[argyllcms] Re: Perceptual intent black problems

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Fri, 30 May 2008 20:09:35 +1000

Mihai Zahorski wrote:
I am trying to make a profile for Epson SP4800 ( Luster 300g paper ) and i
have big problems with printing shadows using perceptual intent.
I used Adobe1998 as source profile for gamut mapping and I loose details in
the shadow areas because of the black generation.

If this is specific to perceptual, then I doubt it is due to black generation,
since the separation will be common to all the intents in an ICC profile.

I am attaching two screenshots of the test image displayed with the black
channel turned off :
-the first is one is using a profile generated with black limit 90
-the second one is using a profile generated with black limit 70.
The other parameters used are: 240 total ink limit, 0-100 black generation
with 0.5 shape ( I didn't want to 'hide' the black generation ).

Sorry, I'm not at all understanding what you are hoping to illustrate.

For me it seems there is something wrong in the black generation, in how it
substracts the levels of CMY.

What leads you to think that ?

How can I control better the black generation?
I only want a smooth curve that would not destroy the details.
Using Gretag ProfileMaker showed a much more moderate black generation
although I have used the same parameters. I also saw the following controls
in the separation page: "Black width" and "Neutralize". Are there similar
functions in Argyll?

There is no equivalent currently to "width" in Argyll, the black generation
is always "wide". You control the level of black using the -k parameters,
as illustrated here <http://www.argyllcms.com/doc8/Scenarios.html#PP6>. This
control is not as easy to use as perhaps it should be, and a number of runs
using xicclu is generally require to arrive at a reasonable curve.

It's hard to comment any more than this without the .ti3 file, and the
parameters used to create the profile. Generally the black generation
doesn't change the color at all (it's simply choosing between all the
combinations of CMYK that can generate a particular color), but it
affects secondary aspects, which can show up as color artifacts indirectly.

Graeme Gill.

Other related posts: