Speakerplans.com Homepage
Forum Home Forum Home > General > General Forum
  New Posts New Posts RSS Feed - 12 line array cab plan
  FAQ FAQ  Forum Search   Events   Register Register  Login Login

12 line array cab plan

 Post Reply Post Reply Page  <1 56789 13>
Author
Message
toastyghost View Drop Down
The 10,000 Points Club
The 10,000 Points Club
Avatar

Joined: 09 January 2007
Location: Manchester
Status: Offline
Points: 10920
Post Options Post Options   Thanks (0) Thanks(0)   Quote toastyghost Quote  Post ReplyReply Direct Link To This Post Posted: 10 January 2019 at 2:40pm
Just to continue the 'data' fun, here's a hang of 6x RCF HDL6A, flown 6m trim, with autosplay. Input signal is AES Broadband pink noise, and the complex summation is enabled up to 16KHz. By default they don't have it up this high because it's heavy on the CPU. All views are 1/3 oct. I've only flown this high to make the dispersion clear, but it's a realistic trim for a gig in say a 1000 pax theatre venue:

16k


8k


6.3k


4k


1k


315hz


Anybody see any kind of cylindrical wavefront effect in play? At what frequencies? Not a trick question, I'm genuinely asking if someone can point out where you get only 3dB loss per doubling of distance. Feel free to recreate the sim yourself, it's just using free EASE Focus 3 and the data on RCF's site.

So to all those people who say the inter-box interactions aren't that noticeable, look at the plan view, and the side view. Now imagine you have important audience members in some of those nulls.

Or, imagine that there's a reflective, concrete wall behind the stage, and a tin roof above, at the height the sim stops. Check out the spurious side lobes that come in hard at 6.3KHz and again at 4KHz. The rear beam and side lobes at the midrange. How nice would your mix be in a difficult venue, let alone with open mics roaming in the space?

This is my main issue, they're useful tools but there are clear flaws that have serious impact in the real world. But of course, smoothed measurements, broad views of multiple octaves, filtered stimulus all makes this less obvious, and often a sim and design isn't even done, it's just thrown up and the boxes aimed roughly where the crowd are.

For what it's worth, I never, ever use autosplay for a real gig, but many, many people do.


Also sorry Matt, I'm not picking on you or your new baby, it was just a box at the forefront of my mind and RCF actually share data in an easy to find place

Edited by toastyghost - 10 January 2019 at 2:47pm
Back to Top
MattStolton View Drop Down
Old Croc
Old Croc
Avatar

Joined: 04 September 2010
Location: Walthamstow
Status: Offline
Points: 4234
Post Options Post Options   Thanks (0) Thanks(0)   Quote MattStolton Quote  Post ReplyReply Direct Link To This Post Posted: 10 January 2019 at 2:51pm
Originally posted by toastyghost toastyghost wrote:


Also sorry Matt, I'm not picking on you or your new baby, it was just a box at the forefront of my mind and RCF actually share data in an easy to find place


None taken, don't worry!

And, as you mention, the RCF data is freely published by them, and it does reflect what you hear.

As you also said, flexibility and ease/cleverness of rigging was really important, even to the detriment of sound. And, what I paid per box was silly cheap!

However, in their defence, you should do the same EASE work with some of the competition! I did, and they were even more horrific.

I presume the 800Hz horizontal drop out is a function of the 6" drivers being 300mm apart, but the HF horn comes in from 900 Hz. Perhaps they should have got it to work lower?
Matt Stolton - Technical Director (!!!) - Wilding Sound Ltd
"Sparkius metiretur vestra" - "Meter Your Mains"
Back to Top
toastyghost View Drop Down
The 10,000 Points Club
The 10,000 Points Club
Avatar

Joined: 09 January 2007
Location: Manchester
Status: Offline
Points: 10920
Post Options Post Options   Thanks (0) Thanks(0)   Quote toastyghost Quote  Post ReplyReply Direct Link To This Post Posted: 10 January 2019 at 3:00pm
Oh trust me, I've done the same sims and real world measurements with plenty I'd advise anyone looking at using, let alone buying, an array to do the same.

