Alan Goldhammer wrote:
ArgyllCMS but I want to see if I can do better). My question is the utility of using the '-g' parameter in creating a set of gray patches. I do all my B&W printing using the Epson ABW print driver and have profiles for that driver that Eric Chan has created by reading a step chart of B -> W densities. Is there value of including a set of gray steps when creating a profile that will be used only for color printing on this printer?
Possibly. Some people have said they get better results if they add some explicit grey patches. The general answer is that you will always get a better result if you measure the color(s) you want to be most accurate. For input profiles this can be pretty simple (i.e. often in proofing there are test colors that are defined in terms of the input device space values), but for output devices there is something of a catch 22. If (say) you want a good neutral grey of Lab 50,0,0 then you should print a test patch of that color. But you don't know what device values will create that color without a profile. So to do this sort of thing you really need a two step (or iterative) approach - create a profile with no special patches, lookup the device values for the colors that are most important, add them to the test set, and then re-profile. There's no simple path for doing this in Argyll currently, but it is possible using xicclu and some editing of the .ti1 file. If you assume that your device has roughly known behaviour, (ie. and "RGB" printer where R=G=B give reasonably close to grey output, or you know your K only is neutral and you are going to use a lot of it in the profile), then you may get away with adding simple device grey values in targen. Graeme Gill.