Which is better for a DBA (Swarm); powered subs or unpowered?


I want to start building a swarm (starting with 2 subs), on a budget.  Starting with $1000, am I better off buying two used powered subs, three less expensive used powered subs, or a subwoofer amp (eg Dayton SA1000) and two (less expensive) used unpowered subs?  What is the advantage of having a discrete subwoofer amp?  Room size is 13'x22'. 
128x128cheeg
So, I am the new Swarm(er) noble100 mentioned.

3 ceiling facing subs (3 inches from ceiling) and one on the floor.
10’ x 14’ room with minimal acoustic treatments.

So far, outstanding improvements with Magnepan LRS and Ologe 5 speakers.

Switched over to KEF LS 50s non wireless last night and have been listening to them all day today.

Makes the somewhat diminutive LS 50s sound like true floor standing monitors. As with the other speakers, my listening room seems (sounds?) bigger.

Like they say, "Seeing is believing". So is hearing and listening!

Want to believe it?
Go listen to a system with a properly set up Swarm.
Then listen to that system without using the Swarm.

There's just no going back.


Tympani are typically tuned to about C3 ≈130Hz with a 65Hz sub harmonic. Not really much in the sub range.
While tympani aren't really big on deep bass notes, the rest of this statement is incorrect. Tympani are tuned as needed, not just to C...
@audiokinesis 

"My recollection is that Toole was referring to a Harmon subwoofer integrator processor which optimized the gain, frequency response, lowpass filter, phase, delay, and equalization based on in-room measurements. I assume it does what they claim."  

Duke, thanks for the answer. I think Toole was referring to an integrator, but I don't know it was the integrator process. On chapter 13 "Making (bass) waves - below the transition frequency" he shows several examples where 4 subs are used in a rectangular room and optimized (through an optimizator) to minimize seat to seat SPL variation at 5 seating positions. The optimizator adjusts overall level (dB), delay (ms), and EQ at a given point (frequency, Q, level). It's not clear to me if said delay was introduced through DSP or just a "phase" adjustment.

I can measure with REW and optimize with MSO for example, but if actual delays are needed then I would need a digital processor to incorporate these.