[argyllcms] Re: Linearization
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Sun, 21 Jun 2009 17:59:03 +1000
Roger wrote:
What I'm trying to find is can something similar be done *without* a linear
press ICC profile? Based on just four cyan, magenta and yellow step wedges,
from 0% to 100%. My hypothesis is, if, through 1D curves, it is possible to
colorimetrically match the individual CMYK channels, or within a very close
tolerance, THEN in my view, the Lab values in the Reference print conditions
in the rest of that colorspace ought to match also. Since the two sets of
colors are linearly related. That is my hypothesis.
You would need a process specific mixing model to estimate
ink combination values from pure colorant readings. Spectral
readings may give more to go on, but the result is likely
to have noticeable inaccuracies. Various mixing models
have been suggested in the literature, but they all have
parameters that would need tuning to the type of process
involved, and validates as having sufficient accuracy.
Graeme Gill.
Other related posts: