Re: CMYK to Lab and back
Re: CMYK to Lab and back
- Subject: Re: CMYK to Lab and back
- From: Harold Boll <email@hidden>
- Date: Thu, 31 Jul 2003 10:32:02 -0400
>
Could someone explain something to me please? I have a CMYK file with an
>
embedded ICC profile, if I convert it to Lab in Photoshop colorimetrically
and
>
then convert again back to my original CMYK colourspace again
colorimetrically
>
should the Lab values remain the same at points of the process? MY
>
understanding is that they should, however when I try this the values
change
>
sometimes considerably. Anyone brainy enough to tell me why? Cheers!
In principle, if you do the following series of color transforms, the
starting and ending
Lab values should be the same for either Absolute or Relative Colorimetric
intents.
CMYK1 -> Lab1 -> CMYK2 -> Lab2
When there is a large disparity between Lab1 & Lab2, there are a number of
reasons
why, most of which ultimately boil down to the fact that you're trying to
model non-linear
behavior with linear interpolation in the Profile.
1) Interpolation error due to non-linearities. If the relationship between
ink/color is highly
non-linear, there will be inherent interpolation errors in the 3D table
which computes via
linear interpolation. Inadequate sampling of the color space exacerbates
this error.
2) Interpolation error at the gamut boundary. This type of error is
distinct from the
the non-linear errors described above. Colors near the gamut boundary are
prone
to clipping in the linear 3D grid table of the profile. There are ways by
which this
can be mitigated (something called Extended Range), but not all vendors do
this.
3) Poor modeling of the relationship between color/ink. Its relatively easy
to model
the relationship between color & ink in this direction: CMYK->Lab. For a
variety of
reasons the inverse direction, Lab -> cmyk, is more challenging. Some
vendors
do a better job at this than others.
One point to note, there is no expectation that the CMYK1 & CMYK2 inkings
will be the same
due to Gray Component Replacement relationship between CMY & K.
Hb...
Creo
Harold Boll | 1.978.439.7096 | email@hidden | www.creo.com
IMAGINE CREATE BELIEVE " The sunlights differ, but there is only one
darkness. " Ursula K. Leguin
_______________________________________________
colorsync-users mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/colorsync-users
Do not post admin requests to the list. They will be ignored.