[argyllcms] Re: Custom illumination

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Wed, 20 Jan 2010 12:22:18 +1100

Klaus Kompatscher wrote:
Did try to create profile for custom illumination:
- Measured light with I1 with special diffuser; Got measurements in cxf
format (file attached)
- Made a profile with Profile Maker White point looks good
- Copied spectral values to ref. D50_1.5.sp file and renamed DalTile...
(file attached)
- made a profile ./colprof -v -i
/Applications/Argyll_V1.1.0/ref/DalTileLichtI1.sp -kp 0 .3 1 1 .8 -r 2 -L
100 -l 280 G60_G07_5c

White point calculated by Argyll is fare off from the real.
Visually it does not correspond at all.
PM white point: Lab 87 0 2
Argyll white point: Lab 88 5 30

Why this big difference?

What's going on I suspect is that you are asking Argyll to do
a non-ICC standard thing and it is accommodating you, whereas
Profile Maker is doing a sort-of standard ICC thing. See the note
at the bottom of the paragraph here in the Argyll doco:


"Note that if an illuminant other than D50 is chosen, the resulting ICC profile 
 not be standard, and cannot be freely interchanged with other profiles that 
that use
 the standard D50 illuminant, particularly if the absolute rendering intent is 
 Profiles should generally be linked with other profiles that have the same 
 and observer."

What I guess Profile Maker is doing is computing the media colors as if they
are viewing in the given illuminant, and then creating a chromatic adaptation
matrix to convert the white to D50 to make it (sort of) conform to the standard
ICC conditions. I say sort-of because standard ICC is actually to measure under
D50 illuminant where possible, and a non-D50 illuminant + chromatic matrix is 
the same thing, and Profile Maker could have simply ignored the given illuminant
and used actual D50 if it is really trying to create a standard ICC profile.
[The option in the ICC spec. of using a chromatic adaptation matrix for the 
 is intended for the situation where measuring under D50 is not possible.]

I think this rather defeats the purpose of Absolute colorimetric rendering 
though, since a standard CMM will not reproduce the absolute appearance if the
illuminant has been adapted to be D50 white point.

No doubt I could do the same with Argyll as an option, but I'm sure it
would increase the confusion factor, and I'm not sure I see the point. If
you want ICC standard D50, then give it D50 spectrum. If you want absolute
intent under an alternate illuminant, then that's exactly what Argyll
provides, even though a non-standard ICC profile will result.

What means Spectral_Norm in the Spectral illumination file?

It's a number to scale the spectral values. Used as an illuminant
this shouldn't have any effect (An illuminant is always normalized
to 1.0).

Graeme Gill.

Other related posts: