Earl, [snip] I can now put each [subwoofer] driver anywhere I like. So I could have 4 separate subs firing down from my ceiling ... From what I've read of the multiple sub theory, the more the better and you only recommend 3 as a pragmatic solution and because you believe there are only small incremental benefits by going to more than 3 subs. Am I right in saying that and if so would you agree that having the 4 separate drivers dispersed across the ceiling should be better than mounting them and wiring them in pairs? In case its important, all these drivers are backing into the same void/roof space.
Just saw this interesting thread. At CES Earl, Pat Turnmire and I talked briefly about the approach my team developed to provide physics-based answers to exactly this kind of question. Strikes me that at least a few readers here might be interested as well. To keep it brief, I'll stick to the most basic questions, "How many subs, and where, exactly?", and ignore other relevant ones like "How should the delay settings, filters, gains, polarity etc. be set on individual subs?"
Goal is to model the actual physics of what's happening so the optimum subwfr config (quantity & locations) and LF acoustic treatments can be found for reducing the severity of seat-to-seat response variations. We can do simple optimization, where we assume all subs will play the same signal at the same time (as if they were all paralleled off the same virtual amp), or we can include in the optimization process any number of electroacoustic features, like individual sub settings for delay, shading (different output level settings), EQ, polarity, etc.
In crude outline, our process runs like this:
1: Create 3D CAD model of the room, including all the "geometric inconveniences" that simple mode calculators can't deal with, e.g. L-shape room plans, openings into hallways etc., alcoves, niches, doors, windows, seating platforms, soffits, sloping/cathedral ceilings, pony walls, big fluffy sofas, and so on.
2: Tag all locations in room where a pressure source (subwfr) can be physically accommodated and aesthetically tolerated, and set a max practical limit on the number of subs the client would consider buying (say, 3-6). (Side note: The commercial availability of high-quality enclosed, in-wall subs -- Wisdom S90i, Genelec 5041A, JL Audio Fathom IWS, and JBL Synthesis S4S spring to mind -- makes it a LOT easier to avoid lo-fi bass afflicting most rooms because they make it possible to drive the room's pressure map cleanly from multiple locations.)
3: Bring the model into an FEA (finite-element) package, feature/de-feature it as needed, mesh it and assign physical properties (Young's modulus, Poisson's ratio, acoustic impedance, etc.) to the walls, floors, ceilings, doors, etc.
4: Transfer the model into a CFD (computational fluid dynamics) package and create the resulting pressure plot throughout the listening zone at all frequencies below the Schroeder freq in whatever resolution (e.g. 15-120Hz in 0.1Hz increments).
5: Bring the results into mathematical analysis package running optimization routines to find the highest-performing sub configuration -- for example, 5 subs total, with 1 each in locations 7, 12, 23, 29 and 34.
6: Since we know WHERE in space all the antinodes are of all modes (axial, tangential & oblique), we can model the efficacy of specific LF treatments to further reduce standing wave ratios and increase spatio-pressure uniformity.
All of this is before we show up with the DSP/equalizer. In our experience, there's no substitute for getting the basic physics right!
--Keith
