Array Configuration

B

Brent Venter

Freshman
Mar 31, 2011
22
0
1
So while attending the EASE training seminar an array optimisation technique was presented that was new to me. Instead of splaying speakers at their nominal opening angle in a point source the speakers are tight packed. The quantity of speakers should then be uneven e.g 3 speakers. The two outer speakers are on a different processor channel than the center one. Ease then has a function to delay the speakers relative to a virtual point/speaker behind this cluster.

I have seen the 1 kHz polar balloon of this and how the lobing improves dramatically after the processing is applied.

Anyone have experience with this. Can this hold over the frequency spectrum or is it wavelength dependant?

I will try get some graphics up soon.
 
Re: Array Configuration

Brent Venter said:
So while attending the EASE training seminar an array optimisation technique was presented that was new to me. Instead of splaying speakers at their nominal opening angle in a point source the speakers are tight packed. The quantity of speakers should then be uneven e.g 3 speakers. The two outer speakers are on a different processor channel than the center one. Ease then has a function to delay the speakers relative to a virtual point/speaker behind this cluster.

I have seen the 1 kHz polar balloon of this and how the lobing improves dramatically after the processing is applied.

Anyone have experience with this. Can this hold over the frequency spectrum or is it wavelength dependant?

I will try get some graphics up soon.
Click to expand...
Of course it depends on the particular loudspeaker.

And you HAVE to look at other freq. What looks great at 1K may completely fall apart at other freq.

With any model YOU HAVE to look at the entire freq range-not just getting hung up on a single freq.

The splay angle will depend on what the ACTUAL pattern of the loudspeaker is-NOT just the stated angle. Of course if the angle changes with freq-then there is no way you can properly array them POLARS ARE important. NOT just a single published coverage number

In my opinion the cut angles on the cabinets have to cut at 1/2 the rated coverage angle-if there is any chance of it actually being "arrayable".

You must be able to get the HF and mid drivers very close to each other if it is going to be arrayable.
 
Re: Array Configuration

Ivan Beaver said:
In my opinion the cut angles on the cabinets have to cut at 1/2 the rated coverage angle-if there is any chance of it actually being "arrayable".
Click to expand...

The first set of speakers I spent considerable about of time on was the KF650. This drove me crazy. All the marketing lit said "tight pack these boxes" ie. virtual array. Running the math (way before a got my hands on EASE) and even a call to EAW said they needed and additional 20 some degrees between boxes. I even had carpeted wedges made to place between the boxes to eliminate the guess work.
 
Re: Array Configuration

Hmm... I don't want to speak out of line here, but improving a plot at one frequency or for one mic position is easy. The trick would be doing it broadband. If you think about it, the array is already an arc whether flat packed or not - if you draw back to an imaginary "origin" source all the speakers are already equidistant from it. So what are you then delaying to? The problems in the array behavior aren't because of apparent source, but because of multiple sources arriving at disparate times and varying with physical position.
 
Re: Array Configuration

Bennett Prescott said:
Hmm... I don't want to speak out of line here, but improving a plot at one frequency or for one mic position is easy. The trick would be doing it broadband. If you think about it, the array is already an arc whether flat packed or not - if you draw back to an imaginary "origin" source all the speakers are already equidistant from it. So what are you then delaying to? The problems in the array behavior aren't because of apparent source, but because of multiple sources arriving at disparate times and varying with physical position.
Click to expand...


And this is exactly what I questioned with the instructor, and got a reply relating to the sources only having to be in time with each other. This method apparently brings the drivers closer together electronically to approach a theoretical point source.
 
Re: Array Configuration

Brent Venter said:
And this is exactly what I questioned with the instructor, and got a reply relating to the sources only having to be in time with each other. This method apparently brings the drivers closer together electronically to approach a theoretical point source.
Click to expand...

In time with each other where? Adding delay to separated sources will always make them less in time somewhere.
 
Re: Array Configuration

Brent Venter said:
And this is exactly what I questioned with the instructor, and got a reply relating to the sources only having to be in time with each other. This method apparently brings the drivers closer together electronically to approach a theoretical point source.
Click to expand...

This not the case as the instructor describes, unless he/she was doing it schematically at one frequency to make a point. Even then, the description of bringing the drivers together is suspect at best.

What is generally the case, irrespective of the sound source:
  1. A curved, finite arc source of sound will have a narrowing of directivity in the proximit of a wavelength whose dimension is near the total arc curve length. "Line" array, front loaded box, horn loaded box, doesn't matter; that the overall directivity subtends a finite curved angle is what matters.
  2. Delaying the outer corners of the array, in a specific window of frequencies, can help retain the arc's directional consistency through the region of narrowing directivity.
  3. Reducing the output at the periphery via level tapering can achieve similar results. This can be done without as dramatic effects on the polar response at other frequencies.

I've only successfully applied the staggered delay technique a couple times in the wild. In my experience, it only works with large horn-loaded boxes: the correct delay time to broaden the array at, say 500Hz, will cause havoc at 2kHz unless the directivity of the devices is such as to be largely independent at higher frequencies.
 
Last edited:
Re: Array Configuration

Brent Venter said:
Good reply as always Ivan, can you maybe explain this in dummy terms for me to grasp :blush:
Click to expand...
Lets say the cabinet is rated at 60° horizontal coverage. The angle on the trap side must be 30° (1/2 of 60)

The common angle that many trap cabinets are cut at is 10 or 15° That would mean that the coverage angle would have to be 20 ro 30°. You don't see that very often.

The whole idea is to get the drivers close together. If the angle is not "hard enough", then you can't get the drivers close enough to array without interference.

That is the first thing you need to be able to have the cabinets able to actually array- (and not just have array in the name)
 
Re: Array Configuration

Bennett Prescott said:
In time with each other where? Adding delay to separated sources will always make them less in time somewhere.
Click to expand...
Exactly!

I have tried playing with delay in EASE on various manufacturers products and have yet to find a case in which the delay made it any better.

Sure you can get it "better" at one seat-but it will be worse for all the others.

So I guess it depends on what is most important-a single seat or the "general public".
 
Top Bottom