Elena wrote: > For a good job, > one would better create a second additional target with many many dark > patches, > possibly even duplicating ones, so that colprof can average them, all randomly > spreaded on the chart. May be the (an example) targen -d4 -G -fxxxx -c"preliminary.icc" -A1 -f0,0,0,25 NewTarget will help to automate the process? As I understood the targen behavior, it will optimize the chart by device responcse curvature first of all (if the -A parameter is 1 - default for preliminary profile use. The -f key can help to filter out dark patches. As I understood the colprof, it computes different error values for the different device space (or color space?) points. And it can do this not only by repeated patches, but by adjacent patches too. So I didn't see the need to duplicate the patches. The higher number of patches per volume unit can be enough. The colprof will be able take this into account and account the errors more precisely. Am I mistaken?