconceptions.

Understanding the basic principals goes a LONG way-much further further than "wives tales".

Agreed - my point was that this behavior is the same for both a line array and a point source.