[argyllcms] Re: What colorchart I should buy?
- From: Klaus Karcher <lists@xxxxxxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Fri, 01 Jan 2010 19:39:16 +0100
Klaus Karcher wrote:
IMHO, they are useless for users that have not spectrophotometers
can you explain why you think they are useless?
Konstantin Pyanzin wrote:
There are obvious reasons, I think:
1. An X-Rite ColorChecker SG is not supplied with individually measured
reference file.
With regard to the apparently tight manufacturing tolerances for the
CCSG, I don't expect /individually/ measured reference files (it's
questionable whether they had any advantages in this case IMHO) -- but
of course CCSG should come along with the reference file or at least a
hint where to get it (without downloading and installing a huge
dispensable software package) -- but finally Xrite took heed of our
moaning and seems to be willing to resolve this grievance.
2. There are several different "general" reference files from X-Rite. With
different values.
... and measured with different devices. From my experience, the 2005
ProfileMaker reference file is pretty reliable. Probably more reliable
than many do-it-yourself measurements.
[...] Having spectrophotometer a user does not suffer from such limitations.
I've seen too many inaccurate instruments in the field to agree to this
statement.
for example, my own ColorChecker SG
differs from both reference files. The difference between my
ColorChecker SG values (measured by two i1Pro spectros) and the
ProfileMaker reference values reaches dE94=3.6; between my
ColorChecker SG measured values and the i1Match reference values
reaches dE94=3.8
DE94 max I guess. What are the DE94/DE00 avg values?
Yes, you are right, they are max dE94. Avg dE94/dE00 are 0.67/0.61 (for
ProfileMaker) and 1.06/0.91 (for i1 Match).
So also your measurements confirm that the 2005 ProfileMaker reference
file goes better with your measurement values than the i1 Match file.
An avg/max DE94 of 0.67/3.6 is largely the ballpark of the usual inter
instrument differences (note that the instrument data sheets specify
values for only few BRCA tiles measured under laboratory conditions)
I daresay it can be even helpful to use an CCSG as "poor man's BRCA tile
set" -- and in case of discrepancies I'd rather start troubleshooting my
measurements and the instrument than the ColorCheker ;-)
Klaus
Other related posts:
