PDA

View Full Version : Getting mixed results from DeviceLinking


rogerB
01-26-2011, 10:29 PM
We got an average of 2.30 dE and a max of 9 between our "confirmation" test run and SWOP2006_C3, which isn't bad. I used a FullReseparation setup with an AbsCol RI. The DL fixed the broken hue angles on both the Magenta and the Yellow, but didn't do anything to better match the Cyan, which is a little on the bluish side. I'll have to continue investigating because the results are in the right direction overall except solid cyan, which is odd.

Roger

Newton
01-27-2011, 02:09 AM
For the "confirmation" run are you simply running the colorflow charts with the new plate curves?

It never occurred to me to do that...

rogerB
01-27-2011, 02:34 PM
New Gray Balance curves + Device Link Profile, not ink optimized.

Best / Roger

david.herder
01-27-2011, 04:47 PM
Roger,
You're using a Full Reseparation DVL to go out to press?

rogerB
01-27-2011, 06:26 PM
For testing purposes, only. Here's the deal. If I stick with an ink optimizing setting, the solids are never reseparated or matched to the PCO. So, using a FullReseparation is the only way to see how good is ColorFlow's color matching engine.

Which brings me to a question. Suppose a person's face. In the cheek, one finds 10% magenta and 10% yellow. Am I right in understanding that, because the software insists on keeping secondary colors "pure", those 10% magenta and 10% yellow will remaind unmanaged, in an out of the device link, in an ink optimized setting? Possibly being nudged up or down a few percentage points, to match the incoming L* value? Ignoring the incoming a* and b* values?

What about the case where the tint composition is different, like 10% magenta and 30% yellow. Do different rules apply then? I still have not had found the time to reverse-engineer how the device link gamut maps all colors.

Best / Roger

david.herder
01-27-2011, 07:34 PM
Am I right in understanding that, because the software insists on keeping secondary colors "pure", those 10% magenta and 10% yellow will remaind unmanaged, in an out of the device link, in an ink optimized setting?


Yes.


Possibly being nudged up or down a few percentage points, to match the incoming L* value? Ignoring the incoming a* and b* values?


No L* adjustment is made, the input tint of primaries and secondaries are preserved entirely.


What about the case where the tint composition is different, like 10% magenta and 30% yellow. Do different rules apply then?


No, the same rule applies, it is still a MY build.

rogerB
01-27-2011, 08:02 PM
So, effectively, because of the way the device link is designed, many incoming colors remain unmanaged on the way out to the SCO? Like all those skin tones where the composition is only made of two colors? Regardless of the hue "error" in the solids?

In effect, can you say that only tertiary colors get managed? For the purpose of ink reduction and color matching?

Roger

david.herder
01-27-2011, 08:14 PM
With Ink Optimizing DVLs, and the fully constrained DVL, yes, this is true.