Hi Wim, this is correct but you have to realise that commercial profile packages don´t offer a calibration procedure at all. That is because the vendors in most cases asume that their products will be used in a professional production environment. That means that output devices are driven by software that offers a calibration utility. This is IMHO also the reason why for home-printers this solutions are worthless. When you create a profile as I described, apply the profile using Photoshops Convert to profile utility, Photoshop simply converts your RGB image to your RGB-profile compensating the visual shift, or keeping the images original look. At this step you could apply softproofing by creating an apropriate setup. This ist basicaly the way profiling in PS works. Applying the calibration curve changes the image because PS can´t compensate the curve-behavior. This what you get is the image "how your calibrated printer need´s it" in order to produce the correct result. A workarround for softproofing (or checking the quality) could be: 1. Open your original RGB-image 2. Create a softproof-setup with your argyll-profile and apply it to the image. Create a copy and open it in PS 3. Convert to your output profile, apply the calibration curve and print 4. Compare the print to the softproof-image. The results should be very close to each other. if not -you may have a problem with your monitor profile. OK. The whole process may be confusing in the begining. But, the first serious colour-shift will convince you very fast of it´s advantage. It´s faster, easiser and cheaper to create a new printer calibration, than creating a new profile from scratch. Cheers Vladimir 2011/10/20 Wim Hertog <nertog@xxxxxxxxx>: > Hi Vladimir, > > I'm a bit confused now. Alan just wrote that he only profiles his Epson 3880 > and after that his prints match his calibrated monitor. My experience is > that the profile just characterizes the printer but does not change the way > it prints, in other words, it can not match print to monitor. Am I wrong in > this? > > The process you described is what I did and I have the same results. After > the extra step of applying the correction curves (either in the profile or > as a PS curve) the whole image changes but prints quite ok. > > In every other profiling package however, I can match my print to monitor > using just 1 step: converting to the printer profile. This profile includes > the correction curves and applies them without making the softproof in PS > unuseable. In other words: 1 confusing step less when printing. > > Could you clarify this a bit for me? > > Thanks, > Wim > > > > 2011/10/20 Vladimir Gajic <vgajic67@xxxxxxxxxxxxxx> >> >> Hi Wim, >> >> the Idea of calibrating the printer before building a profile is simply to >> keep your device in a constant printing condition. The process could also >> look like this: >> >> 1. You create a printer calibration witch results in a .cal-file. During >> the calibration process you also can create a Photoshop curve, witch can >> illustrate the whole procedure IMHO mutch better >> >> 2. You generate a RGB-target for your profile, applying the .cal you >> created in the previous step. You also can skip adding the .cal using >> printarg, simply open the file in PS and apply the generated PS-curve BEFORE >> printing the file. >> >> 3. Generate the profile >> >> 4. The printing process could look like this: >> - open your image and convert to the device profile. You will notice that >> the image looks correct. This is also your softproof. >> -now apply your curve. The image changes in a strange way, but will be >> printed correctly. >> >> The Idea behind: profile once, calibrate many. >> >> Your printer may change in time. Any cartrige replacement, even if you are >> using original inks, can produce colour shifts. The same applies to the >> substrates you are using. In that case it's enough to recalibrate your >> printer generating a new .cal and PS-curve. >> >> There are also different ways to work with an calibrated workflow (e.g. >> applying .cal using cctiff, or linking the .cal directly to the profile). >> Anyway, the described procedure was for me a good starting point for >> understanding the whole stuff. >> >> Hope this helps. >> >> Cheers >> Vladimir >> >> >> -- Gesendet von meinem Palm Pre >> ________________________________ >> Wim Hertog <nertog@xxxxxxxxx> schrieb am 20.10.2011 10:02: >> >> Hmm, so the profiling step alone should do the trick then? I thought >> profiling only characterized the printer and you needed the calibration step >> in order to actually change the printing behaviour. If the profiling step by >> itself is enough to create prints matching my (with argyll) calibrated >> monitor, I must be doing something wrong somewhere... >> >> After following the tutorial and profiling the printer the gamut shape and >> softproof look perfect. Very similar to what I get from PM5. The printout >> using this profile results in a horrible yellow-brown cast though. I follow >> my usual workflow while printing: windows CM is turned off in the canon >> driver and photoshop manages colours using the generated profile. I'm pretty >> sure it's not double profiling anywhere. >> >> I must be doing something wrong somewhere but I literally read the >> tutorial a 100 times and tried everything and always get the same result: a >> strong yellow brown cast together with totally blocked shadows. >> >> Anyone has any idea what's happening or....a link to another tutorial to >> double check? >> >> Wim >> >> >> 2011/10/20 Graeme Gill <graeme@xxxxxxxxxxxxx> >>> >>> Wim Hertog wrote: >>> > Now, the above workflow results in some strange outcomes: the colours >>> > of >>> > the softproof in photoshop are completely off (the same happens when I >>> > convert to above generated icc file). The image prints ok (ok doesn't >>> > mean as good as I want though), nothing like the softproof shows. >>> > However, when I don't add the .cal file to the icm (last step), the >>> > softproof is perfect but the actual printed image is horribly wrong >>> As suggested in the tutorial, get just profiling working first. There >>> are too many variable otherwise, and the first thing you do in diagnosing >>> a problem is break things down into individual steps anyway. >>> >>> Graeme Gill. >>> >> >> >> > >