My main issue with this sort of tiny box is the HF horn is really only working like a line array horn from 8k to 16k. Below that it's into normal point source horn territory, with all the comb filtering you'd expect, unless you go for extreme angles that then mess up the lower frequencies.

Edited by toastyghost - 10 January 2019 at 3:02pm
Back to Top
MattStolton View Drop Down
Old Croc
Old Croc
Avatar

Joined: 04 September 2010
Location: Walthamstow
Status: Offline
Points: 4234
Post Options Post Options   Thanks (0) Thanks(0)   Quote MattStolton Quote  Post ReplyReply Direct Link To This Post Posted: 10 January 2019 at 3:02pm
Originally posted by toastyghost toastyghost wrote:

For what it's worth, I never, ever use autosplay for a real gig, but many, many people do.

Agree on autosplay being a bit mewh.

Tend to use it for first guess, and then fiddle to see what changes as angles change, and then do my own "iterations". As ever, it is about adjusting until the compromises are optimised to real life.

Funny you picked 315Hz as one of your lobe pictures, my first use was just a couple on sticks and EASE did predict the lobe aimed directly at mic at 315Hz, and, on site, it was there! Forewarned was forearmed, so no great issue, and gave me some faith that the EASE data was true.
Matt Stolton - Technical Director (!!!) - Wilding Sound Ltd
"Sparkius metiretur vestra" - "Meter Your Mains"
Back to Top
gen0me View Drop Down
Young Croc
Young Croc
Avatar

Joined: 20 February 2016
Location: UK
Status: Offline
Points: 999
Post Options Post Options   Thanks (0) Thanks(0)   Quote gen0me Quote  Post ReplyReply Direct Link To This Post Posted: 10 January 2019 at 7:34pm
Basicly looking at this there is no chance you would get as huge tonal variation from well design horns sprayed by each other (f1 style spraying).
Are you sure they are properly tilt? How with the manual positioning?

Notice on the second post tonal variety on 1khz. On the view from the top you can clearly see the centre of spl on audience is moved from another frequency sims. 
This:

Is unweighted. Keep in mind 0db on axis to keep it into perspective that for it you still need dsp work.

Originally posted by toastyghost toastyghost wrote:


My main issue with this sort of tiny box is the HF horn is really only working like a line array horn from 8k to 16k. Below that it's into normal point source horn territory, with all the comb filtering you'd expect, unless you go for extreme angles that then mess up the lower frequencies.
What could be better hf waveguide choice for this splaying? This one clearly has issues around 1k and under 2.5k
Or things would change with better manual splaying?

I appreciate every like :)) https//www.facebook.com/genomesoundsystems
Mixes: https://www.mixcloud.com/gen-ome/
Back to Top
gen0me View Drop Down
Young Croc
Young Croc
Avatar

Joined: 20 February 2016
Location: UK
Status: Offline
Points: 999
Post Options Post Options   Thanks (0) Thanks(0)   Quote gen0me Quote  Post ReplyReply Direct Link To This Post Posted: 10 January 2019 at 7:42pm
F1 style spraying: 2 bottom ones splayed 40deg give the top one:
1 colour is 2db.
Its intended bandwidth is up to 2.4khz. Higher you would have to deal with hf horns.


Edited by gen0me - 10 January 2019 at 8:10pm
I appreciate every like :)) https//www.facebook.com/genomesoundsystems
Mixes: https://www.mixcloud.com/gen-ome/
Back to Top
toastyghost View Drop Down
The 10,000 Points Club
The 10,000 Points Club
Avatar

Joined: 09 January 2007
Location: Manchester
Status: Offline
Points: 10920
Post Options Post Options   Thanks (0) Thanks(0)   Quote toastyghost Quote  Post ReplyReply Direct Link To This Post Posted: 10 January 2019 at 8:47pm
hen0me: these are boxes with built in DSP. The data is collected with the DSP and all correctional tools from factory already applied. The splay is ‘correct’ as much as any given splay is ‘correct’ - it’s down to design goal for the hang, and there are things you can do to help mitigate these issues but as Merlijn says, pick your battles because you don’t have long even if you have a million channels of DSP.

These issues occur on ALL line arrays, but at different frequencies depending on box size, arrangement, J curve and line length.

efinque: the data is a simulation calculated from GLL balloon data, which is a full 360 degree measurement of magnitude, phase and impulse taken using a single mic in one position, and the speaker rotated in all directions in one degree increments using a robot arm. Usually done in full anechoic or semi-anechoic environment, with the latter being a windowed measurement.

To be really effective at the entire range of stated frequencies a line array HF horn just needs to be really, really deep. Like more than a metre deep.

Edited by toastyghost - 10 January 2019 at 8:48pm
Back to Top
fatfreddiescat View Drop Down
Young Croc
Young Croc
Avatar

Joined: 15 October 2010
Location: N.E.Wales
Status: Offline
Points: 1081
Post Options Post Options   Thanks (0) Thanks(0)   Quote fatfreddiescat Quote  Post ReplyReply Direct Link To This Post Posted: 10 January 2019 at 9:02pm
Originally posted by efinque efinque wrote:

Originally posted by toastyghost toastyghost wrote:

Just to continue the 'data' fun, here's a hang of 6x RCF HDL6A, flown 6m trim, with autosplay. Input signal is AES Broadband pink noise, and the complex summation is enabled up to 16KHz. By default they don't have it up this high because it's heavy on the CPU. All views are 1/3 oct. I've only flown this high to make the dispersion clear, but it's a realistic trim for a gig in say a 1000 pax theatre venue

Cool.. are those measurements/graphs hard to take?

I mean, do you need a special measurement mic to plot from several points or something?


Those graphs are using data in the form of 'ease' files which are supllied by the manufacturer so users can model the acousti c response of different arrays to work out what would work for their needs.The graphs can only be as accurate as the data provided ther efore measurements would normally be done with a reference grade measurement mic, the amount of data to be collected is not insubstantial, I have not persoanally looked at the criteria but would frequency response and phase response at multiple positions over a 360 sphere around the loudspeaker if I'm not mistaken.

Gen0me , narrowing at 2.5Khz may well be 'waistbanding' caused by the shape of the horn flare, at 1Khz likely to be the 2 * 6” drivers narrowing due to spacing as full width of the cabinet is a radiating surface around the xover frequency.
Back to Top
gen0me View Drop Down
Young Croc
Young Croc
Avatar

Joined: 20 February 2016
Location: UK
Status: Offline
Points: 999
Post Options Post Options   Thanks (0) Thanks(0)   Quote gen0me Quote  Post ReplyReply Direct Link To This Post Posted: 10 January 2019 at 10:05pm
Originally posted by toastyghost toastyghost wrote:

hen0me: these are boxes with built in DSP. The data is collected with the DSP and all correctional tools from factory already applied. The splay is ‘correct’ as much as any given splay is ‘correct’ - it’s down to design goal for the hang, and there are things you can do to help mitigate these issues but as Merlijn says, pick your battles because you don’t have long even if you have a million channels of DSP.
Dsp in terms of eq which is the same on each box? Or dsp as applied delays to boxes, or even doing fft on every box input, and using linear phase filters. Composing frequencies back from fft with different delay applied.
First one?
Im curious how far can go the second one.

Originally posted by fatfreddiescat fatfreddiescat wrote:

Gen0me , narrowing at 2.5Khz may well be 'waistbanding' caused by the shape of the horn flare
See that 8" flare dont have it in its usefull bandwitch or rather it can be equed flat independently from listeners position.
I appreciate every like :)) https//www.facebook.com/genomesoundsystems
Mixes: https://www.mixcloud.com/gen-ome/
Back to Top
fatfreddiescat View Drop Down
Young Croc
Young Croc
Avatar

Joined: 15 October 2010
Location: N.E.Wales
Status: Offline
Points: 1081
Post Options Post Options   Thanks (0) Thanks(0)   Quote fatfreddiescat Quote  Post ReplyReply Direct Link To This Post Posted: 10 January 2019 at 10:20pm
Originally posted by gen0me gen0me wrote:

Originally posted by toastyghost toastyghost wrote:

hen0me: these are boxes with built in DSP. The data is collected with the DSP and all correctional tools from factory already applied. The splay is ‘correct’ as much as any given splay is ‘correct’ - it’s down to design goal for the hang, and there are things you can do to help mitigate these issues but as Merlijn says, pick your battles because you don’t have long even if you have a million channels of DSP.
Dsp in terms of eq which is the same on each box? Or dsp as applied delays to boxes, or even doing fft on every box input, and using linear phase filters. Composing frequencies back from fft with different delay applied.
First one?
Im curious how far can go the second one.

Originally posted by fatfreddiescat fatfreddiescat wrote:

Gen0me , narrowing at 2.5Khz may well be 'waistbanding' caused by the shape of the horn flare
See that 8" flare dont have it in its usefull bandwitch or rather it can be equed flat independently from listeners position.

Not possible to eq it flat if it's due to the waveguide geometry, if it's due to filter/phase interaction then maybe possible to do something with dsp.
Back to Top
toastyghost View Drop Down
The 10,000 Points Club
The 10,000 Points Club
Avatar

Joined: 09 January 2007
Location: Manchester
Status: Offline
Points: 10920
Post Options Post Options   Thanks (0) Thanks(0)   Quote toastyghost Quote  Post ReplyReply Direct Link To This Post Posted: 10 January 2019 at 10:28pm
Default EQ which is the same on each box, and some factory presets for some generic hangs I would assume.

At this level of box I’ll be very, very surprised if anyone is remotely deploying it in close coupled coherent sections, each with their own filtering, independently measured.

Typically on a ‘proper’ system you break the hang into three or more clusters, aiming at specific parts of the audience, and sharing inter-box angles. You can then measure each ‘cluster’ and treat it as one box.

Often EQ does very little here, because it will make other issues worse at other positions, and it has a serious law of diminishing returns. In fact it’s often better to drop the hang and change angles based on new predictions and verification.

However some careful allpass placement or delay can help, but again, you need to measure in minimum three positions, on axis. When you’re happy, you verify off axis, and copy to the other hang.

The funny thing is, at this point, you’re actively trying to reduce the tonal variance in response because of HF absorption, LF buildup, and the inconsistent ‘cylindrical wavefront’ behaviour. Which means you basically push them back closer to a vertical hang of big point source ‘boxes’. So why not just use big, true point source boxes in the first place, rather than this mish mash of different design components?

The other issue is, how much time do you have on a real gig to do this? With shorter load in times, most gigs needing sound checks for multiple bands, and other people working on light, screen, etc? Let alone shows with more than a left and right. This is why so many people whack the boxes up by eye, or at best, spend ten minutes in the prediction software. Particularly at the level of smaller, cheaper cabinets. E.g. smaller than 1500 cap.

Hence the drive to preshow design and FIR calculation with uploads to the boxes on site. FIRmaker, etc all do this, and rely on good venue and box measurements. Very few are verified properly in the field, the tech is trusted, and frankly, the work of the algorithm is kind of beyond a field measurement anyways. Which makes more and more ‘engineers’ become operators, really, which is a fundamental problem.

Edited by toastyghost - 10 January 2019 at 10:30pm
Back to Top
smitske96 View Drop Down
Young Croc
Young Croc


Joined: 16 February 2016
Location: The Netherlands
Status: Offline
Points: 1085
Post Options Post Options   Thanks (0) Thanks(0)   Quote smitske96 Quote  Post ReplyReply Direct Link To This Post Posted: 10 January 2019 at 11:29pm
Originally posted by toastyghost toastyghost wrote:

 Very few are verified properly in the field, the tech is trusted, and frankly, the work of the algorithm is kind of beyond a field measurement anyways. Which makes more and more ‘engineers’ become operators, really, which is a fundamental problem.

+1
Back to Top
 Post Reply Post Reply Page  <1 56789 13>

Forum Jump Forum Permissions View Drop Down

Forum Software by Web Wiz Forums® version 12.06
Copyright ©2001-2023 Web Wiz Ltd.

This page was generated in 0.157 seconds.