Well, that's impressive.
Anyone who uses a hi-resolution film scanner (expecially Nikon SuperCoolscans and Minolta 5400) knows that even slightly curled film produces bad scans, with either the center or the borders out of focus.
Those scanners have very thin DOF and it shows, even for apparently flat originals.
Up to now, there was no cure except glass holders when available (which are a pain) and/or wet mounting. Most of us adopt a compromise, focusing half-way between the center and the border.
Enter Helicon Focus.
I scanned 6 times my test slide (35mm on Minolta 5400) , with focus from optimal in the center to optimal in the borders (I manually stepped the focus), and combined the scans (35+ MP each) into the final image with Radius = 3 and Smoothing = 2.
Results (100% crops, heavy unsharp mask to better show the differences). Please forget the wrong label: where you read "Lower-Right" it's really "Lower-Left":
I think they pretty much speak for themselves.
Yes the HF image is slightly blurred compared to the optimal scans in the two regions, but it's consistent from border to border!
That's impressive in my book. And maybe I could tweak HF parameters a bit more (R=2 S=1 maybe?).
BTW, the elaboration did not have any hiccup; HeliconFocus took up to 1.3 GB during the process (I have a 3 GB machine running XP 32-bit).
Fernando