Discussion: View Thread

Target / Result Lab value correction

  • 1.  Target / Result Lab value correction

    Posted 06-19-2019 12:10 PM
    Does anyone know of a software based tool for correcting L*a*b* formulas to better match a target? - similar in practice to CurveX to adjust tone reproduction curves. I have a Lab target, a Lab formula and a Lab measured result. I could make a couple of "best judgements" and experiment but I thought there may be a tool that will do the calculation, possibly taking into account an .icc profile.
    -dan

    -- 


  • 2.  RE: Target / Result Lab value correction

    Posted 06-20-2019 03:48 AM
    Hi Dan-

    That describes ColorAnt (ColorLogic) pretty well. Correction to reference profiles, G7 and ISO 20654, TVI editing, averaging, smoothing, error correction, merging data sets, print order and channel editing with overprint recalculation, white and black point correction. CxF/X4-compliant. It applies edits directly to remission (spectral) data, too. It's not expensive. Distributor in NA is CrossXcolor. There may be other solutions out there, so check. This is the one I know.

    ------------------------------
    Mike Strickler
    MSP Graphic Services
    ------------------------------



  • 3.  RE: Target / Result Lab value correction

    Posted 06-20-2019 06:08 AM
    Edited by Juan Martorell Climent 06-20-2019 07:07 AM
    Hi Dan.

    I've a software (Color Lab) that does exactly this. You can download it from the next we-transfer LINK: Install it (Mac or PC) and I'll generate you a free testing license for 1 month.



    Best regards.

    ------------------------------
    Juan Martorell Climent
    ColorInLab Color Consulting S.L.
    Villarreal
    +34 964 523 278
    ------------------------------



  • 4.  RE: Target / Result Lab value correction

    Posted 06-20-2019 12:25 PM
    Hi Dan,

    If this is related to a brand color, then our ChromaChecker software has an iteration capability that allows you to predict what CMYK build is required to create the color (and if out of gamut it will flag it). Then it will allow you to export the predicted CMYK builds in a format that can be ingested by many of the RIPs, then you print, and measure using ChromaChecker Capture software, and it any colors are still off, ChromaChecker will iterate the CMYK builds by considering the original LAB, versus the new LAB and update the CMYK table...

    Is this what you are wanting? f so, email me and we can walk you through the steps. Thanks, Dave Hunter david@...

    ------------------------------
    David Hunter
    Principal
    Pilot Marketing Group
    Forest Lake MN
    (651) 717-0590
    ------------------------------



  • 5.  RE: Target / Result Lab value correction

    Posted 06-20-2019 02:16 PM
    Dave: Now I am also unsure what Dan is asking for! 

    Dan: What is the application? Are you trying to improve your printed match to a reference? Is it a conventional or digital device? Is the reference process or spot colors--both? Are you trying to improve your characterization of devices? Improve your process control? 


    ------------------------------
    Mike Strickler
    MSP Graphic Services
    ------------------------------



  • 6.  RE: Target / Result Lab value correction

    Posted 06-20-2019 03:30 PM
    Hi All, and thanks for your replies.
      So, I think what I was really looking for was something that works similar to how Curve's virtual press run is described. The environment is digital, Jetmaster 1890, CMYKOV.  If I have a source color (given Lab), a known output of that color (measured Lab) and the .icc profile the color was processed through, could I input new values and get feedback on the predicted live output so that I have the best chance of getting the color output needed without multiple iterations - using printer time/materials.
      Color Lab will build a test form of variations of a given color and I could then choose the swatch that best matches the target - as long as I have access to the values used for the various patches. This would work fine and probably only require a single print out of the machine.
      Of course, there are other considerations when a color does not match as expected. . . but I was looking for a good mallet to whack this particular mole. I suppose if I did color correction in Photoshop in Lab mode on a regular basis I'd have a better mental projection of an outcome. Maybe I should practice this a bit also.
    Best, Dan

    ------------------------------
    Dan Pierce
    Abbott-Action Inc.
    ATTLEBORO MA
    (781) 821-8200
    ------------------------------



  • 7.  RE: Target / Result Lab value correction

    Posted 06-21-2019 03:23 AM
    OK, so perhaps I understood more or less correctly after all. I use Colorant for this. One scenario is that I want to match particular reference colors--could be spot or process colors. I can input any output profile and see  report, patch by patch, with interactive displays showing the build needed to make the match and how the match will be affected by adjustments  in the build--a fancier form of the old "ring around" chart. It also calculated the most stable build based on the data, let's you limit the number of output channels, allows you specify dE00 or dE76 for the computation. It's very fast--calculates an entire Pantone library in maybe 10 seconds. Or a big reference set like GRACoL. 

