19 Replies Latest reply on Sep 19, 2017 4:18 AM by NB, colourmanagement

    Color Fidelity and Cross Polarization Copy Photography

    Auto-Repro

      Dear Adobe and Lightroom users.

       

      I´m using Adobe Lightroom as  raw conversion software for a number of different makes of cameras and lenses. Some of my tasks include photographing misc. metal objects such as coins and medals and also archaeological artifacts as well as two-dimensional graphic art and paintings with misc. different types of reflective surfaces, such as e.g. oil paintings with problematic reflective surfaces. The aim is to reproduce the color with high fidelity.

       

      I use misc. gray cards to to obtain neutral reproduction, but the problem of high reflection of highlights in eg. the above mentioned oil paintings and coins has brought me to experiment with cross-polarisation scenarios, where I use e.g. Kaiser reproduction light banks (fluorescent tubes) fitted with Kaiser  linear polarization filters - or similar tubes or small LED lights of misc. make with LEE (gel) filters fitted, in combination with misc. makes of polarization filters on the lenses I use.

       

      (Ideally I would not use different lights, lenses and cameras, but concentrate of one type of illumination, one camera lens and one type of image acquisition sensor (camera body), but the variation of sizes and shapes of objects or art, two dimensional or three dimensional makes it difficult to achieve such  scientific accuracy).

       

      The results in respect of reducing reflection in the highlights vary depending on painting  techniques and angles of light, but a main problem is the increased contrast and saturation of colors achieved in this process, which make the reproduction of colors appear dramatically different from shooting without the cross-polarization technique, on account of increased contrast and saturation caused by the filters (and to some minor  extent the variation of lights, filters, lenses and cameras in use).

       

      Now I´m thinking, that it must be possible to measure the reflection of colors from any given subject BEFORE application of cross-polarization in lighths/lenses and later compare this measurement with a second measurement of color reflection with the cross-polarization applied. Thus  making it possible in Adobe Lightroom to subtract the change of color tones in the process, and in this way render the color production more natural and similar to the original result without the negative effects of the polarization.

       

      Could anyone help me out in pointing to which color tone measuring device and measuring techniques that could bring me further in this effort of reproduction of color fidelity (saturation and hues) in combination with cross-polarization? Thank you very much for any replies/suggestions.

        • 1. Re: Color Fidelity and Cross Polarization Copy Photography
          NB, colourmanagement Adobe Community Professional

          Hi

          A few years ago I did some colourmanagement work with Christie's the Art Auctioneers in London.

          It was a big project, designed to improve and streamline the workflow for catalogue production - all the way from capture, editing and proofing to creating improved plate curves for their offset printing presses.

          There's an article here you may find interesting: http://www.colourmanagement.net/PDF/Christies_PrintMediaManagement.pdf

           

          But, back to photography, amongst other objects, the photographers at Christie's capture literally hundreds of paintings, many of which are captured with cross polarisation to minimise unwanted reflections - for example in brush textures on oil paintings.

          The project involved creating icc input (camera) profiles, in this case for use with Leaf software and Adobe Photoshop.

           

          How does this relate to you? Well, I found that although the profiles were accurate and streamlined workflow with standard lighting, cross polarisation changed the appearance significantly and so a second set of custom icc profiles were needed - created from Colorchecker SG targets captured with cross polarised lighting.

           

          That's a solution you could consider.

           

           

          I hope this helps

          if so, please do mark my reply as "helpful" and if you're OK now, please mark it as "correct" below,

          so others who have similar issues can see the solution

          thanks

          neil barstow, colourmanagement

          1 person found this helpful
          • 2. Re: Color Fidelity and Cross Polarization Copy Photography
            D Fosse Adobe Community Professional & MVP

            This isn't as complicated as you think.

             

            I work as photographer at an art museum and I use cross-polarization all the time. Yes, it increases contrast and saturation, but that's no more "correct" or "incorrect" than working without filtration.

             

            In both cases you need to set the correct tonal curve and endpoints with the help of a Colorchecker card. These cards have standardized Lab values for every patch, and it's fairly simple to correct for the gray patches in Lightroom. Problem solved.

             

            I leave the color patches alone. There's no way you will ever be able to achieve a full colorimetric match. If you ever did, the result would not be internally consistent and lose credibility. What you aim for is equivalent color. That will be accepted as "accurate" everywhere.

             

            ---

             

            If the result is going to print, you may need to compensate for paper color and max ink. This is getting a bit more complicated. Assuming your display calibration is targeted for those specific print conditions, you may need to pull the black and white points a bit (increasing total contrast range), away from their true Lab values, to get a good visual match from original to printed copy.

            1 person found this helpful
            • 3. Re: Color Fidelity and Cross Polarization Copy Photography
              D Fosse Adobe Community Professional & MVP

              One more thing: avoid fluorescent tubes or LED if you can. Spectral distribution is highly irregular, with sharp spikes and deep troughs. If you have to use them, they require custom camera profiles to account for this.

               

              With continuous light sources like tungsten or flash, you'll likely get the best and most predictable results with the Adobe Standard profile.

               

              If you do this a lot, a pair of studio flash units is a reasonable investment that quickly pays back. For three-dimensional objects you need to work with soft-boxes and other kinds of light diffusion, and this is much more easily workable with flash.

              • 4. Re: Color Fidelity and Cross Polarization Copy Photography
                NB, colourmanagement Adobe Community Professional

                Hi Dfosse, All,

                 

                 

                I agree about LED lighting, it doesn’t seem to be a mature technology as far as photographic lighting goes - as yet.

                [Maybe its down to cost to manufacture, LED does fine on a small scale, it seems, in internal monitor display illumination, e.g. Eizo Coloredge].

                Significant research also suggests that LED light may also contribute to macular degeneration. Not good given its increasing Ubiquity.

                 

                 

                Contrary to your own experience, I’ve found that SOME fluorescents can work well. Really well.

                I did some extensive testing with a London based freelance fine art photographer and we ended up using Kino Flo lamp units with tubes from a proofing luminaire manufacturer.

                I am told that they use Kino’s with Kino tubes at the Van Gogh Museum in Holland.

                 

                 

                Long exposures can be a disadvantage with fluorescents though, fine for a studio where all can be locked off and natural light excluded - So he does use flash when working on location.

                Again, as you mentioned with regard to fluorescents, all flash is not equal in its ability to make a painting look like it does in daylight.

                 

                 

                 

                 

                 

                 

                I hope this helps readers of this resource 


                if so, please do mark my reply as "helpful", so others who have similar issues can see the solution


                thanks

                neil barstow, colourmanagement

                2 people found this helpful
                • 5. Re: Color Fidelity and Cross Polarization Copy Photography
                  Auto-Repro Level 1

                  Thank You very Much for your detailed description of Christies solution, which may be beyond my budget. Unfortunately I don´t have the means of controlling all segments of the variations of print or web use my pictures may be used for, which ofcourse would be an ideal situation. I´d love to read a step by step manual for the work carried out in establishing all relevant input device icc profiles including the second measurement you mention, with the cross-polarisation applied, specifying the measuring hardware/software used in the process. I don´t see any references to equipment used anywhere in the link you have provided, but again thanks for your input. When I figure out which hardware/software I need to create input/output device profiles, I´m going to try this out, and conform my own humble equipment to the ISO 12647-2 (and THEN see if our external printers have any idea of what I´m talking  about :-D - Or maybe initially, I should ask them which ISO they comply to if any). But first I´ll get to next level of actually seeing this approach in function with my own Epson printer and calibrated monitor.

                  • 6. Re: Color Fidelity and Cross Polarization Copy Photography
                    D Fosse Adobe Community Professional & MVP

                    Make it simple. Use a ColorChecker passport when shooting. Here are the numerical Lab values for each patch. Setting Lightroom to show Lab values is the most convenient. Ignore the color patches - as I said you'll never get a colorimetric match, and you shouldn't try because it simply wouldn't look credible if you did. Work with the neutral patches.

