[argyllcms] Re: High quality printer profile?

  • From: Yves Gauvreau <gauvreauyves@xxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Mon, 21 Mar 2022 06:20:47 -0400


On 3/20/2022 10:15 PM, Ben Goren wrote:

On Mar 20, 2022, at 1:45 PM, Yves Gauvreau <gauvreauyves@xxxxxxxxx> wrote:

Do you do only the self calibration and or you do an Argyll calibration on top 
of that?
The point of either calibration is to linearize the printer, and Argyll’s is 
only there for printers that can’t do this themselves (or that don’t do it 
well). The big Canon (and Epson and HP) printers are pretty darned good out of 
the factory, and their built-in linearization routines are really good. Good 
enough such that, once you’ve got your profile built for a particular paper, 
you never have to do anything other than re-linearize every now and again 
(maybe but generally not after changing inks, certainly after changing 
printheads). You should have zero need to make a new profile in the future, 
unless the paper changes (etc.)
I've continued to read about this calibration stuff and though I'll have to trust that the printer can self calibrate correctly for now, in the specific case of my printer I installed the am1x file and I did the calibration for this paper specifically. This seem a much better way, I think. I'll try to see if I can create and or at least verify the calibration. That would be even better I would think.

I wonder if I should try to make a special version profile for highly saturated 
colors by supplying either a gamma for a specific image or a set of image with 
similar colors?
The challenge of the profile isn’t attaining the extremes of the gamut so much 
as it is mapping the interior shape. The primaries establish the perimeter, and 
there really isn’t anything a profile can do to expand beyond that perimeter. 
Plus, the borders of the gamut where the saturated colors lie aren’t 
necessarily any more prone to inherent weirdness than the neutral axis.
Funny you say that about the neutral axis but I get the idea, if you want to put emphasis on some region of the gamut, give it more patches to work with and all should be fine.
You’d really only need to optimize especially saturated colors if the printer 
is poorly linearized in that region, and you’d know this if you had excessive 
“muddiness” or loss of detail in those colors — or if profcheck reported high 
Delta-E values for saturated patches of colors you particularly cared about. 
Even then, in general, the solution is to throw more patches at the profiling 
engine.

I don't know the kind of printing you guys do, but capturing (and printing) the colors of flowers on a bright sunshiny day, that spells troubles 99.9% of the time and it's the main reason I got interested in studying color management. Muddiness, lost of details, that's the story with flower photography. I've learn to use a diffuser and other mean but still.

I'm more interested in printing because the papers I use have a gamut that protrude out of the Adobe RGB space in some region, so I find myself using the paper that is best for this or that flower shot. All this without clipping of course. If it's to much for the paper I use, then I want the best gamut mapping I can get, I even use the image dependent gamut mapping strategy we can find on the scenario page.

I understand quite a few photograph don't need special treatment like this, but mine do.

Many thanks for your time and effort.

Yves


So, in general, I wouldn’t worry about “leaving something on the table” or the 
like. You should be on the path to great prints with minimal fuss by basically 
following the well-trod path.

b&

Other related posts:

  1. Home
  2. argyllcms
  3. Archive
  4. 03-2022