Can I use DRs/OTops/Jacks in a cluster array?

Frequently asked questions and the answers to them.

Moderator: Bill Fitzmaurice

User avatar
Bill Fitzmaurice
Site Admin
Posts: 27834
Joined: Tue May 02, 2006 5:59 pm

Can I use DRs/OTops/Jacks in a cluster array?

#1 Post by Bill Fitzmaurice » Sun Oct 29, 2006 2:24 pm

Yes, but you shouldn't. So-called cluster arrays put speakers side by side, the idea being that wide placement gives wide dispersion. 'Trap' arrays splay the cabs outward, ostensibly for even wider dispersion. It doesn't work, and here's why:

In the diagram below A and B are listening to a cluster array. Listener B is on center to the two cabs, so the ten foot path from both cabs to him is the same. No worries. Listener A is off to one side, twelve feet from one cab, fourteen feet from the other. That two foot difference is a half-wavelength at 280 Hz, so at that frequency the two signals will cancel each other out. Two feet is one wavelength at 565 Hz, so at that frequency the two signals will perfectly combine. In between the two extremes the two signals will mix at various degrees of phase, and response will be all over the place. This is called Comb Filtering, so named because a response chart looks like the teeth of a comb.

What's worse is that as listener A moves the difference in the path lengths between he and the two speakers changes, so response is different everywhere he goes.
This shows what happens with a single frequency when you have a single cab, two cabs side by side and three cabs side by side:

With one cab the result is uniform as you go across the sound field. With two cabs there are hot and cold spots, with three cabs it's even worse than two.

If you absolutely must have speakers horizontally placed crossfire them, as shown in the next diagram.

Here the path differential is shorter, which pushes the problem frequency bandwidth up, where it's less obvious. Even more significant, as listener A moves farther to the side he moves out of the coverage zone from the one cab and into the coverage zone of the other. This reduces comb filtering to a minimum, and also smooths the on-axis versus off-axis response of the system. This is the exact technique used in the cross-firing tweeter arrays of the DRs, Jacks and OmniTops to deliver wide dispersion and smooth response.

But the best way to eliminate comb-filtering is to stack cabinets vertically. That way no matter where one is the path distance to the individual speakers in the array remains constant. Just as significant, vertical arraying reduces the vertical dispersion angle, sending less of your sound down at the floor and up at the ceiling, and more into the audience. If you need wider dispersion than stacked cabs provide when all aimed the same direction rotate them, creating a spiral array. This preserves the narrowest possible vertical axis.

Just as significant as the comb filtering issue in the highs is dispersion in the mids. In the midrange where side by side cabs are close enough together to couple, acting as a single source, two cabs side by side will have half the horizontal dispersion as two cabs vertically stacked. In short, there are many reasons why cabs should be vertically stacked, and none to support horizontal placement.