# [argyllcms] Re: Black turning down problem - help!

• From: Nikolay Pokhilchenko <nikolay_po@xxxxxxx>
• To: argyllcms@xxxxxxxxxxxxx
• Date: Sat, 08 Jan 2011 15:20:13 +0300

```Fri, 07 Jan 2011 20:27:49 +0100 Elena [service address] wrote:

> I have spent good hours (sich) playing with xicclu -g interactively, and I
> partially save you from the trouble of answering my questions (partially ;).

I'd recommend You to play with high resolution profile for xicclu. My profile
for playing is

colprof -v -al -qh -bn Basename

> I never played so extensively with it. I understood the -k and -K behaviours
> are all but intuitive: one has to try with all 5 parameters until he obtains
> BOTH the desired K behavior AND smooth curves.

Unfortunately it's true. But with several hours of experience with different
CMYK devices behaviors, You'll be able to achieve desired curve if the device
behavior permits it.

> Playing with -K or -k is a real art.

From some point You'll get the feeling how to manipulate with -kp and -Kp. The
best principle of forming black generation curve (-K or -k) is depends on a)
the device behavior and b) the desired black curve.

The aim of trying the parameters is smoothness of the black curve at the
joining point between -kz and -kx. The second aim - the smoothness of CMY.

> As for the limits, I understood that one is better to chose -l being the sum
> of the plotted c,m,y,k at 100% as for -kz, and -L as the max k allowed
> at 100% as for -kx.

I can't confirm or disprove this. I'm checking the effective limits (TIL and
black) by resulting gamut. At this stage the iccgamut is useful, because it can
compute the gamut only by A2B table of the profile. For example:

iccgamut -v -d3 -pj -w Basename.icc - here take look at default ink limits by
iccgamut judging.
iccgamut -v -d3 -pj -w -l300 -L100 Basename.icc
iccgamut -v -d3 -pj -w -l300 -L65 Basename.icc
iccgamut -v -d3 -pj -w -l270 -L65 Basename.icc
iccgamut -v -d3 -pj -w -l280 -L65 Basename.icc
iccgamut -v -d3 -pj -w -l290 -L65 Basename.icc

By this information You can determine minimum practical ink limit for Your
paper as a compromise between gamut, paper wetness and (in some cases) profile
smoothness.
The similar check is checking the deepest black dependensy on ink limits, for
example:

xicclu -v -fif -ir -pj -l 300 -L100 Basename.icc

and vice versa.

xicclu -v -ff -ir -pj Basename.icc

Particular case:

-l300 -L100
0 0 0 [Jab] -> Lut -> 1 1 0.809165 0.090437 [CMYK] Lim 2.899602 (clip)
1 1 0.809 0.090 [CMYK] -> Lut -> 23.66 -4.170 -4.514 [Jab]

Here we see, that deepest black we can get only by 290% total limit. By the
way, the K level at the deepest black is only 10%! So, the black can be
strictly limited on 65% as iccgamut said without gamut decreasing (if
preliminary profile isn't wrong)
Another computing:

-l290 -L100
0 0 0 [Jab] -> Lut -> 1 1 0.809165 0.090437 [CMYK] Lim 2.899602 (clip) the same
result as with 300% TIL. The limit of 290% is OK for black.

Conclusion: I recommend set for this printing process the next limits:

TIL=290% with 0.018% decreasing of gamut by whole 10% decreasing total inks. I
suppose it's expedient.
BlackLimit=65..100%, depending on smoothness of resulting profile curves.

Appropriate black limit should be determinated whilte xicclu "playing"

> I'm improving my results but... at what a time cost!!!

Yes, the time is the cost...

```