>A question for Dan and Tyler: how did you arrive at these curves? Is further >adjusting them the way to solve this problem? If so, can you give me any >advice on doing it sensibly? Finally, is my calibration-by-stepped-grayscale >a realistic approach? I think the curves posted are Tyler's finessing of my first rough partitioning curves but the answer does lie in tweaking the curves (as you suspected). As for advice -- I think the best thing to do is make the info pallette visible and set the second color to "total ink." Then as you adjust the curves run the cursor over various areas to see if there are any supposedly smooth areas that drop or raise total ink suddenly. If look closely at the info pallette first color (it should be set to "actual" which becomes CMYK) running the cursor over the same area to see which ink channel is posterized. Then tweak the curve for that channel to move the area of posterization under an area of heavier ink (you may see the discontinuity bands move as you adjust the channels). Usually you need to iterate between curves and channel mixer for the best result. It is best to work in 16 bit as much as possible since this process *challenges* the individual 255 level channels of 8 bit. Of course in 16 bit you can't use layers so tweaking involves a lot of do and undo and saving and reloading curves settings. The channel mixer layer (or adjustment) removes the light color inks from under the dark color restoring detail to the shadows that would be lost with heavy ink spreading. But it is easy to overdo it. Try setting the layer opacity lower (or in 16 bit try adjusting the amounts of ink removed in each channel of the channel mixer adjustment). It is better to get a light partitioning of the channels in 16 bit with very conservative curves and channel mixer adjustments, then convert to 8 bit and add the adjustment layers for final tweaks. Did you get my action which does the first partition automatically in 16 bit? If not email me off line and I will send the files. I'm not sure calibration buys you much since each image is so different and the channels really are stretched to their limits. But it will give you a good start. Try pasting a narrow step gradient along the top of the image (per recommendation of Blatner and Fraser for similar types of activities). Dan Culbertson - Turn off HTML mail features. Keep quoted material short. Use accurate subject lines. http://www.leben.com/lists for list instructions.