13 Replies Latest reply on Mar 25, 2014 1:01 PM by Alp Er Tunga

    Linear workspace

    Alp Er Tunga Level 1

      Is there any negative side effect to use linear workspace with ProPhoto primaries in Ps (as in Lr)?

        • 1. Re: Linear workspace
          jdanek Level 4

          None that I know of.

          • 2. Re: Linear workspace
            Alp Er Tunga Level 1

            What about using Rec. 2020 primaries with linear gamma as the RGB workspace in Ps?

             

            Its primaries are in spectral locus and so has a little bit smaller gamut than the ProPhoto, but bigger than AdobeRGB. We will never get ProPhoto printers and displays, but I hope we will all have Rec 2020 displays some day, maybe it will be the next standard after sRGB.

             

            As I'm just printing with my inkjet at my home as a hobbyist, I know I really don't need to think about my present workspace so much.

            Just want to learn your thoughts, especially about the disadvantages of using a custom linear workspace, if there is one.

            • 3. Re: Linear workspace
              jdanek Level 4

              How did you determine "Rec 2020" has a larger gamut than "Adobe RGB", but not as wide as "ProPhoto RGB"?  ProPhoto should not be assigned to your monitor or display.  There are photographic printers that can take advantage of ProPhoto's fairly wide gamut.  I'm not familiar with Rec 2020, where does that come from?  When discussing linearization, linear profiles are built when calibrating a printer and creating a custom profile for the RIP ( where applicable, Wasatch comes to mind ).  Which RGB colorspace you use depends on what type of files you print.  A photographic workflow would definitely benefit from a ProPhoto RGB profile ( application ), as well as the correct Rendering Intent.  If this is a hobby, feel free to experiment along and see how different color settings and profiles affect output.

              • 4. Re: Linear workspace
                Alp Er Tunga Level 1

                Thanks for your thoughts jdanek.

                 

                Rec. 2020 is the specification for UHDTV. Its gamut is in the following figure.

                 

                476px-CIExy1931_Rec_2020_svg.png

                 

                 

                With ProPhoto, there is always a possibility that we save our images with some unvisible colors. Our eyes should evolve more for seeing them, but I don't expect such a thing in my life.

                But with Rec. 2020 or any other thing like it, wide enough and primaries inside the human spectral locus, I think that I may see all the colors in my files in five or ten years, maybe less.

                 

                I decided to use a linear workspace, but I can not decide which primaries I should use. What about the following primaries (white line)? ))

                 

                476px-CIExy1931_Rec_2020_svg.jpg

                • 5. Re: Linear workspace
                  G.Hoffmann Level 3

                  (1)

                  http://www.google.de/url?sa=t&rct=j&q=&esrc=s&source=web&cd=7&cad=rja&uact=8&ved=0CFwQFjAG &url=http%3A%2F%2Fphoto-lovers.org%2Fpdf%2Fcolor%2Fromm.pdf&ei=8kMYU7LyKqKlyAPC6IHABQ&usg= AFQjCNFg6qPEVMKMs5XhcwwHFWZqzjarkg&bvm=bv.62577051,d.bGQ

                  (2)
                  http://docs-hoffmann.de/swatch16032005.pdf

                  p.17

                   

                  ProPhoto RGB is the same as Kodak ROMM RGB, see (1). With gamma=1.8 and a linear slope

                  at zero. These were the design guidelines:

                  a) reproduce additionally to sRGB and Adobe RGB vibrant oranges and yellows. This requires

                      that the green primary is outside of the horseshoe.

                  b) cause very little hue variations for saturation manipulations of the nonlinear data.

                      Perhaps the inventors thought at that time that it might be recommended not to transform

                      into the linear space because of speed considerations and round-off errors.

                  c) make available with different bit depths per channel, 8 bpc and more.

                   

                  Your black system is reasonable, practically the same as my OptiRGB, see (2) p.17,

                  which contains almost all surface colors, a gamut which is nicely represented by the set of

                  Pantone spot colors.

                  See as well (1). There are no surface colors in the upper green area of the horseshoe.

                  Therefore, your white system is, please excuse me, very bad. You're loosing the yellows and

                  oranges, but winning not much, just a better cyan reproduction.

                   

                  In my humble opinion this demand for huge RGB spaces is somewhat unreal, perhaps guided

                  by the hype about breathtaking sunset photos and other crude effects with 'pushed' colors.

                  Under all circumstances it's necessary to understand, which colors appear in real scenes

                  and how the preferred reproduction (Hunt) should be – really as colorful as possible?

                   

                  Best regards --Gernot Hoffmann

                   

                  Edited by me

                  • 6. Re: Linear workspace
                    Alp Er Tunga Level 1

                    Three color spaces with the Pointer RGB.

                    ProPhoto RGB, Rec. 2020 and Optimal RGB, respectively.

                    figure18_ProPhoto%20RGB%20and%20pointer%20CIE1931.jpgfigure20_Rec.jpgfigure_a-optimal pointer cie1931.jpg

                     

                     

                    By the way, the gamut of Pointer RGB is the theoretical boundary for printer profiles, isn't it?

                    I mean, there is no reflective color in real world at the outside the gamut of Pointer RGB?

                    • 7. Re: Linear workspace
                      G.Hoffmann Level 3

                      Hello,

                       

                      never heard the name Pointer RGB, but this somewhat irregularly shaped contour can be explained.

                      It is the gamut of optimal surface colors under a certain illuminant (for printers D50).

                      Optimal surface colors have reflectance spectra with value 1 as one solid block (and 0 elsewhere)

                      or as a combination of two solid blocks for magentas (as well one solid block, if we bend the spectrum

                      to a circle).

                      So far we would get a smooth contour, but together with a Dxx Illuminant the contour looks rough

                      (should be smooth for Illuminant Equal Energy or Illuminant A).

                      Now we see immediately, that this surface color gamut changes, depending on the Illuminant,

                      or simply on the scene light, of course not exceeding the horseshoe contour.

                       

                      Some remarks on the choice of primaries:

                       

                      For a working space (like sRGB, Adobe RGB, ProPhoto, CIELab) any coordinate system can be used.

                      The accuracy is affected by the number format (bit depth for integers or type of floating point format).

                      It's probably annoying, that in some cases arbitrary numbers (though elements of the set of numbers

                      of binary coding) are nonsensical, like blue and red primaries and their neighbours for ProPhoto

                      or many numbers in CIELab, for instance 0L* 10a* 20b* or 90L* 30a* 20b*.

                      Nonlinearities affect the accuracy further.

                       

                      For monitors the situation is very different. The primaries have to be chosen so, that for each primary

                      with limited radiometric power a significant photometric output is achieved. For instance red at 700nm

                      is almost invisible.  

                       

                      Best regards  --Gernot Hoffmann

                      • 8. Re: Linear workspace
                        Alp Er Tunga Level 1

                        I have one more question and then I will not ask any other before finishing "Measuring Color" )

                         

                        A general question.

                        As a working space, do you advise Optimal RGB with linear gamma over any other working space commonly accepted and used today (like sRGB, AdobeRGB, ProPhotoRGB, etc.)?

                         

                        Or more generally, which color space should be used as a working space for photographic work and for the optimal results?

                        • 9. Re: Linear workspace
                          G.Hoffmann Level 3

                          I'm really happy about these questions. My private RGB space OptiRGB is as arbitrary as

                          – many years ago – BruceRGB, which had been propagated as a big progress.

                          OptiRGB was an exercise, how to pack the information into an ICC profile, and of course

                          with a reasonable choice of the primaries.

                           

                          A friend of mine is a well-known photographer, who had contributed very much in the field

                          of landscape photography to calenders and travel books:

                          https://www.google.de/search?hl=de&as_q=Fotograf&as_epq=Norbert+Kustos&as_oq=&as_eq=&as_nl o=&as_nhi=&lr=lang_de&cr=&as_qdr=all&as_sitesearch=&as_occt=any&safe=images&as_filetype=&a s_rights=

                           

                          Some years ago we had all these discussions, and my part had been mainly the introduction

                          into  Lab editing, soft proofing and generally into the production of print-ready CMYK files,

                          or in other words, how to respect the limits of offset printing.

                          In cooperation with editors of calendars and books it turned out as the best choice, starting

                          with RAW, to prepare the files in AdobeRGB, but of course after modifying and verifying the

                          printability by soft proofing.

                          By the way: the rendering of these photos is always done by preferred reproduction (just as

                          we like, but our opinions differ occasionally), never by colorimetrical reproduction.

                           

                          Best regards --Gernot Hoffmann

                          • 10. Re: Linear workspace
                            thedigitaldog MVP & Adobe Community Professional

                            G.Hoffmann wrote:

                            My private RGB space OptiRGB is as arbitrary as

                            – many years ago – BruceRGB, which had been propagated as a big progress.

                            IF you're referring to the Bruce RGB built by the late Bruce Fraser, he abandoned it fairly quickly after the idea of RGB working spaces were introduced in 1998 in favor of ProPhoto RGB. Not sure what progress you refer to.

                            • 11. Re: Linear workspace
                              Alp Er Tunga Level 1

                              Thank you Gernot.

                               

                              I have many more questions in my pocket I would like to ask here ))

                              But, let me finish the book first of all.

                              There are so many new terms for me, I need to go slow and digest them correctly.

                               

                              Best wishes.

                              • 12. Re: Linear workspace
                                Bill_Janes Level 2

                                One disadvantage with a linear working space is that it is not perceptually uniform: a small change of levels at lower luminosities produces a larger perceived change than a similar change at higher luminosities. The sliders in LR would be less precise and "touchy" at the low end. This was commented on some time ago in these forums by

                                Thomas Knoll himself.

                                 

                                The human visual system is approximately logarithmic rather than linear, and gamma encoding causes the steps in the color space to be more perceptually uniform. LR does use a linear working space, but the sliders, readouts and histograms are in terms of the virtual Melissa color space which uses a sRGB tone curve. With this setup, one is working in a linear space but the user interface is that of a sRGB space.

                                • 13. Re: Linear workspace
                                  Alp Er Tunga Level 1

                                  Thanks Bill.