                     

                    Macbeth_CC_Calc_Lab.jpg

                     

                    Next, make sure your monitor is calibrated to match the print conditions. This means setting a monitor white point that is a visual match to paper white, and a monitor black point that is a visual match to max ink density for the paper.

                     

                    With the monitor thus calibrated, what you see is what you get. What you see on screen, is what you will see in the final print.

                     

                    Now you can compensate visually for paper color and max ink. This is necessary because you are working from absolute Lab values, but these values will be modified by paper color and max ink. If you were working strictly visually from the start, this step would be redundant.

                     

                    Adjust color as needed. "Accurate" is a pie in the sky, you'll never get there. Think equivalent color, and it will be accepted as accurate.

                    1 person found this helpful
                    • 7. Re: Color Fidelity and Cross Polarization Copy Photography
                      Auto-Repro Level 1

                      Thanks for your reply. I had the impression it wouldn´t be too hard correcting the increased contrast and saturation resulting from Cross-Polarization using Lightroom, I just haven´t figured out how to measure the altered values, so I can subtract the difference. I´ll try your approach and see how it works out for me. Which particular Color Checker card are you using? I´d be very grateful for a short step by step workflow description of your practise using cross-polarisattion applied to an oil painting.

                       

                      Regarding your advice on light sources. I have visited another photographer, who also recommends tungsten over fluorescent tubes. After switching to mirrorless camera bodies, I have abandoned flash for photographing artefacts that need cross-polarisation, because continuous light enables me to see my output on a monitor while composing the light. I exclude most other light in the studio. I still use flash extensively for portraits or less complicated 3 dimensional artefacts. The photographer I mention using tungsten is specialising in art e.g. oil paintings and dislikes cross polarisation precisely on account of my problem :-D So for him, it´s much simpler to use tungsten bulbs in relatively small, round reflectors with no polarisation involved at all. He has a very special studio, which he has painted in a variation from dark, neutal (matte) grey on the floor to medium light grey ascending to the ceiling. He then points his reflector (often just one) indirectly to the walls/ceiling, and the light is then bounced back onto his canvas. He´s able to modify his angle of light until his reflecting highlights in the oil are practically gone! :-D I´m very envious, but I haven´t got as much space available to create such a specialised studio. I´m forced to work in much less space with no chance of designing such a large, specialised, indirect, neutral bouncing box. So the next best solution is: Fluorescent bulbs. The bulbs used in Kaiser´s repro light banks are specially designed for repro work, and don´t display nasty spikes as regular bulbs do, so I achieve very nice "equivalent color" as you say, but the Kaiser banks are too small for larger canvas photography, so I also use KINOFLO bulbs as recommended below. It would be possible to arrange a vertical array of tungsten reflectors, which might distribute lights as even as the KINO/Kaiser bulbs, but it would practically impossible to apply the Lee polarisation gel I have to use on account of my direct lighting as opposed to the ideal indirect solution I described above. Because of the heat emisson from tungsten, actually some conservators wouldn´t let me get near their arcahaological originals of e.g. textiles, if they were aware of how hot tungsten gets. If you use tungsten for photographing, it´s necessary to have quite long distance to the subject and preferably indirect lighting as described above, but if you work for some hours with tungsten, the heat emission has a consequence on the room tempperature. So I have been left to use fluorescent, continuous light, which is much better than it´s reputation using KINO or Kaiser. I´m sure you have a point about LED´s, but as you said, for most puposes, what´s necessary to achieve is equivalent light, not necessarily dead accurate for all applications.

                      • 8. Re: Color Fidelity and Cross Polarization Copy Photography
                        D Fosse Adobe Community Professional & MVP

                        https://forums.adobe.com/people/NB%2C+colourmanagement  wrote

                         

                        LED does fine on a small scale, it seems, in internal monitor display illumination

                        It works because it's an output device, and the eye interprets it in a specific way (metamerism).

                         

                        A camera sensor is a different animal. Try to photograph your Eizo, and you'll see the difference.

                         

                        Flash output is spectrally continuous and more or less uniform, like daylight. A small temperature shift is very different from a highly fluctuating spectral curve like that of LED or fluorescent. You don't need specific camera profiles for midday, sunset, flash or tungsten - those are just differences of white balance, all handled by the same profile. But for LED and fluorescent you need custom camera profiles to compensate for the spikes and valleys.

                        2 people found this helpful
                        • 9. Re: Color Fidelity and Cross Polarization Copy Photography
                          Auto-Repro Level 1

                          I agree, but it´s so much easier to light a painting with 4 KINOFLO bulbs in two reflectors, and there´s no manufacturer that makes affordable pol gels that would fit e.g. a large striplight reflector. I have Chimera striplights btw, and the distribution of light from the centre of a striplight reflector to the far sides is nowhere near as even as a fluorescent bulb...I see you use Gretag Macbeth color checker. I thought Gretag Macbeth had gone out of business?

                          • 10. Re: Color Fidelity and Cross Polarization Copy Photography
                            D Fosse Adobe Community Professional & MVP

                            It's just an image I just found on the web, probably old. The numbers haven't changed, and they are still valid.

                             

                            I have all this written out in my office, also recalculated to Lightroom % readouts, but I'm at home now.

                             

                            But yes, large sheets of polarizing film aren't easy to come by. I had to order a 3x1 meter roll from abroad somewhere (can't recall) - but it arrived quickly and wasn't really expensive. Google it.

                             

                            Anyway, for flat painting surfaces (sans frame), you really want to use point light sources. That in itself minimizes reflections. Anything three-dimensional, from frames to thick impasto paint to sculptural objects, require diffusion of some sort, and that's where you need larger sheets.

                            1 person found this helpful
                            • 11. Re: Color Fidelity and Cross Polarization Copy Photography
                              Auto-Repro Level 1

                              Thanks Neil, I alredy have a fine set of Kaiser repro banks 5400 K, but they are too short for even distibution of light on large canvas´s. I didn´t get the Kinoflo banks approved in my budget (the city needs to save :-D), so I tried to mimic a simpler set using KINO bulbs in modified reflectors designed for mounting in the ceiling :-D but I made them hang from tripods instead. It´s not an ideal setup, but they work well with diffuser and polarizer gel clipped on the metal cabinet :-D I do get fairly long exposures though, so maybe I do have to get the KINOFLO sets with four bulbs. First I must sort out the contrast/saturation problem :-D

                              • 12. Re: Color Fidelity and Cross Polarization Copy Photography
                                Auto-Repro Level 1

                                Thanks, I´d love to get a hint on where to order polarizing gel which is "not really expensive" :-D. I have been googling a lot, but I found only two rather expensive manufacturers of polarizer gel, Lee being the "cheapest", so it would be great, if you could come up with the name of that "foreign" manufacturer.

                                • 13. Re: Color Fidelity and Cross Polarization Copy Photography
                                  D Fosse Adobe Community Professional & MVP

                                  Ah, to be perfectly honest, I'm employed to do this stuff, and my employer picks up the bill. But I usually remember the invoices that cause trouble, and this wasn't one of them

                                   

                                  It's several years ago and I'm still using the same sheet, cut up to fit my soft boxes and reflectors.

                                   

                                  As for the colorchecker, B+H sells it for $100.

                                  colorchecker.jpg

                                  • 14. Re: Color Fidelity and Cross Polarization Copy Photography
                                    NB, colourmanagement Adobe Community Professional

                                    Hi Auto-Repro

                                    This is getting to be a long thread so I’ll quote your question here.

                                     

                                    D. Fosse is providing some good real user advice, he does this every day it looks like.

                                    you write:

                                     

                                    Thank You very Much for your detailed description of Christies solution, which may be beyond my budget. Unfortunately I don´t have the means of controlling all segments of the variations of print or web use my pictures may be used for, which of course would be an ideal situation.

                                    Inability to control the print does not mean its not worth getting decent input and using a calibrated screen.

                                     

                                    VERY few people control process end to end, that’s what device independent colour management is all about, however, this instance was a privilege to work on as it proved [to me] once and for all that full end to end works fabulously. IMO belief is great, proof is better.

                                     

                                    I´d love to read a step by step manual for the work carried out in establishing all relevant input device icc profiles including the second measurement you mention, with the cross-polarisation applied, specifying the measuring hardware/software used in the process.

                                    It took many days of experimentation, I used the colorchecker SG [semigloss] (still available from my German contacts although Gretag were swallowed by X-Rite years ago] .

                                    They used 5’ strip flash, with and without polarisation depending on the job.

                                    The SG target was measured [by me] with and without polarisation to create each reference file [yes, it made a significant difference].

                                    At that time measurement was done with a Spectrolino.

                                    On that project I used Profilemaker software, latterly I have preferred basICColor software.

                                     

                                    I don´t see any references to equipment used anywhere in the link you have provided, but again thanks for your input.

                                    Pleasure, see above

                                     

                                    When I figure out which hardware/software I need to create input/output device profiles, I´m going to try this out, and conform my own humble equipment to the ISO 12647-2 (and THEN see if our external printers have any idea of what I´m talking  about :-D - Or maybe initially, I should ask them which ISO they comply to if any).

                                    As you are US, maybe GRACoL, its very close to ISO 12647

                                     

                                    BUT that does not affect input a or screen calibration at all.

                                     

                                    But first I´ll get to next level of actually seeing this approach in function with my own Epson printer and calibrated monitor.

                                    You will

                                     

                                    Feel free tot me directly if I helps - google me

                                     

                                    I hope this helps

                                    if so, please do mark my reply as "helpful", so others who have similar issues can see the solution

                                    thanks

                                    neil barstow, colourmanagement

                                    • 15. Re: Color Fidelity and Cross Polarization Copy Photography
                                      NB, colourmanagement Adobe Community Professional

                                      Hi D.Fosse,

                                      I am sure you realised this, but I don't think you mentioned it.

                                      it's interesting, I feel, to observe that since the camera captures RAW and the RAW converter output is RGB, an input profile has to be used to translate between that and and the Lab info we see in Lightroom.

                                      So the Lab numbers are not "straight lab as captured" there's no such thing - I thought some users may find it interesting to consider that.

                                      Why?

                                      I've heard input profiling software makers talk about setting camera exposure to give certain Lab values when capturing camera targets [e.g. my favourite Colorchecker SG] - but, meanwhile, those lab values are being created [in a conversion from camera RGB to Lab] courtesy of of the existing input (camera) icc profile [or often the working space profile in the case of photoshop assessment.].

                                      As that icc profile is going to be replaced by a new profile made form the target capture, Lab values for un-profiled target readings re rather irrelevant.

                                      • 16. Re: Color Fidelity and Cross Polarization Copy Photography
                                        NB, colourmanagement Adobe Community Professional

                                        Hi D.Fosse

                                         

                                        you write: "Flash output is spectrally continuous and more or less uniform,

                                        like daylight."

                                         

                                        I've seen significant differences in capability between

                                        different manufacturer's flash outputs.

                                         

                                        I've no archived readings, but have come across this which appears to be of

                                        interest:

                                         

                                         

                                         

                                        and daylight albeit D65:

                                         

                                         

                                         

                                        apologies for the resolution

                                         

                                         

                                        neil barstow, colourmanagement

                                        • 17. Re: Color Fidelity and Cross Polarization Copy Photography
                                          NB, colourmanagement Adobe Community Professional

                                          Trying again for the pics:

                                          Hi D.Fosse

                                           

                                          you write: "Flash output is spectrally continuous and more or less uniform,

                                          like daylight."

                                           

                                          I've seen significant differences in capability between

                                          different manufacturer's flash outputs.

                                           

                                          I've no archived readings, but have come across this which appears to be of

                                          interest:

                                          Rare_gas_flashtube_spectral_outputs.JPG

                                           

                                           

                                          and daylight albeit D65:

                                          D65 spectrum_sunlight.jpg

                                           

                                           

                                          apologies for the resolution

                                           

                                           

                                          neil barstow, colourmanagement

                                          • 18. Re: Color Fidelity and Cross Polarization Copy Photography
                                            D Fosse Adobe Community Professional & MVP

                                            https://forums.adobe.com/people/NB%2C+colourmanagement  wrote

                                             

                                            I've heard input profiling software makers talk about setting camera exposure to give certain Lab values when capturing camera targets [e.g. my favourite Colorchecker SG] - but, meanwhile, those lab values are being created [in a conversion from camera RGB to Lab] courtesy of of the existing input (camera) icc profile [or often the working space profile in the case of photoshop assessment.].

                                            As that icc profile is going to be replaced by a new profile made form the target capture, Lab values for un-profiled target readings re rather irrelevant.

                                             

                                            You're not interested in the technicalities of how the sensor records it, or how the data are processed and demosaiced in the raw converter.

                                             

                                            You're interested in what comes out of the raw converter, encoded into an output color space.

                                             

                                            Yes, Lab numbers are entirely relevant.

                                             

                                            I honestly think you're making this way more complicated than it needs to be.

                                            • 19. Re: Color Fidelity and Cross Polarization Copy Photography
                                              NB, colourmanagement Adobe Community Professional

                                              Hi D. Fosse

                                              My post:

                                              I've heard input profiling software makers talk about setting camera exposure to give certain Lab values when capturing camera targets [e.g. my favourite Colorchecker SG] - but, meanwhile, those lab values are being created [in a conversion from camera RGB to Lab] courtesy of of the existing input (camera) icc profile [or often the working space profile in the case of photoshop assessment.].

                                              As that icc profile is going to be replaced by a new profile made form the target capture, Lab values for un-profiled target readings re rather irrelevant.


                                              you write ( am comments I line):

                                                You're not interested in the technicalities of how the sensor records it, or how the data are processed and demosaiced in the raw converter.

                                              Correct

                                               

                                                You're interested in what comes out of the raw converter, encoded into an output color space.

                                               

                                                Yes, Lab numbers are entirely relevant.

                                              OK, now it is getting complex.

                                               

                                              In assessing a capture workflow “as is” i.e. - how good is my capture in the assigned colourspace then, yes Lab has great relevance as it allows comparison to the colour checkers original values (courtesy of the excellent Bruce Lindbloom).

                                              BUT, in this event, when looking at Lab we are, essentially, assessing both capture and colourspace.

                                              [assign an alternative colourspace and you get a different appearance and  different Lab values - therefore the colourspace is definitely involved]

                                               

                                              Lab numbers for target patches are calculated from the captured RGB using an input ICC profile or workingspace.

                                              A RAW converter doesn’t make Lab, it makes RGB . To view or use that RGB [or too read off Lab numbers] we need to assign an icc profile (or maybe the raw converter does that for us).

                                               

                                              Think about it - what I was referring to was when making a camera (input) profile - I need to first assess captured target tonal range to get the best profile results. Lab doesn’t help me here as I have no input profile as yet

                                              The captured target is RGB, so what is correctly used to assess target range is RGB values.

                                               

                                              What I mean is that, in THIS scenario, any Lab values we see at this stage (e.g. in Photoshop] are worthless because they use a current icc profile NOT a correct input profile, we didn’t make that yet.

                                               

                                                I honestly think you're making this way more complicated than it needs to be.

                                              You have a way of doing this which is fine. Your posts here indicate a high level of technical expertise.

                                               

                                              But I am discussing the technicalities of input profile making [because the OP asked me] and I am certainly not going to get into an argument about the relevance of Lab values when assessing target capture when making an input profile.

                                               

                                              I hope my explanation above helps make more sense of my original post

                                              I think we’ve gone far enough into the technicalities now

                                               

                                              Yours

                                              neil barstow, colourmanagement