    Or perhaps I want to smooth or idealize the measurement data as in a "virtual press run." I can apply a G7 correction with one click and see the impact on specific color matches as well as on TVI curves (alternatively available in ISO 20654 or a "default" dE-P visually linear scale). I can change print rotation "virtually" as well and have ColorAnt recalculate overprints, though that obviously is not relevant in inkjet printing. Its a pretty nifty tool one comes to depend on.

    ------------------------------
    Mike Strickler
    MSP Graphic Services
    ------------------------------



  • 8.  RE: Target / Result Lab value correction

    Posted 06-21-2019 08:39 AM
    Hi Dan,

    Are you trying to match a spot color? If so, can you give me the LAB you are putting in, the target LAB that you are trying to match and the output LAB reading you are actually getting? I have been messing around with this constantly and would like to test it out on something I am currently working on.

    Regards,

    Jeff

    ------------------------------
    Jeff Acoba
    Prepress Manager
    Cole Graphic Solutions
    Tacoma WA
    (253) 564-4600
    ------------------------------



  • 9.  RE: Target / Result Lab value correction

    Posted 06-24-2019 02:44 PM
    Hey Dan.
    What RIP are you using? There are different options for matching spot colors depending on the RIP.

    ------------------------------
    Dan Reid
    President
    RPimaging, Inc.
    Tucson AZ
    (520) 323-5855 X103
    ------------------------------



  • 10.  RE: Target / Result Lab value correction

    Posted 06-25-2019 12:16 PM
    Systems for matching spot colors can vary tremendously in their sophistication, from visual "ring-around" charts to those that rely on advanced modeling using spectral data. The more efficient the system the fewer samples it will need to calculate the match. Obviously, a system that requires printing of charts during production to "refine" spot color matches won't be as practical as one that gives acceptable matches automatically. And because package designs frequently have spot color tints and overprints--which appear in an unlimited and unpredictable variety--it is helpful to have a way to model these colors as well. An ideal system might employ methods similar to those used by ink formulation software using spectral measurement; a lesser but still useful system might work from Lab values and a certain amount of "high-low" sampling to interpolate a match. A good system also allows the input of measurements of tints and colors printed over black to capture opacity for calculation of overprints. Very few RIPs have this ability.

    The "virtual press run" is a slightly different idea, as it more about predicting the overall color match, for example, in images, rather than simulation of specific, single colors. Here smooth curve-fitting is the thing, and 2D curve adjustments in fact work really well. This is convenient for conventional printing, where standard press profiles/output conditions are common (if not a requirement), only curve and density adjustments are available, and test printing is impractical and expensive. Digital presses can be easily profiled and tend to give highly repeatable results, so virtual press runs don't apply in the same way.




    ------------------------------
    Mike Strickler
    MSP Graphic Services
    ------------------------------



  • 11.  RE: Target / Result Lab value correction

    Posted 06-20-2019 10:15 PM
    To my humble knowledge (Don Hutchison could step in), Curve 4 "Blend" tool would allow you to do this.
    If you have an ICC profile of your device in a linear state and done some G7 curves work, I believe that's what Blend will do for you, allow you to "preview" the result of the calibration. It is like being able to obtain a new device characterization based on a set of 1D curves without having to go back to re-characterize. Is this what you're after?

    ------------------------------
    Roger Breton G7 Expert
    www.graxx.ca
    ------------------------------



  • 12.  RE: Target / Result Lab value correction

    Posted 06-21-2019 09:49 AM
    Yes Roger, something like that. The problem and situation itself is probably a symptom implying that there's a can of worms somewhere with my name on it. I truly appreciate the active minds that are considering this question. I think the plain truth is that I was looking for instant potatoes, a way of pointing me in the right direction towards a one-iteration color adjustment or to back up my shot-in-the-dark Lab color correction.
    Happy Friday all, enjoy the start of the summer season (northern hemisphere anyway!)
    -dan

    ------------------------------
    Dan Pierce
    Abbott-Action Inc.
    ATTLEBORO MA
    (781) 821-8200
    ------------------------------