[argyllcms] Re: Image dependent ?
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Wed, 14 Nov 2018 10:31:01 +1100
I'm a French speaking person and I don't fully understand
the subtleties of the English language.
I understand, and given my lack of abilities in French,
we had best battle on in English :-)
"What is different is where the source gamut is coming from - i.e.
from the image rather than the colorspace."
Could this be interpreted to be that internally, when it's time to do the
gamut mapping portion of the work collink uses the gamut of the image
instead of the gamut of colorspace?
No, because collink doesn't have access to the images. It only
processes ICC profiles. tiffgamut is the tool that extracts
a gamut from an image, and so makes available the relevant
information to collink.
[ ArgyllCMS is a toolkit rather than having "do everything" programs. ]
As for the V4 profiles, I was thinking the other way around again, my
question should have been are there V4 compatible tools (out there) that
could be used especially with the image gamut like you do with collink. But
don't worry about that since I'll do my own profile in the near future using
A workaround would be to use some ICCV4 compatible image editing tool to
open the image files and then convert & save them in a device independent
format such as L*a*b* encoded TIFF to feed into tiffgamut.
Note also that tiffgamut doesn't automatically use embedded profiles, so
if you know what V4 profile an image is using, you could provide the equivalent
V2 profile to tiffgamut.
Thanks for the suggestions on the spectrophotometer, I think I would go for
an I1 Studio probably, it should be plenty enough for my need and maybe even
a bit overboard.
[ Of course if anyone knows of a competitive instrument, feel free to
suggest it. I personally _don't_ think that the Datacolor SpyderPRINT
is competitive, because (not being a general purpose spectrometer) it is
a lot less flexible, and I have concerns about its practical accuracy.
Some people are perfectly satisfied with the SpyderPRINT, but note that
it's not supported directly by ArgyllCMS. ]
I'm sure a scanner is not a spectrophotometer by far but I wonder
if it could be useful to some degree, if you could share your opinion on
this if only to convince me of not wasting time on this, I would appreciate.
Generally people are disappointed in profiles made using scanners. A scanner
will probably be better than nothing in terms of measuring channel response
(i.e. to profiling tonal response), but is typically poor at measuring the
color response, due to it not being colorimetric.
The best alternative to owning a dependable instrument is probably using
an external profiling service.
Other related posts: