[argyllcms] Whitepoint difference Argyll vs ProfileMaker and ProofMaster

Hello all,

Made a profile for Epson 4800 ( ContractProof Paper ), first with
Argyll and then with ProfileMaker, but using the same measurement
files.

There is a small difference between the whitepoint computed by Argyll
and the one computed by ProfileMaker ( bigger in the Z component
~0.75 for Argyll, ~0.78 others ). Also, I measured the substrate with
ProofMaster and computing the XYZ gave a very close value to the one
of ProfileMaker.

Did a little debugging and the difference seems to come from
sp2cie_sconvert. The Lab computed here from spectral is different than
what the other two apps get.

Actually, the difference is better observed when doing a chromatic
adaptation ( Bradford ) from ISO Coated to the printer profile. The
substrate simulation when using the Argyll profile gets way off,
shifted a lot towards magenta.

When using ProfileMaker, the substrate gets a yellowish hue, which is correct.

The difference between the two is better observed when doing an
UgraFogra certification.

I used Argyll v0.60 when testing.

Can anyone say why this difference?
I can't really attach anything now ( I'm off work ), but I'd be happy
to attach the certify reports and/or the generated profiles on monday.

Regards.

Other related posts: