[argyllcms] Re: OS X 10.8.4 from 10.6.8, deltaE 11 from 3
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Fri, 09 Aug 2013 16:50:50 +1000
Technical Operations wrote:
> Thanks very much Graeme, I narrowed the problem down to trying to use a
> Rec.709 tone
> curve.
Hi,
you mean is dispcal ?
> With it I get the high dE's and visibly off (even with the 1.6 beta codebase)
> and the sRGB coverage is 95% or less. With a standard 2.2 gamma tone curve,
> our dE fall
> right into place (avg<1, max<2.5) and the sRGB coverage is 99-101%.
When you are talking about delta E's, delta E to what ?
colprof self fit ?
Against a target colorspace emulation such as sRGB ?
> Any ideas? I'm testing using dispcalGUI's "verify profile", using either the
> Rec709.icm
> or straight to the ti1 chart, which in both cases give the same results.
Sorry, I don't really know what dispcalGUI is doing, so I can't really help.
If you have a verification workflow using Argyll tools, I can help.
> Rec.709 is strongly preferable for us since most of our work is in that color
> space.
A lot of the advice in the V1.6 Beta video calibration section of the tutorial
applies:
Calibration transfer curves are overridden by the profile linking, so don't
target
the source transfer curve in calibration, use something that is perceptual and
controllable like sRGB or L*.
You don't want a Rec709 transfer curve because it's an encoding space not
viewing space - you want Rec709 primaries with a BT.1886 transfer curve.
You probably want to make a viewing conditions adjustment.
etc.
Graeme Gill.
Other related posts:
