My experience adding subwoofers to 2 channel

     I thought it would be useful to cut and paste a response from Duke Lejeune, owner of Audio Kinesis and advocate of the DBA system concept, from a thread I started awhile ago on the DBA concept. Here's a link to the full thread followed by Duke's relevant response:

https://forum.audiogon.com/discussions/anyone-else-using-an-audio-kinesis-swarm-or-debra-distributed...

Duke:

"Just for the record, I got the distributed multisub idea from Earl Geddes. It happened like this:

I was driving Earl to the airport after CES in January of 2006 and we were stopped at a stoplight. He said to me "Duke, I’ve figured out how to get good bass in a small room. Use four small subs and distribute them asymmetrically. Each will produce a different room-interaction peak-and-dip pattern, and the sum of the four dissimilar peak-and-dip patterns will be much smoother than any one alone." A lightbulb went off in my head as I immediately realized this was "the answer", so I said, "Can I license that idea from you?" He replied, "You can just use it." And then the light changed. It was that fast.

Now let me explain what I meant by "the answer": For years I had been trying, off and on, to build a subwoofer that was "fast enough" to keep up with Quads and Maggies. I was a SoundLab dealer, and the SoundLabs are big enough to have good low-end extension, but they are too big and expensive for most people. I figured the first guy to make a sufficiently "fast" sub would have a potential market among Quad and Maggie owners. Being a longtime amateur speaker builder, I tried pretty much everything except for a horn: Sealed boxes, aperiodics, low-tuned vented boxes, transmission lines of many different geometries, equalized dipoles, and isobarics. None of them could "keep up" with Maggies and Quads. I was stumped.

But the problem was not the "speed" of the subwoofer - it was the peak-and-dip pattern that the room inevitably superimposed on the sub, which totally dominated the result. Yes some subs were audible improvements over others, but none could blend with Maggies or Quads without there being a distracting disconnect between the subjective lack of "speed" in the low bass and the rest of the spectrum.

You see, woofer(s) + room = a minimum phase system at low frequencies. What this means is, the frequency response and the time-domain response track one another. In English: Where you have an in-room response PEAK is where the decay is SLOW, and it doesn’t matter how "fast" the woofer is. The good news is that, when you fix the frequency response you ALSO fix the time-domain response, and vice-versa! So, SMOOTH bass is FAST bass. This is why the lightbulb went off in my head when Earl described his idea to me.

A distributed multisub system like the Debra or Swarm or your own set of four subs results in four (inevitably nasty) peak-and-dip patterns, but their sum will be much smoother than any one of them. Actually you will end up with MORE peaks and dips, which will be much smaller and much closer together. This "closer together" part is quite beneficial - the ear tends to average out peaks and dips that are within 1/3 octave of one another. So the subjective improvement is often greater than one would expect from merely eyeballing before-and-after curves.

One of the things to be aware of if you’re going to "roll your own" distributed multisub system is this: At the upper end of the bass region the outputs of the subs will be combining in semi-random phase, but at the bottom end of the bass spectrum their outputs may well be combining in-phase or nearly so, if the longest room dimension is a small enough fraction of a wavelength. This results in a rising response as we go down in frequency (and remember that the time-domain response tracks the frequency response). So if you already have two subs that go real deep, you might do better by adding two more that do not go as deep. Or vice-versa.

In the Debra and Swarm systems, the subs’ native response curve falls by about 3 dB per octave from 80 Hz down to 20 Hz. This approximately compensates for typical room gain. Then in most rooms we reverse the polarity of one of the subs (usually the one farthest from the main speakers), and this extends that semi-random-phase addition down into the bottom of the bass region, further smoothing the response and offsetting the rising bottom end that we would otherwise have. Also the down-firing ports on the Debra and Swarm modules are pluggable, converting the enclosure into a low-tuned sealed box, to give further adjustability (given that every room is different). And finally, the amp that we use has a single band of parametric EQ, in case there is still a peak at some frequency.

Speaking of EQ, why not use a single sub with EQ to fix the frequency response curve, simultaneously fixing the time domain response? EQ works great at a single listening position, and can work well within a small listening area. But the larger the listening area that we try to improve with EQ, the smaller the improvements are going to be. This is because the afore-mentioned room-induced peak-and-dip pattern changes dramatically as we change locations within the room ,so EQing a single subs works best for a single location. Unfortunately when we fix the frequency response in one location, we are making it MUCH WORSE in other locations! This is because the peaks and dips are at different frequencies at other locations. But with a distributed multi-sub system, the spatial variation (difference in frequency response from one location to another) is also greatly reduced. So if by chance there is still a peak with a distributed multi-sub system, chances are it’s a GLOBAL (room-wide) rather than a LOCAL peak, and therefore is a very good candidate for correction via equalization.

Big rooms have smoother bass than small rooms because they have a more dense modal behavior: The room-induced peaks and dips are more numerous and more closely spaced. A distributed multisub system can make a small room mimic a larger room in the bass region by doing the same thing. And this will sound counter-intuitive, but the smaller the room, the more it will benefit from a distributed multi-sub system.

One other bit of trivia: The other main researcher behind distributed multisub systems is Todd Welti of Harmon International. He investigated symmetrical rather than asymmetrical arrangements. Anyway Earl and Todd were developing their ideas AT EXACTLY THE SAME TIME, but each was completely unaware of the other’s work. So I guess it was an idea whose time had come!

Robert E. Greene of The Absolute Sound gave the Swarm a Golden Ear award again this year (It got a Product of the Year award back in 2015). So despite the passage of time, apparently the distributed multisub idea hasn’t become outdated.

Duke

dealer/manufacturer/caught in limbo?"

      I think this explains why the time-domain is much less important on deep bass response than some seem to believe it to be.  It also explains the importance of how our brains process the presence of multiple bass peaks and dips at multiple frequencies in a room, referred to as psycho-acoustics, which is a bit counterintuitive but is a crucial factor in how a DBA so successfully utilizes the differentiation between the physics of how deep bass very long soundwaves behave in a given room and how we perceive those soundwaves.  

     I admit I don't understand this subject in detail but I can state with certainty that the concept produces what I consider state of the art bass response in my system and room without the use of any room correction software or hardware, DSP, equalization and bass room treatments.  

     I know I'll always have this excellent deep bass foundation that seamlessly integrates with not only my current main speakers on music and ht, but it will do so equally well with any main speakers I may use in the future.  I've found my huckleberry.

Tim