Graeme Gill wrote: > Hmm - they really shouldn't. A2B is the natural direction of an output > device: device values in, CIE values out, so it's "forwards". > B2A inverts the natural direction, so it's "backwards". > (It's consistent with A and B alphabetical order.) > > Graeme Gill. > Goog to know! Confusion is not only in my head as I thought :-) Graeme, let me try to explain what I was thinking when I first posted my plots (probably I'm wrong but it will takes just a minute for you to realize it!). From running colprof I noted the target white and black point White point XYZ = 0.845282 0.876724 0.746296, Lab = 95.022845 -0.010481 -2.015543 Find black point K only value (Lab) = 24.460331 -0.002464 -0.214489 Black point XYZ = 0.019814 0.020550 0.016944, Lab = 15.773101 -0.000874 0.007522 Then started to look at the "rollercoaster curves" and investigated the TAC values 35.560000 0.000000 0.000000 [Lab] -> Lut -> 50,443696 41,711687 40,067612 59,870632 [CMYK] -> TAC 192,093627 31.620000 0.000000 0.000000 [Lab] -> Lut -> 49,424223 40,768571 38,909611 68,558099 [CMYK] -> TAC 197,660504 27.660000 0.000000 0.000000 [Lab] -> Lut -> 44,350140 36,607080 34,194777 78,359684 [CMYK] -> TAC 193,511681 23.695000 0.000000 0.000000 [Lab] -> Lut -> 40,055614 32,925542 30,424117 85,407089 [CMYK] -> TAC 188,812362 > To hit the deeper blacks while not going over the > total ink limit, more black needs to be used, and less CMY. I expected CMY values to start decreasing at some point but it seems that this is done when there is still room to add more black without concerning about total ink limit. From L*a*b* 31.62 0 0 to 27.66 0 0, for example, you could maintain CMY values and increase black without exceeding the TAC limit. I'm not saying that colprof is doing anything wrong. Hopefully, understanding what's wrong in my reasoning will help me getting the best out of Argyll. Thanks Fabrizio