[argyllcms] Re: Checking profile with GrangerRainbow
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Mon, 03 Oct 2016 23:10:55 +1100
posts@xxxxxxxxxxxxxxxxxxxxxxxxx wrote:
Hi,
following the advice by Fraser et al. (2005) I have assigned one of my
scanner profiles
created with Argyll to a GrangerRainbow image. The profile was for a Fuji
slide using
colprof with the options -qh -Zp -as -bn.
I'm not really sure what you mean by "assigned to". A scanner profile maps
between
the scanner RGB gamut and PCS values. Due to the nature of the type of profile,
the
gamut explored by the test chart is limited. Any RGB values beyond those
exercised by
the test chart have to be extrapolated, and hence are inaccurate. So there
is no surprise that if you explore the extrapolated RGB space, that you
may see undefined behavior.
There is a sharply defined rectangular range at upper left, which should not
be there (see
attachment). It vanishes if I use the option -ax, where the Rainbow image
looks as expected.
A matrix model or one in a different PCS space to L*a*b* will tend to different
extrapolations.
It's all rather irrelevant unless you have things you are going to scan
that fall into these areas, in which case you really need a test chart
that covers the gamut you want to use.
Cheers,
Graeme Gill.
Other related posts:
