Re: [Lcms-user] Getting the profile whitepoint

[<mar...@li...>]

Hi,

> As an aside, I did wonder about the documentation for
> cmsCreateExtendedTransform.
>
> cmsCreateExtendedTransform(cmsContext ContextID,
>                cmsUInt32Number nProfiles, cmsHPROFILE hProfiles[],
>                cmsBool  BPC[],
>                cmsUInt32Number Intents[],
>                cmsFloat64Number AdaptationStates[],
>                cmsHPROFILE hGamutProfile,
>                cmsUInt32Number nGamutPCSposition,
>                cmsUInt32Number InputFormat,
>                cmsUInt32Number OutputFormat,
>                cmsUInt32Number dwFlags);
>
> What I expected was nProfiles = 2, hProfiles[2], BPC[2], Intents[1],
> AdaptationStates[2], i.e. you set (nProfiles-1) worth of transform
> data. If you do this it doesn't work, and so I have to do something
> odd like:
>
>     cmsUInt32Number intents[2] = { INTENT_ABSOLUTE_COLORIMETRIC,
>                        INTENT_ABSOLUTE_COLORIMETRIC };
>

Ok, this function allows you to specify intents *per profile*, that means
you can use relative colorimetric when going from RGB to PCS in the first
profile and perceptual when going PCS to second profile. Mixing intents
is generally a bad idea, but there are situations where it can be
useful. So here is the functionality.


> Also I don't think BPC makes sense for absolute-colorimetric, and I
> kinda expected to be able to pass NULL there. If you do that it blows
> up :)


Right. I will try to fix that in next release (maybe NULL would mean
no BPC?)

Regards
Marti