This is a great topic.
Similarly, this has been discussed here
Since that discussion, I've experimented with Audyssey XT32. It can be unpredictable, but it is nice, really nice. However, I prefer having manual control especially in the time domain w/varying sub type/location. My cheapo, $99 MIC2200 bested the XT32 in optimizing and management between my 4x18 IB, and (2)2x15 small sealed. I've since ordered a custom 4x8 dsp piece, for added processing capability.
I don't care for XT32 speculating on above the transition
EQ'ing, which is real touchy proposition in the first place. As others have stated, the multi-channel alignment between the surround, LCRs etc, is great. I couldn't get it to properly align and subsequently sum the two sub-systems ideally in the bottom octaves. The post Audyssey FR was respectable, however I was suspect. Upon further examination and experimentation, I discovered turning off my IB actually raised the output
I've since discovered regardless how many XT32 passes, this mis-alignment in the time domain occurs. I've got several options for subwoofer EQing (multiple parametrics, Velo SMS-1, some pro EQs and others), but none as powerful as the 4x8 custom piece I've ordered. It'll enable me to manage up to five separate sub outs, with full manipulation in both time and freq domain. My plan is to possibly integrate a tactile system as well, and the processor would certainly help make it more fun.
I plan to use the other 3 input channels for manipulating the LCRs below the transition freq, as long as the transparency isn't affected (I've got no reason to think it'd be an issue). I'm really looking forward to the essentially unlimited flexibility and capability of the system.
Audyssey is amazing and has brought a more enjoyable experience to untold numbers of enthusiasts. Systems of the highest caliber can benefit from the XT32 processing. I'm looking for even more control and flexibility. Room acoustics is where one should really focus one's attention. It is oft stated, and many realize the fact. But it's hard to have too much LF damping/bass trapping in most every HT room. If too much critical MF/HF energy is potentially removed, there's simple steps taken to mitigate that loss yet maintain the damping needed. Yes, XT32 is fantastically helpful, but it's no replacement for well understood, best practice acoustic treatment techniques and approaches that aren't necessarily an expensive proposition.
I recently discussed this very issue with Mark Seaton. Mark suggested that the 4x8 processor was fine, but perhaps the ideal solution was also allowing XT32 to do it's thing, and carefully examining it's influence w/measurments. Perhaps utilizing Audyssey in a single input channel config of the DSP processor, could be helpful as an approach toward my end goal means to an end.
Good question, great topic.