[argyllcms] Re: Camera profiling with bracketing

  • From: Iliah Borg <ib@xxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Sun, 13 Mar 2011 08:17:34 -0400

Dear Nikolay,

> Iliah, did You check the diag.tif (-dipn option) in both cases? What the 
> difference is?
> 

For scanin - yes, I checked it. For manual ROI selection there is no such thing 
as diag - but one has the benefit of WISIWIG. So, the difference is that while 
selecting manually I voided more of a border zone where shadows are cast.


> I've prepared custom *.cht file for my target with smaller sampling areas for 
> the patches (wider spacers between them) and I have no problem with X-Rite 
> Digital Color Checker SG even without a tripod. The sampling area I've 
> choosed is 50% from whole area of each path, so the x/y dimensions of 
> sampling area are about 29% smaller than full patch size.

When I select manually I avoid any chart imperfections - so for autoscanning 
one needs not only to make the ROI smaller, but to filter off deviations based 
on a more discriminating criteria.

--
Iliah Borg
ib@xxxxxxxxxxx




Other related posts: