Hi Dave, 

For an out-of-the-box option it's worth taking a look at beastMulch, although I find it an overkill (it's mostly vbap) and too complicated for a smaller system, so I've never used it beyond experimentation at home. 

I would suggest a multichannel interface with a real analogue mixer. If you get a mixer with direct outs and a few auxes then you can just route the signals into the desk and then use the direct outs and some of the auxes to feed the speakers. Often I patch together two mixing desks to get more channels. 

Another option that I've explored a bit is using midi faders (the beringer works best I think). But then you need to do all the processing in pd or max (or Csound I guess, although it would be more cumbersome). In this case I use Ardour to play the track/s and then, using Jack, I route the tracks into pd or max. I've had to do a lot of messing about to get a high resolution and smooth control signal from the midi faders though, and they're never as responsive as potentiometers on an analogue desk, and if something goes wrong it's more difficult to debug (say, in the middle of a gig). So I'd still go with the analogue mixer set-up. 

You can, of course, have your own OSC faders built (this is what they did at birmingham and manchester), but it's too expensive. Having said these, I encourage you to actually just get a 6 channel set-up (I'll be happy to go into more detail about the philosophy of physically arranging the speaker array) and actually composing the piece in multichannel. This gives you so much more control and it's actually more economical and efficient in the long run. You can get a really good spatial feel with just 8 speakers in the concert and almost no diffusion, save the odd balance tweaking. And if you have more than 8 speakers, doing some play with the sense of enclosure and distance (by cross fading between closer and more distant speakers). I use 6 channels because front/back amplitude panning doesn't do what it's supposed to do (because we only have one ear on the sides (!) you don't get a stereo image on the sides. So I find that you can only use the front/back configuration to disperse the image (widen it), making it more fluid. Source localisation never works between front and rear and can only be psychologically suggested. So I don't see the point of using side speakers, which is what you get with a conventional 8 channel ring. I do, however, double up the rear signals into a pair of speakers on the sides (but pointing them more towards the stage, not directly into the listener's ears, and positioned slightly higher to avoid any direct sound). This is needed so that the listeners sitting near the front of the auditorium get a rear image: otherwise the rear speakers will be too far away from them and they won't have any sense of engulfment or whatever the technical term for it is. To me it seems that the art of sound projection is largely to do with making the speakers as invisible as possible, thus creating a sense of space that transcends the geometry of the auditorium.       

I would go as far as to suggest that 'conventional' diffusion was born before even stereo production was invented, and that it's rather obsolete now. There is no need for it unless you're performing a stereo piece. But if you have a multichannel system in your studio then why compose in stereo!?  As much as I love diffusion, I class more as a sort of period performance for some of those great classics. And trust me, once you move to multichannel, stereo just feels crude, other than for home listening!

As for EQing, sometimes, depending on the space and the system you get, you need to EQ the piece. I only use the desk EQs when I have smaller, not so powerful speakers that aren't full range (like the 8030 genelecs), because certain lower frequencies introduce nasty sound characteristics in those speakers at higher amplitudes. And I then use the subs to carry most if the bass. But this is not ideal. I prefer having no subs, and full range speakers. I would never add reverb but delaying can be useful sometimes (in the region of 20-40 ms). I often delay the rear speakers by 25ms, but I do this by just nudging the audio region in pro tools or ardour. In a really large reverbrant space you need more speakers, including some near-field ones to get a good sense of spatial articulation. Otherwise all you get is a wishy washy reverb and almost no ability to do anything interesting with the piece.       

I'll be happy to answer any specific questions about diffusion techniques and set-ups, as well as anything related to multichannel composing. 

Hope that helps.

Peiman