Has anyone dabbled with Firefox's rendering intent?

>
> Came across a thing that slightly surprised me when I had a look at Mogrify ( http://timothyarmes.co​m/lr2mogrify.php?sec=q​uickguide )
> It has the option to set a rendering intent.
> Since, as far as I know, the only available rendering intent when converting between matrix profiles (most RGB working spaces) is relative colorimetric, I did some digging.


Hi Rene. Please note that you can't obtain relative colorimetric with
matrix shaper v2 profiles because they lack an offset, so you will
obtain XYZ=(0,0,0) for RGB=(0,0,0) Obviously this is not true since
no device can render an L* of zero. I think matrix-shaper should
be avoided today, since we have more powerful and accurate ways
to store the information. Unfortunately, many vendors still are
using this method to build the profiles.

>
> Couldn't find any info on the littleCMS website, so googled a bit, and found this: http://ufraw.sourcefor​ge.net/Colors.html
>
> "INTENT_PERCEPTUAL:
>
> In lcms: Default intent of profiles is used.
>
> INTENT_RELATIVE_COLORI​METRIC:
>
> In lcms: If adequate table is present in profile, then, it is used. Else reverts to perceptual intent."
>

That means, if BToA1 LUT table is found, then it is used. Else it
reverts to BToA0 table, which is perceptual. Matrix shaper is used
only if everything else fails, as this model only works well on
syntetic spaces and certain CRT monitors. ICC is spec is very
explicit about precedence.

> Sounds to me like it should be exactly the other way around, right? Unless a table based profile is used, perceptual is impossible.
> http://www.colorwiki.c​om …agement_Myths_2​1-25#Myth_. 2321:_There_are_percep​tual_rendering_intents​_available_when_conver​ting_f rom_scanner.2Fcamera-.3Eworkspace_or_worksp​ace-.3Eworkspace.


Humm... that may be true for old v2 profiles, but today's V4 certainly
allow that. Latest ICC addendums have addressed this issue. Please see
those papers coming from ICC:

http://www.color.org …es_with_camera_​images.pdf
http://www.color.org …l_rendering_use​_cases.pdf


> At least, that's how it works in Photoshop.

CS4 fully supports V4 standard. Even more, it supports MPE which is the
ICC proposed solution for digital cameras. But I realize it is hard to
locate such profiles.

>
> Also:
>
> "INTENT_ABSOLUTE_COLOR​IMETRIC:
>
> Outside the gamut; hue and lightness are maintained, saturation is sacrificed."
>
> Is this always the case? I though that this; http:// www.brucelindbloom.com​/index.html?MunsellCal​cHelp.html#BluePurple also applied to colorimetric rendering intent?
>
> Maybe I'm missing something here?

Well, the behaviour for out of gamut is not defined by ICC. So, not
the case as each vendor may choose its own gamut mapping strategy.
Also, V4 has changed the meaning of absolute colorimetric. I agree this
is highly confusing.

Best regards
Marti