Re: custom targets in ProfileMaker Pro 5
Re: custom targets in ProfileMaker Pro 5
- Subject: Re: custom targets in ProfileMaker Pro 5
- From: Graeme Gill <email@hidden>
- Date: Tue, 13 Jul 2004 18:07:48 -0700
Terence L. Wyse wrote:
I don't know about the RGB stuff but for press/inkjet profiles, you'll
see a big jump in quality going from a 400-500 patch chart (TC3.5) to
something like the 900+ patch IT8.7/3 chart. Jumping from the IT8 to the
ECI2002 (abouut 1,500 patches) chart gives you slightly better results
but with diminishing returns. The only significant difference I've see
We've noticed differences in profiling CMYK inkjet devices between
the ECI2003 and 3000 or so patches when aiming for high accuracy.
Say you're device has a total gamut volume of 500000 cubic delta E, then
even 3000 patches can at best share 167 cubic delta E each. If this is a
sphere (ideal case), then the radius (hence furthest distance of any
point in the colorspace to a sampled point) will be 3.4 delta E.
For 1500 patches, it will be 4.3 delta E.
For 900 patches, it will be 5.1 delta E.
for 500 patches, 6.2
So if you assume that whatever interpolation algorithm gets used
has an interpolation error that decrease the closer it is to a test
point, then the smaller the maximum unsampled point distance,
the less error there will be. (In real profiles there are other
sources of error as well, including any introduced in inverting the
profile.)
In the real world, none of the numbers are this good, because most
test charts patches aren't uniformly distributed or packed in
perceptual space.
Graeme Gill.
_______________________________________________
colorsync-users mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/colorsync-users
Do not post admin requests to the list. They will be ignored.