[argyllcms] Re: FWA overcompensation bug in spec2cie

  • From: Gerhard Fuernkranz <nospam456@xxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Thu, 15 Jul 2010 15:11:04 +0200

Am 15.07.2010 12:18, schrieb Alexey Gribunin:
> So I consider this feature as a simulation of UV cut filter.
> And now the question. How do you use spectral measurements from UV Cut
> device?
> Do you create ICC profile with the regular D50 light source or with
> the "special UV Cut" light source?
> I think, you will say "regular D50", right?

Well, this may be common practice, but if FWAs are the play, then XZY
numbers computed for D50 from UV-cut readings are no longer supposed to
reflect the actual XYZ colors of the samples under a light source with a
(regular) D50 spectrum (i.e. what you compute from the UV-cut readings
is not what you see).

The XYZ number you get from spec2cie -f are on the other hand supposed
to be an estimate for the actual XYZ color of the samples when
illuminated with the light source that was specified with the -i option
(of course just an estimate - as good as the FWA model can do).

The idea of the implemented procedure is that you measure the samples
with a non-UV-cut instrument, tell spec2cie under which light source
spectrum you want to view the samples, and you get the estimated XYZ
colors of the samples under this particular light source (and this
estimate includes the color shift induced by the FWA due to UV stimulation).

[Btw, this is not specific to spec2cie - exactly the same applies if you
call colprof with the -f and -i options]


> Ok, that's why I'm trying to get either filtered spectrals (which
> simulates UV cut measurements) or Lab values, calculated from that
> filtered spectrals with STANDARD D50.

Other related posts: