[argyllcms] Black turning down problem - help!

  • From: Elena [service address] <1007140@xxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Wed, 05 Jan 2011 19:42:19 +0100

Hi Graeme,
I'm stuck with this problem since days, I hope I have now collected enough
info to provide you with.

I'm still making experiments profiling plain paper which, I know, sucks
(in every sense ;) - but as I told, I don't feel to proceed expending efforts 
time in profiling quality papers until I'm sure that everything is working just
fine (from my side also, if that's the case).

Any way I proceed, I'm more or less getting bluish bumps in the deep blacks.
And xicclu seems to confirm some serious problem in that area, when looking with
it at the kplot. I included details, using a ofps 1300 target as reference.

 targen -v -d4 -e4 -s21 -g21 -f1300 -l310 ofps

I measured it with my i1 pro, making sure that everything went fine. I then
generated a profile:

 colprof -v -D"Argyll ofps kx qm" -y -qm -kx -ssRGB.icm

here is the command output:

and here is the profile, if you want:

here are some SRGB to CMYK test converions made with photoshop from a test
picture with color gradients, using the above generated profile. So you can
visually see the problem in the blacks. TIF files are original CMYK conversions,
JPGs are the same converted to RGB for quick watching:

Relative colorimetric:




Here are xicclu plots from the above profile (see the options inside the 


As you see, the generated Rel intent does show this abrupt ramping down
of K - and I honestly don't see why, since I have specified kx! Colprof 
do that in any case, it should obey my black choices and adapt c,m,y 
NOT turning the black down! Afterall, this behaviour is well visible in the 
pictures above.

What I don't explain, instead, is why the -fb -ip kplot doesn't show such a K
behaviour, while the perceptual conversion still visually shows the bluish bump!

Here are kplots of the same profile exploring table inversion with -fif, and
of course -l300 -L100

  this one with -kp 0 0 1 1 .5

The K ramping down is always predicted.

I can't now flood you with data. It sufficies to say that I made dozens other
tests with profiles generated with custom K, different targets (cubic grid
centered space and also regular grid, -m6). I also made a test, converting
with a custom written utility a testchart generated with Profile Maker to
.ti1 as target, with matching grey generation options. I also tried
iterating one step, i.e from the original OFPS target I made a profile, I then
made a second OFPS target passing it as pre-conditioning profile. No 

In every case I get pretty the same results: bluish bumps in the dark blacks.
Sometime more, sometime less, sometime it's more visible with the RC intent
than with the Perc intent, sometimes else it's the opposite, but always there.

I don't think that about 1300 patches aren't enough to generate a preliminary
but overall correct profile for plain paper. Or just they aren't ??
Colors are right and pretty good after all, test photos printed are ok, except
the bluish bump in the darks. I know that plain paper is bad and can't reach
a decent dmax, but how is that I can't profile it ? Why colprof insists making
the K ramping down ?? Is it my fault ? Where ??

Thanks for you support,

Other related posts: