Hi!
In (most of) my recordings I use a technique called near-coincident (mixed stereophony) that records both time arrival differences (phase) and intensity differences. Usually phase differences are not problematic until you try to mix down both the channels to mono (often done for radio transmissions). Any real organ builder will tell you that the pipes in a real organ might have phase issues with each other, there's even music pieces written (for instance by Bengt Hambraeus - Constellations/Interferences) that exploit the phenomena. It's certainly nothing to be scared to death of, nor is it a goal to only use the x-y stereo recording technique to avoid the phase differences. It actually creates much more space and depth in the sound to record as I do.
As Martin explained, there shouldn't be any problems specifically with phase as GO now handles reverb (a mono (in effect) ir-recording is equally added to both the channels), but what does happen is that the ir-reverb "colours" the sound and will amplify certain frequencies and suppress others. Also my ir-recordings are different from each other because balloon popping is really a bit random in how it will excite the room (and it's certainly not flat over all frequencies) from one pop to the other, but I think they are usable for different effects.
I can only agree that for the time being this implementation is good enough, but eventually we'll want to add the multiple channels of the recorded ir too. I think that it would be nice to be able to decide how the reverb should be routed too. For instance one could want to have front speakers output the original recorded audio while surround speakers (or back) would output the convolution reverb sound only... Just a thought. Of course, in theory it could be multiple channels of recorded audio that (perhaps only some) will be processed by the reverb into their own audio streams that could be routed to any number of different audio devices (different) output speakers. It quickly gets complicated, but for most common use cases it will rarely be. Many will be happy to just process the normal recorded stereo audio with respecive channel of the stereo impulse response reverb that then will get mixed together and enjoyed over simple stereo speakers connected to a single audio device... Besides, there are no free multi-channel recorded samples available (except the normal stereo of course, yet...) compiled into a GO sampleset, nor does GO support reading such files (again, yet).
Kind regards
Lars P