Why the fascination with subwoofers?

@rauliruegas wrote: "we not necessarily need 4 subs we can do it with two true subs. The Harman white papers proves that.... My target is to have the best solution ( rigth now with two true subs. ) at one and only one seat position where the mids and higs are spot on."

You can get good bass in one sweet spot with two equalized subs.  If you want good bass over a significantly larger area, four subs intelligently distributed can make a worthwhile improvement. 

(Note that Todd Welti et al did not investigate asymmetrical sub placements, and they also made the assumption that rooms are acoustically symmetrical at low frequencies.  This totally made sense for the paper they were writing.  But in practice rooms are never symmetrical low frequencies because unless the room has no doors or windows or AC vents or other "soft spots" in the room surfaces.   Even a heavy recording studio door is a significant "soft spot", changing the room's effective length in that dimension at low frequencies.) 

Rauliruegas also said, "Btw, I take a look in your site and I don’t find out which are the 3 +,- db points in those 10" units. Could you share it?. Appreciated."

The response of the individual Swarm units is the approximate inverse of "typical" room gain.  "Typical room" gain has been given by several authors as being about 3 dB per octave below 100 Hz or so, so my response curve is -3 dB per octave from about 100 Hz down to about 20 Hz, and then below 20 Hz the rolloff accelerates rapidly.

In other words I look at subs + room as a system, and my target is all about what the system does, not what the part I make does all by itself.  The Swarm system is highly adjustable to work well with a wide variety of rooms.   Ports can be plugged, polarities reversed and/or phase manipulated, and the amp includes a single band of parametric EQ along with a +3 dB "bass boost" switch.  In practice, the output of the four distributed subs combines in semi-random phase at the upper end of the bass region, gradually transitioning to approach in-phase at the bottom end of the bass region as the wavelengths become long relative to the room dimensions.  In most cases it is beneficial to offset this additional gain as we go down in frequency by reversing the polarity of one of the subs, or if we are using two amps, by manipulating the phase of one amp relative to the other. 

If I were to tell you that my subs are "-3 dB at such-and-such Hz", none of the above information would be conveyed by that spec.   Unfortunately people compare subs based on who has the lowest -3 dB spec, and therefore manufacturers compete on the same basis, and what REALLY MATTERS (which is, what happens when you put the sub into a room) is not given much if any consideration.   

Duke

@rauliruegas wrote, about the AudioKinesis Swarm:

"For a 10" woofer (as the four subs array posted here. ) is almost imposible to handled frequencies below 20hz at over 100dbs ( SPL. ) and with low THD kind of distortion."

I didn’t design the Swarm to go below 20 Hz, but several customers who have measured the in-room response report -3 dB in the upper teens. For deepest loudest bass at the same price point (three grand for the system), a single monster sub is the way to go. I have designed a Swarm that can do well over 120 dB at 13 Hz, but it’s not very practical and I don’t think there would be much market for it.

Each individual 10" Swarm unit can do about 100 dB at 20 Hz in-room at one meter without approaching x-max or amplifier clipping, assuming "typical" boundary reinforcement. I don’t know what the THD number would be, as imo that’s not a problem that needs solving.  The in-room frequency response is of vastly greater audible consequence as long as neither woofer nor amp is driven beyond its linear limits.

* * * *

My fascination with subs dates back about twenty-five years, to my simultaneous fascination with Quads and music with bass.

Duke

Hi @lewinskih01, you asked, "What is your point of view about time alignment of a SWARM or other DBA systems?"

Imo time alignment is at best a secondary consideration in a distributed multi-sub system, from a perceptual standpoint.

The ear has very poor time-domain resolution at low frequencies, and you are aware of Geddes’ thinking on the subject (which is based on AES papers). On the other hand the ear is very good at hearing loudness differences at low frequencies once the low frequencies become loud enough to be audible. This is why equal-loudness curves bunch up south of 100 Hz. A 5 dB difference at 40 Hz can be perceptually as big a difference as a 10 dB change at 1 kHz! The implication is that getting the in-room frequency response right matters more.

Also, since speakers + room = a minimum phase system at low frequencies, when we fix the frequency domain we have also fixed the time domain, and vice-versa.

That being said, imo you bring up something which intuitively makes sense: Preserving the initial pulse of bass energy, that which "whaps" you. It would seem that precise alignment of the arrival of the energy from multiple subs would best support that initial impulse, but how precise is "precise"? Within 1/4 wavelength? According to a paper I read, the ear cannot even detect the presence of bass energy from less than one wavelength, so the "precision" required might not be as great as our intuition would lead us to believe.

This is just anecdotal, but every time I have reversed the polarity of one of the four subs in a Swarm system, there as been a subjective improvement, despite the fact that the initial pulse has been obviously degraded.

This past October at RMAF an industry veteran manufacturer with decades of experience came into our room and played his reference recording of Fanfare for the Common Man. He said, "that’s what a tympani sounds like." He went on to say that our system (in a normal hotel room) did the best he had yet heard on that recording. We were using two amps and had manipulated the phase of the two left-side subs relative to the two right-side subs.

Now it is theoretically possible to use four time-aligned and equalized channels of amplification and achieve precise time alignment and excellent in-room response smoothness simultaneously, and this would probably be even better. But at the price point I’m working, focusing on room-interaction related issues seems to give good return on investment.

Once we relax cost constraints, it might make more sense to build a planar array into the front wall and a corresponding array into the rear wall, reverse the polarity of the rear wall array, and time-delay it such that it cancels the signal from the front wall when it arrives.

Duke

@noble100 wrote: "But my main reason for posting is to ask about your quote above.
Was " and time-delay it such that it cancels the signal from the front wall when it arrives." accurate and not a typo?"

That’s what I meant to write.

My understanding is that the idea is to arrange four subs in a centered, half-the wall’s-scale pattern on the front wall. So if the front wall is 18 feet wide by 8 feet tall, then the subs would be at the corners of a rectangle 9 feet wide by 4 feet tall, centered on the middle of the wall. This mimics the best four-sub geometry Todd Welt found in his study of symmetrical sub arrays, but it’s on the wall instead of on the floor. If the wall reflections cooperate, they should cancel out standing waves in the vertical and side-to-side dimensions.

To cancel out standing waves in the front-to-back dimension, the idea is to use active cancellation based on an identical array on the rear wall. This array is in reverse polarity, with a time delay that corresponds to the length of the room.

So the subs in the front of the room to create a planar wave that moves to the back of the room and then disappears as the active cancellation array cancels it when it gets to the back wall. The "no bass" you mentioned happens at the rear wall, not out in the room, or so the theory goes.

I do not know how well this would work in practice. The answer is probably some variation on "it depends". Nor do I know whether it would sound better than a distributed multi-sub system. I haven’t really investigated it and will never build anything like it myself, but it MIGHT be a competitive "cost no object" approach.

Duke