Re: Screen profiling

Hello,

In my application I also do screen and print profiling,
although with much less success for the screen.
I understand Bob's wish to directly compare emission and reflectance
measured D65 CIE Lab values.

Tom Lianza wrote:
2. When filling out the CLUT, you must not include L* = 0; The lowest L*
should be the minimum L* found in the media. If you are previewing in
Photoshop and you have set your minimum value to 0 (either in the
MatrixShaper Lut or the CLUT) you will see a real increase in luminance
when proofing to the screen.
I agree with that.
In my application, I scale the screen L values to realistic values
I can get from reflectance measurement, that is roughly between
14.0<L<92.0
Probably this is some kind of gamut mapping.
If I don't do this, the screen colors are completely off,
just like Bob describes it.

Gernont's patterns are very nice
http://www.fho-emden.de/~hoffmann/caltutor270900.pdf
but I always found his L=0.0 and L=100.0 patches disturbing.
Certainly a backlit LCD does not have L=0.0
more like L=25.0
If not, I can't see how come a swatch of black woolen yarn
with L=14.0 is way way darker then an LCD displaying R=G=B=0
with L=0.0 from Gernot's PDF?

I know a little about appearance models and have
read some of Mark Fairchild's book.
But in my industry (textile CAD) people expect to see
screen colors matched to actual yarn sample they have
in front of them, and then printout to look the same.
They don't sit in a light box, and don't have Barco monitors.
So telling them that the color is equal, they just
see it differently, doesn't work.

Best regards,

Dušan Peterc
http://www.arahne.si
.