[argyllcms] Re: Camera Profiling using ArgyllCMS
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Thu, 27 Nov 2008 12:20:34 +1100
Pascal de Bruijn wrote:
The following archive should contain everything of interest:
http://files.pcode.nl/temp/argyll/canon_eos400d_testcase.zip
Thanks for that, but it is far from enough information to really
track down the issue. If it's any consolation, this is not the first
time someone has raised this particular issue, but the difficulty
is in tracking down exactly what the situation is, so that the source
of the problem can actually be identified. I was unable to do
so with the previous report.
From the information you've provided, there would seem to be no
problem. The Argyll and lprof profiles are both a better fit to
the IT8 data than the profile maker profile, and therefore would
be expected to produce the most accurate result.
Now this doesn't mean that there is no problem, just that there
is nothing to go on with what you've been able to provide so far.
It may be that you aren't in a position to help track down
what is really going on, because you don't have the gear.
Information that would help:
An example of what the issue is, for instance a photograph
in raw format together with the instructions for applying
the profiles to create a (say) sRGB result, and a description
of why the profilemaker result is closer to expectations
than the other profiles. There would need to be some
explanation as to what is setting those expectations
(ie. are they purely subjective from memory, are they in
comparison with the appearance of the real world, or what ?)
Information on the lighting conditions. Was the same illuminant
used for photographing the IT8 chart as the example photograph ?
If so, what is its spectrum (ie. the output of spotread -s or
equivalent measured with an Eye-One or equivalent.)
If not, does the problem occur with photographs shot
under the same illuminant as used for the IT8 ?
Was the reference file for your IT8 chart measured from your
actual chart, or is it a batch average ? (from the header it
appears to a batch average ?).
Is there a spectral reference file for your IT8 chart available ?
(I can see some reference files on www.targets.coloraid.de have
spectral data, but can't find your batch of charts there.)
What is the spectral reflectance of a sample that is being
poorly reproduced ? What is the spectral reflectance of
the IT8 samples that most closely resemble that sample ?
cheers,
Graeme Gill.
Other related posts:
