[Openicc] Open Source compatible AdobeRGB profile

[Graeme Gill]

Gerhard Fuernkranz wrote:
> Graeme Gill wrote:
> 
>> Given the issues with ICC V4 display profiles not supporting Absolute
>> Colorimetric Intent, I'm not sure a V4 profile is all that desirable.
> 
> 
> Graeme, why "not supporting absolute colorimetric intent"? I understand 
> that a correct V4 profile does contain all information needed by a CMM 
> to compute both kinds of absolute intent, a "true" absolute reproduction 
> (CIE absolute colorimetry), and also the (illuminant relative) 
> ICC-absolute reproduction as defined in the ICC V4 spec. It's up to the 
> user of the profile to do the desired thing with the information 
> recorded in the profile.

It may contain the information, but it's not available using
any of the ICC defined intents, whereas with a V2 profile,
it is immediately usable in a very wide range of systems
using the icAbsoluteColorimetric intent.

> The question is probably rather, can all CMMs handle V4 profiles 
> properly yet? And of course, as the Adobe RGB 1998 profile is obviously 
> a V2 profile, the compatible one should probably be a V2 profile too, 
> for best interoperability with V2-only CMMs, shouldn't it?

Probably. To be fair, the issue with V4 display profiles
and absolute isn't terribly relevant to Adobe RGB 1998
since it will almost never be the source or destination
colorspace where soft proofing is the aim.

> IMO the problematic ones are rather those V2 display profiles, which 
> generally record D50 in the 'wtpt' tag (i.e. the true WP adapted to D50, 
> which gives always D50) and which furthermore do not contain a 'chad' 
> tag. The information recorded in these profiles is indeed insufficient 
> to reconstruct the true white point chromaticity and to compute "true" 
> absolute colorimetry.

I've rarely come across such display profiles. Of course, V4 Display
profiles are all meant to be 'problematic' ones (albeit with the
'chad' tag recording what went on to get a D50 "absolute white" point.) :-)

> (But Adobe RGB 1998 does not belong to the latter class anyway, but 
> records D65 in the 'wtpt' tag, as also does the old sRGB profile from 1998)

Yes.

Graeme Gill.