6 Replies Latest reply on Jul 20, 2008 4:50 AM by (TW_Miller)

    How close is close enough?

      How close is close enough?

      I'm wondering how close is close enough, when trying to match our vendor's colors.

      We have been tasked with matching our vendor who is matching towards Gracol.

      Our vendor has previously been fairly out of spec with their gray balance being overall too red, but they are moving much closer to the Gracol standard.

      We have been using our HP z3100 to match their proof outputs. So far, after having them run out a target for me to profile and match, I now have a Delta E of 1.88 average when comparing our P2P25 G7 target output to theirs.
      Our deep blues and dark Magentas however, are about a max of 7 Delta E off of theirs.
      Comparing their proof sheet to ours is visually very close. It's just in the deep blues where their colors have a slight magenta cast to them.

      My concern is that we are matching our vendor's representation of Gracol and we don't have a real third party P2P25 target from a tightly controlled press sheet for verification as to who is closer to the standard.
      The vendor might be a little too red, but we might be a little too gray balanced. Our three color gray might be too neutral.

      I'm using a spectro with Profilemaker Measure tool for comparing Delta E. I'm also using no RIP, just custom profiles through the RGB printer's internal RIP which doesn't give me much, if any adjustments for calibrating linearity.

      Our company is willing to buy a dedicated RIP so that we can match the vendor better, but my concern is that we might not get much better than the Delta E of 1.88 that we have now, and if we're going to spend that extra $15k, I want to make sure we will be closer.

      There is also the possibility that we can "toy-with" the printer using different paper settings to get a better match to the vendor's linearity by choosing a paper setting which gets us closer to theirs.

      The real main difficulties are trying to speak color management to the people who make the cost decisions and have no comprehension of what we're talking about.

      Does anyone have thoughts on whether I'm doing it the difficult way, or should a new RIP make it easier for us?
        • 1. Re: How close is close enough?
          John Danek Level 4
          I take it you do not own the HPz3100ps/GP? The gray balance could be critical, especially when you are trying to match a particular color. It would be nice if you spoke about your vendor's printer/proofer. The HP is a 12-cartridge photographic inkjet printer and, often times, a photograph's color gamut is going to be much wider than a graphic arts typical offset press. If their proof is showing warmer grays, then it could be their color gamut is not as wide as yours. I'm not sure which version of the z3100 you're using, but I do know that the z3100ps/GP is a very accurate proofer and because it has an internal spectrometer and RIP, it can calibrate without you having to put out more cash for a third party RIP. If you have the basic z3100, then perhaps you should look at a third party RIP ( you can probably get an upgrade through HP ).
          • 2. Re: How close is close enough?
            Level 1
            We have the HP z3100ps/GP with the internal spectro. We also have an EFI ES1000 profiling package. this is what I use for comparing the P2P25 targets.

            The vendor is using an Epson. I'm not sure if they're using K2 or K3 inks.

            The calibration in the Z3100ps is automatic. No user defined controls other than the Color options dialog in the print driver.

            We're an average 1.88 Delta E off of the vendor and 7 Delta E max off of only one color. There are about five patches which are over Delta 5. All in the Deep blues and Dark magentas.

            I think we're really close. I'm just not so sure a new RIP will bring us even closer.
            • 3. Re: How close is close enough?
              John Danek Level 4
              I'm not sure how many patch swatches you're talking about, but they say the-more-the-better. If you are talkning about 6 out of 24 patches being off, then you've got a complaint. But, if it's 6 out of 600, then that's forgiveable. I've been in shops where there are two identical printers, side-by-side, calibrated, same paper, etc., and they do not match eachother. You're talking about two different manufacturers using what could be very different inks. So, you may be getting the best results you can within the technology. If you've got unlimited time and money, then it would be an interesting test to see if you could get a better match with a third party calibration system.
              • 4. Re: How close is close enough?

                one standard for the validation of proof prints is
                the UGRA/FOGRA MediaWedge:

                The tolerances are fairly large, here named 'Max' in
                units deltaE.

                A qualified proof print contains the printed MediaWedge
                (in fact a couple of patches) and the report for the
                measured results, as shown in the mentioned doc.

                It's required that the RIP prints the MediaWedge together
                with the images and that it provides an integrated
                measuring module, for instance for Eye-OnePro, the error
                calculations and the test report.

                Best regards --Gernot Hoffmann
                • 5. Re: How close is close enough?

                  You SHOULD be able to get significantly better results using a RIP. I would say that on the P2P25 target you should be able to get an average well under 1 and max patch around 3. You can get an excellent RIP for under 5K, so don't think you have to spend 15K.

                  Although the P2P25 is a great target for matching, you can download a free ISO control strip from IDEAlliance (http://www.idealliance.org/swop/ ) that comes with reference data for GRACoL_Coated1, SWOP_Coated3 and SWOP_Coated5. If you and your vendor both output the strip, then you can compare to each other AND to the spec to see who is closest. Maybe you are spot on!

                  Good Luck
                  -Todd Shirley
                  • 6. Re: How close is close enough?
                    Print the IT8 7/4 target. Read with spectro in measure tool and compare to the GRACol data set. This will tell you how close you are proofing to the specification. Have your vendor proof the IT8 7/4 measure and compare to the GRACol data set and this will tell you how close your vendor is to the spec.