Announcement

Collapse
No announcement yet.

Vray GI and Color transmittance tests [maxwell] involved

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Vray GI and Color transmittance tests [maxwell] involved

    Hey every one. I have desided to do some testing, on recently new topic of vray and maxwell comparisons and etc. I want to say from the start that I am not making this post to bash vray or maxwell, as a professional I respect all renders and their developers. The reason for this test is to study multiple aspects of vray such as gamma correction, color transmittance and gi transmittance.

    I've started by doing the most simple tests. I placed 3 spheres into a scene with white gi environment. The shaders are vray mat's and maxwell diffuse mat's the colors are identical in both maxwell and vray. In this first test my goal is to see the amount of color which is being sent out by vray's default gi and how it would behave after gamma correction.



    In the first image I have corrected maxwells gamma to 1.0 which is what vray is by default. We can see that maxwells image is much darker but the colors are very close between vray and mx.
    Next set of images I have reverted the gamma in maxwell to 2.2 default and in vray's color mapping using egz gamma calculator I have input 0.45 value for bright multiplier. We see that while in vray everything is lifted shadows, brights but the colors are close to maxwell. However in maxwell the shadows are preserved and the colors are not as washed out. In the close up we see color bleed from spheres to the plane. This is much accentuated when the gamma correction is not applied as you can see on image 1 of vray. However in maxwell no color bleed happens.
    Both sets are rendered with ppt.

    Second test is a more complicated one. I have taken a middle point, so one cannot say which is more correct, vray maxwell and so on. I have taken a photograph with a specific lighting condition and have replicated it in vray and maxwell. My objective is not to replicate the photograph 100
    % but to examine the expected behavior from a photo and how vray and maxwell will perform.

    Photo


    Maxwell


    Vray


    Scene for testing. Need max 7 or greater.
    http://www3.telus.net/public/vinnik8...well_tests.rar

    By default vray has rendered a much darker scene, with settings of qmc 100/100
    When lifted gamma to 2.2 we notice a great deal of information being lost. Shadows become bright along with highlites. I also noticed much greater color transmittance from the sphere to the floor. While there is none on the photo or maxwell. What I have noticed in maxwell is that while illumination wise it had done a better job, the shadows are also bright. Which is vrays situation works better.
    More to come...
    Dmitry Vinnik
    Silhouette Images Inc.
    ShowReel:
    https://www.youtube.com/watch?v=qxSJlvSwAhA
    https://www.linkedin.com/in/dmitry-v...-identity-name

  • #2
    Nice studies here bud. I'll try and contribute once I get my current project done.

    Comment


    • #3
      Nice,

      I wonder if you have adjusted your materials to work with in the gamma corrected environment. If the answer is no, then I understand why things look washed out. Maybe maxwell corrects for this automatically...
      You can contact StudioGijs for 3D visualization and 3D modeling related services and on-site training.

      Comment


      • #4
        How would you adjust mats to work with gamma?
        Dmitry Vinnik
        Silhouette Images Inc.
        ShowReel:
        https://www.youtube.com/watch?v=qxSJlvSwAhA
        https://www.linkedin.com/in/dmitry-v...-identity-name

        Comment


        • #5
          Mind sharing the ball file? I am sure it will be no problem to adjust V-Ray to look more like the photo than the Maxwellrender.
          Eric Boer
          Dev

          Comment


          • #6
            To be honest, I dont think the focus here is to see how close we could get the renders to the images but instead to see how they individually handle aspects like color bleed.

            Comment


            • #7
              exectly. Rerender i have placed a link to the scenes in my first post.
              Dmitry Vinnik
              Silhouette Images Inc.
              ShowReel:
              https://www.youtube.com/watch?v=qxSJlvSwAhA
              https://www.linkedin.com/in/dmitry-v...-identity-name

              Comment


              • #8
                Nice tests Morbid.
                LunarStudio Architectural Renderings
                HDRSource HDR & sIBL Libraries
                Lunarlog - LunarStudio and HDRSource Blog

                Comment


                • #9
                  Yes, but that is my point. V-Ray is a biased renderer at heart. With the
                  right tweaks and artists eye it can produce exactly the results one
                  wishes in a fraction of time required by a non-biased renderer. So
                  whether or not the results are "exact" out of the box is inconsequential.
                  Eric Boer
                  Dev

                  Comment


                  • #10
                    nice test morbid.
                    i think this topic will keep going for a while.........
                    Chris Jackson
                    Shiftmedia
                    www.shiftmedia.sydney

                    Comment


                    • #11
                      I'm not sure I understand a direct comparison either. These are two different render engines. Of course they will be different. I agree with rerender that tweeking by the artist is whats important, not how it looks at default settings. But, it's interesting to see what maxwell can do.

                      I'm curious how do the render times compare? To me, that is more important than default setting images. But hey, I'm not looking to switch either, so this could effect my comparisons of the two.

                      Perhaps there may be advantages in a studio setting, I don't know, I work alone. Speed is everything to me since it's just me trying to do it all by yesterday.

                      Tony

                      Comment


                      • #12
                        interesting tests.

                        I await more

                        Comment


                        • #13
                          Very interresting test Morbid!
                          =:-/
                          Laurent

                          Comment


                          • #14
                            Originally posted by Morbid Angel
                            How would you adjust mats to work with gamma?
                            In general you can say that if you want to have a midgray as output, you have to input a rgb 47 47 47. For a new scene where you have enabled gamma correction in display, your materials look much lighter than they would without gamma correction. This means youo have to correct for it. Compare with textures, where you also have to revert from gamma correction in order to get correct results.
                            You can contact StudioGijs for 3D visualization and 3D modeling related services and on-site training.

                            Comment


                            • #15
                              Forgot to mention, when you turn on display gamma correction while whatching the material editor, you will also notice that colors get desaturated. So saturation level should be increased, lightiness should be reduced.

                              Here is the rendering after 'correcting' the materials:

                              lightcache for first and secondary bounces 2000/0.01
                              QMC: 0.002 global subd. 8


                              You can contact StudioGijs for 3D visualization and 3D modeling related services and on-site training.

                              Comment

                              Working...