[argyllcms] Re: Manual channel limits doesn't work in printcal for RGB-printer

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Wed, 16 Dec 2009 23:34:55 +1100

Nikolay Pokhilchenko wrote:
In RC2 The problem is still here. Even with 100% user channel limits in 
printcal, the
printtarg result didn't translate R=B=G=0 to zeros. For example, the RGB=0,0,0 
printtarg with calibration become 0,21,0. It's heavy restriction for 
RGB-printer gamut.

A possible theory to explain this is that the measured data isn't monotonic,
so printcal is being told that the maximum density for a give channel
is at less than 100% device maximum (== device minimum in the case of RGB).

So if instead of forcing device maximum with -x, you reduce it to match the 
response (or let printcal automatically choose a value closer to the maximum)
you will get a darker black.

To put this another way, it's bad to force printcal to use the non-monotonic
region of the device response.

It might be that the non-monotonic response is not real and due to noise,
in which case it might help if I increase the smoothing a bit.

Graeme Gill.

Other related posts: