Thanks for your explanations. I will give the LUT version of AdobeRGB a try
with Argyll to see how the various intents influence the results for my
scanner.
So is it really the math of matrices which limits the intents to
colorimetric for matrix profiles?
Hermann-Josef
--- Begin Message ---Hermann-Josef Röser wrote:
- From: "Graeme Gill" <graeme@xxxxxxxxxxxxx>
- To: <argyllcms@xxxxxxxxxxxxx>
- Date: Wed, 28 Mar 2018 05:50:05 +0200
Hi,
may I come back to this thread with another question?limited
If your destination is the normal AdobeRGB profile though, you are
to Colorimetric intents because it is a matrix profile.
Is this due to the fact that the required operations for the other intents
cannot mathematically be realized with matrices?
The above depends on the details of the CMM being employed, as well as
the two profiles. Many CMM's have a simple organization in which the
given intent is used to select the same intent for both source and
destination profiles. In contrast, ArgyllCMS's CMM (cctiff) allows
different intents to be chosen for source and destination.
In either case, whether or not there is a difference in overall
conversion depends on whether either of the profiles does a conversion
that is specific to the intent selected for that profile.
If so, would it be possible
to create an equivalent LUT-profile for AdobeRGB which would allow these
operations?
It's certainly possible to create a cLUT AdobeRGB or sRGB etc. profile,
but it's less certain what you would do with it.
[ Basic outline using ArgyllCMS: create a target sample set using targen,
create measurement samples using fakeread, create profile using
colprof with a source gamut selected. ]
When a destination profile is for a device that is of a different
type to the source (i.e. source is additive, destination is subtractive)
or has a noticeably smaller gamut than the source (i.e. again, RGB to
print),
then gamut mapping has work to do, and perceptual and saturation intents
can be expected to be significantly different to colorimetric.
If your source is a scanner and destination is AdobeRGB, then the
role of gamut mapping is less clear. Scanners are a bit like cameras
in having a gamut that is harder to define in itself, although
you can certainly do so if you are scanning a particular type
of media (i.e. a certain photographic print film, or printer output.
Note that the gamut of a scanner in this sense is not the same as
the gamut of the profile defining its response - you would need to
establish the source gamut separately.)
But if you can define or assume it, and it is significantly
different from the destination space gamut you want to convert
to, then creating a cLUT based destination profile with
gamut mapping may well make sense.
Cheers,
Graeme Gill.
--- End Message ---
