[argyllcms] [argyllcms] Re: colprof and proofing purposes

  • From: Nikolay Pokhilchenko <nikolay_po@xxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Mon, 28 Sep 2009 19:54:08 +0400

Michael, my advice is "IMHO". There may be another opinions and experience. :)
It's would be normal to profiles to be compute with -qu quality in 20-40 
minutes for RGB and till several hours for CMYK.

One more IMHO: You should avoid the channel and channel mix clipping by RIP ink 
limit control. It will the best if per-channel limits will be scaled to 100% 
without saturation. For example yellow channel ink limit 80%:

Y=50%in => Y=40%out
Y=80%in => Y=64%out
Y=100%in => Y=80%out.

If the RIP limits the channel with saturation, like this:

Y=50%in => Y=50%out
Y=80%in => Y=80%out
Y=100%in => Y=80%out,

the profiler will work harder with the saturations. The unwanted transitions on 
image gradients may appears at the limit point.
It would better to set total limit in the profile (colprof -l 360), nor in the 
RIP. If the printer can not be damaged by accidentally total ink limit 
override, I'm preferring to set only linearizion with per-channel limits but 
turn off total ink limit in RIP.
This approach gives the profiler the best condition to work.

One more IMHO: when the good linearization is achieved and total ink limit is 
defined, You should print the Argyll targen generated targets with the defined 
TIL+10%, nor the standard ISO targets. If target TIL will be closer to desired, 
You'll get more useful patches in the area of interest. The patches behind Your 
practical TIL will be almost useless. Especially if You'll not try to predict 
printer behavior behind printer TIL.

Michael Schulz wrote:

> Hello Nikolay,
> another good advice - thank you. Now I have to find the time to do all
> this...
> Regards
> Michael
> Am 28.09.2009 15:37 Uhr schrieb "Nikolay Pokhilchenko" unter
> <nikolay_po@xxxxxxx>:
> > Hello, Michael!
> > 
> > I'm preferring the -qh or even -qu profile quality for inkjet printers. But
> > You should use a high number of different test patches for high precision
> > mode. For example, I print about 2700-3600 patches for -qu. There may be a
> > difference: the higher quality mode have more freedom to curve the channels,
> > that's why there may be some strange effects. It may be noticeable when test
> > patch quantity is insufficient for the device characterization at certain
> > region.

Other related posts:

  • » [argyllcms] [argyllcms] Re: colprof and proofing purposes - Nikolay Pokhilchenko