[argyllcms] Re: Camera profiling with bracketing

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Fri, 25 Feb 2011 13:43:46 +1100

Rishi Sanyal wrote:
Don't know if this'll help or is relevant, but when I experienced
flare in IT8 target scans due to scanner optics, the flare was
generally dependent on how light the patch to any side of a patch in
question was (the light the patch, the more light that bled onto an
adjacent patch).

Yes, that is to be expected. It's a difficult to avoid, unless you
make a chart with well spread test patches on a matt background.

So what I did, manually, was to take such problematic patches, select
them using the marquee tool, then copy&  paste over the original patch
rotated 90º using the 'darken' blend mode in Photoshop. I did this
also for 180º&  270º for each patch. This got rid of flare on most
patches&  resulted in a profile that showed less clipping of dark
colors (since flare artificially brightened a lot of dark patches,
which the profile tried to compensate for by darkening those colors).

I guess this could be automated by looking at the spread of values
over a single patch, and just choosing the darker values, but it's
equivalent to using smaller patches, with all the disadvantages that
entails.

Another approach might be to try and characterize the inter-patch
flare for a particular chart layout, but this wouldn't account
for all the sources of flare.

Graeme Gill.

Other related posts: