[argyllcms] Re: black generation drives me nuts

Martin Weberg wrote:
2007/9/17, Klaus Karcher <lists@xxxxxxxxxxxxxxxxx <mailto:lists@xxxxxxxxxxxxxxxxx>>: 

    you know that I am sold on argyllcms, but to puzzle out the right black
    generation parameters sometimes drives me nuts.


Me too.

    here is a screenshot of "xicclu -g -fb -ir SC_paper_bas.ICC":
    
<http://digitalproof.info/argyll/black_gen/xicclu_-g_-fb_-ir_SC_paper_bas.png
    
<http://digitalproof.info/argyll/black_gen/xicclu_-g_-fb_-ir_SC_paper_bas.png>>


I get a reasonable match using:
profile -qm -kp 0 .24 .98 .98 .55 -l270 -L98 -cmt -dpp fogra40l_270_98
xicclu -g -fb -ir fogra40l_270_98.icc


mmmh, looks smother then ever :-)

... but I'm amazed about the low Kmax:

xicclu -fb -ir F40...-l270-L98.icc:
0.0 0.0 0.0 [Lab] -> Lut ->
0.639593 0.592091 0.496326 0.885713 [CMYK]

xicclu -ff -ia F40...-l270-L98.icc:
0.639593 0.592091 0.496326 0.885713 [CMYK] -> Lut ->
18.134765 1.302415 0.341491 [Lab]

I'm looking for a "landing point" close to this patch:

SAMPLE_ID 1401
CMYK_C    60
CMYK_M    45
CMYK_Y    45
CMYK_K    100
LAB_L     15.61
LAB_A     -0.03
LAB_B     0.75
TAC       250
apparently, everything else than -L100 causes Kmax to decrease dramatically, e.g.: 

xicclu -fb -ir "F40 -qm -kp 0 .24 .99 .96 .80 -l295 -L99.5.icc"
0.000000 0.000000 0.000000 [Lab] -> Lut ->
0.769063 0.681621 0.657291 0.885683 [CMYK]
I don't know how the "magic" black generation code works, but it looks like -L and -l are "snaping in" on several particular locations. 


Little more tweaking can be done I think.


after another dozen of trials, I think I got it now.
at least the neutral axis of the perceptual intent looks smooth and reasonable: 
<http://digitalproof.info/argyll/black_gen/F40-qm-kp_0_.24_.99_.995_.78-l285-L100-ip.png>
<http://digitalproof.info/argyll/black_gen/F40-qm-kp_0_.24_.99_.995_.78-l285-L100-ir.png>

and Lmin is pretty low:

xicclu -fb -ip "F40 -qm -kp 0 .24 .99 .995 .78 -l285 -L100.icc"
0.000000 0.000000 0.000000 [Lab] -> Lut ->
0.591070 0.565901 0.552014 1.000000 [CMYK]

xicclu -ff -ia "F40 -qm -kp 0 .24 .99 .995 .78 -l285 -L100.icc"
0.591070 0.565901 0.552014 1.000000 [CMYK] -> Lut ->
15.070250 0.942825 1.467935 [Lab]


Maybe you didn't want a perfect match but I hope this helps anyway.


It did, Thanks a bomb :-)

Klaus

Other related posts: