>Actually, it's pretty much impossible to build a profile for >PC lenses for the reason you stated. All they can profile is >how a lens behaves throughout it's image circle. With a PC >lens the image captured is a smaller area within a larger >image circle. Shifting the lens moves the captured image >rectangle around within the lens's image circle. In order to >correctly profile the result DXO would have to know where the >rectangle is within the lens's image circle - and >shift/rotation information is not available. With tilt the >situation gets even worse. When the lens is tilted the area >of capture is no longer rectangular within the lens's image >circle - it may be trapezoidal, or some other skewed shape >depending on rotation angle and shift. The math to calculate >all this is indeed daunting - but not impossible. Without >information about shift amount, tilt amount, and angles of >rotation, it's essentially impossible. Until shift >amount/direction, and tilt amount/direction EXIF data becomes >available with PC lenses the problem will remain unsolvable >(at least automatically).
I guess you are essentially right: it may really be the technical complexity which keeps DxO from making profiles for tilt-shift lenses. Still, what a pity: in a way, some of the most interesting lenses which play to the craft of photography - and of which you can make so much use on that wonderful d800 sensor - get sidelined.
Just one point: with regard to the tilt amount, I am no certain the situation is as you describe it: assuming the tilt is correctly applied according to the Scheimpflug rule, the plane of sharp focus gets placed other than parallel to the sensor over the motive area, but that should not matter for the sensor, just as it cannot matter for any lens on which part of the image you focus - it may be just that chap sitting in the lower left corner.
I don't know, perhaps a company like DxO should settle for something less than quite perfect in this regard, just in order to promote the craft of photography. The tilt and the shift give so much creative power - sure, you have post-exposure perspective control, focus stacking and all, but that (in my view at least) complements rather than substitutes these old instruments of image control which they all used, Edward Steichen, Alfred Stieglitz, the group f/64, Walker Evans, Edward Curtis, Frederick Evans, Eugene Atget...I apologize if this rant bothers anyone.