Originally Posted by Berland
In my opinion the 2D/3D remapping is one of the key features of Trinnov (even for stereo). Only not needed if you have a perfectly symmetric rectangular room (a door on one side will kill this perfect symmetry). Again, I have never ever seen such a room - only close to.
Speaking just for myself (I mean, who else am I speaking for?
) , I see it this way. If you don't have speakers at spot-on angles both physically and electronically, the minimal advantage of remapping is that you're adjusting horizontal (azimuth) and vertical (elevation) positioning for what would be the ideal. In that sense, minor distortions due to room effects and random variation can be be accounted for in the remapping algorithm. Also, as someone that builds statistical models professionally, I like the idea of having a model predict reality when you've got something less than a perfect fit, and something more than having serious placement issues. Plus in theory, using two or more data points, no matter how subtly, to capture the ideal prediction of one data point inside, say, LCR or between a height and a floor layer on the same vertical plane is an advantage. Assuming it works
My own rule of thumb is that you want to stay within the tolerances of your specs of choices (i.e. Dolby Atmos, SMPTE etc.) and if you're within that range, remapping is icing on the cake. Also if you want to do something like avoid focus point effects of speakers at a straight 90 degrees to a row of MLP. With front Ls1/Rs1 and bed side surrounds (or bed and Ls2/Rs2 sides), remapping can solve that issue for you if the phantom image from the remapping is stable for the critical seats (ideally the listening box).
Where this falls apart is if you're too far off angle - think rears at 170 and side surrounds at 90 - or in a multirow room where remapping might produce distortions and hot spots. Or if you get weird effects like mains getting mapped into height speakers.
What it seems to me is that some old school Trinnov users like me, or some folks that came from the older MC or Magnitude units (not Rur from waht I know in his two channel+subs setup, but at least a couple of others) tend to use it more than some of the mass of later adopters with large, more sophisticated personal cinema-type configurations. That may be a case where manual adjustment with something like a Q-Sys Core and what Adam does for his clients may be a good thing over the automated algorithm, and Curt Hoyt did a setup for me with custom array tuning and upmixers that wasn't intended for using remapping. I have heard from multiple calibrators I've talked to that they're trying to avoid using it when possible, for what it's worth.
Basically I think it comes down to try it for yourself. You may be pleasantly surprised. Or not.