Re: I Never Met A Profile I Liked!
Re: I Never Met A Profile I Liked!
- Subject: Re: I Never Met A Profile I Liked!
- From: Roberto Michelena <email@hidden>
- Date: Mon, 3 Sep 2001 18:51:25 -0500
>
Exactly my point! You can't define accurate because it is not static. To
>
make it even more difficult, accurate may not be the desired result as you
>
pointed out regarding CompassProfile profiles and the opinions of
>
photographers.
>
>
This is why I constantly maintain that our *opinions* about "the most
>
accurate" are rarely relevant.
There are some aspects of a profile's accuracy that can be numerically
evaluated. One of them, that I consider quite important, is internal
consistency or "back and forth" in colorimetric tables.
If you feed some in-gamut Lab values into a profile's abscol PCS->Device
tables, it will give you some CMYK values in return. Then feed those into
the profile's Device->PCS tables, and it should give you the same
original Lab values that you fed into it at the beginning. Within a
tolerance, of course.
You are basically verifying that the profile trusts itself, so to say.
Of course, this depends on CMM also. But between mainstream CMMs, the
differences are not that much. Much more identifiable are the differences
between profiling packages.
There's also the correspondence between measured data vs device->pcs
abscol table. You can measure it, the only caveat being that if measured
data does not come from an averaged set, it is not a fair test.
Both of them give you a very good idea of a profile's colorimetric
precision.
I believe Monaco is the king of numerical match in colorimetric, although
I've not tried latest ProfileMaker versions.
PrintOpen got a big boost in this direction from v3 to v4.
But there's one more caveat: numerical match (in colorimetry) doesn't
necessarily mean visual match. But at least when dealing with similar
inks and substrates, it will. Like for example matching an inkjet printer
to another, even if different brands and models, provided you use the
same inkjet paper for both.
-- Roberto Michelena
EOS S.A.
Lima, Peru
- Prev by Date:
Re: InDesign 1.5.2 prints on Epson, no Postscript RIP needed
- Next by Date:
Inks, inks, inks, pigments, dye-based, your experiences?
- Previous by thread:
Re: InDesign 1.5.2 prints on Epson, no Postscript RIP needed
- Next by thread:
Inks, inks, inks, pigments, dye-based, your experiences?
- Index(es):