Announcement

Collapse
No announcement yet.

V-Ray Scene Optimisation Script

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

  • That would be cool, but it's not in any of the save file dialogs/paths. There is a different exr file in all of the dialogs I can find.

    Comment


    • Originally posted by Morne View Post
      OK if I bring GSM down to 1, then Total Lights Samples is the 1st thing that comes down to 512 (I think)
      So now only place I see 512 is here and it says 512|(1024) / 800|(1600)

      Does that mean it is now 512 SPP? Not sure I understand the min max etc. What is the 512|(1024) / 800|(1600) numbers in bold then?

      And somewhere in the thread you mentioned I should aim for less camera samples vs the rest. So if everything is 512, what should camera samples be then and what part of the info panels tells me that?
      Click image for larger version

Name:	not_for_the_colorBlind_ones.png
Views:	1
Size:	34.2 KB
ID:	854006

      the maths involved is here:
      http://forums.chaosgroup.com/showthr...505#post627505

      I turn off "divide shading subdivs", so each AA step MULTIPLIES the base subdivisions.
      In the case of the Archviz preset, the math is as follow (well, for how it shows up now. not user is identical to what i released, but the math is correct regardless):

      8 subdivisions = 64 SPP.
      (64 SPP * (4*4) (min AA squared) ) * 1.334 GSM * 0.75 (as adaptivity is 0.25, so 75% of the rays are cast regardless.) = 1024 SPP (and change).
      Should the adaptive routines think more rays are needed, at the min AA level of 4 you'd get 1366 SPP.

      Same applies for the Max AA, only in the forumla above 4*4 becomes 6*6, which leads to 2304 to 3073 SPP.
      while this quote comes form here, the last of a set of three posts which may be helpful in understanding the reasons for the suggestion, along with the terminology (or lingo, as some call it) used:
      http://forums.chaosgroup.com/showthr...034#post628034

      So here comes the, by now obvious, I hope, conclusion: in any type of scene setup you will ever face, if speed is a concern, you will want to use the camera rays (AA rays, if you please) as sparingly as you can possibly afford.
      My overarching suggestion (and the script aids in PRECISELY this department) is this: FIRST set up the AA so that your geometric edges and high frequency textures are resolved cleanly (if there is motion blur or depth of field at play, it makes no conceptual difference, you will just have to use more camera rays to get those effects clean) with the LEAST amount of camera rays possible, THEN increase the specialised rays so that your shading effects look clean under those anti-aliasing conditions.
      Lele
      Trouble Stirrer in RnD @ Chaos
      ----------------------
      emanuele.lecchi@chaos.com

      Disclaimer:
      The views and opinions expressed here are my own and do not represent those of Chaos Group, unless otherwise stated.

      Comment


      • Originally posted by AlexP View Post
        That would be cool, but it's not in any of the save file dialogs/paths. There is a different exr file in all of the dialogs I can find.
        Umh, the max standard save dialog, maybe?
        Lele
        Trouble Stirrer in RnD @ Chaos
        ----------------------
        emanuele.lecchi@chaos.com

        Disclaimer:
        The views and opinions expressed here are my own and do not represent those of Chaos Group, unless otherwise stated.

        Comment


        • Originally posted by ^Lele^ View Post
          Umh, the max standard save dialog, maybe?
          Nope.
          Im using a script to submit renders to backburner, but afaics it just renames the file output while sending. Which at the moment is set to a different file altogether.

          Comment


          • If I do NOT switch off "divide shading subdvis", could I then get similar results by simply increasing GSM?
            If so, what would be the math for that?

            For example can I make my AA 2,2 (Resulting in 4 camera samples) and GSM 4 to get rest of the samples back up to 512?

            EDIT: Nevermind, that logic doesnt work hehe. Render time shot up big time
            Last edited by Morne; 19-08-2014, 07:04 AM.
            Kind Regards,
            Morne

            Comment


            • Originally posted by AlexP View Post
              Nope.
              Im using a script to submit renders to backburner, but afaics it just renames the file output while sending. Which at the moment is set to a different file altogether.
              try/catch it is, then.
              Can't fathom where the stuff's coming from, so i see no other way out.
              But boy isn't try/catch THE max scripter "back off" switch :P
              Lele
              Trouble Stirrer in RnD @ Chaos
              ----------------------
              emanuele.lecchi@chaos.com

              Disclaimer:
              The views and opinions expressed here are my own and do not represent those of Chaos Group, unless otherwise stated.

              Comment


              • Originally posted by ^Lele^ View Post
                Very nice!
                Care to test render the scene I linked on google drive a few posts earlier? it's a wee bit cleaner than the GPU one, which looks more in the ballpark of the 22mins render i did last, but i'd be curious to see what's the difference.
                Speaking of which, what's the hardware you're rendering with?

                Here's the RT test of the last scene file you uploaded...It seems that beyond 8 min It doesn't clean up much more...Vlado has LC coming to RT! it should do a much better job with cleaning up the extra noise.


                Click image for larger version

Name:	RT GPU Test 2.jpg
Views:	1
Size:	168.0 KB
ID:	854007
                "I have not failed. I've just found 10,000 ways that won't work."
                Thomas A. Edison

                Comment


                • Script updated to 1.50 (in the first post).
                  Let me know if it sorts out the oddity, Alex!

                  Ah, Eyepiz, i meant for you ro tender it with the offline (ie, standard, CPU) renderer...
                  Lele
                  Trouble Stirrer in RnD @ Chaos
                  ----------------------
                  emanuele.lecchi@chaos.com

                  Disclaimer:
                  The views and opinions expressed here are my own and do not represent those of Chaos Group, unless otherwise stated.

                  Comment


                  • Seems to have worked, thanks

                    Comment


                    • Originally posted by Morne View Post
                      If I do NOT switch off "divide shading subdvis", could I then get similar results by simply increasing GSM?
                      If so, what would be the math for that?

                      For example can I make my AA 2,2 (Resulting in 4 camera samples) and GSM 4 to get rest of the samples back up to 512?
                      The script isn't set up for "divide on", but the math is very similar.
                      the problem with it is that for high AA rays, your subdivs would have to be humongous.
                      Say we look at the example you make:
                      with MAX AA of 2 (min isn't considered at all in the subdivs division), you would need the GSM to be 4 to keep whatever subdivision level you put in your lights and shaders.
                      With a max AA of 16, GSM would have to be 256.
                      With a max AA of 64, GSM would have to be 4096.
                      And that still leaves entirely free what shading gets done at which AA level before the MAX is reached (if it ever gets there).
                      Meaning, in other words, that if you had a min AA of 1, and a max of 64, with a gsm of 4096, you'd likely get very low subdivs per AA ray, so your AA would step up, and use the slower camera rays.
                      So, you'd have to compensate with the "min shading rate", which in turn would not reach the max AA rays, which in turn would make the GSM multiplier used the wrong one to maintain the sampling you wanted, starting a loop for which the actual sampling done is incredibly hard to fathom before the render starts (or after it's finished, for that matter.)
                      I didn't ask for the "divide" option for 6 years for nothing
                      Lele
                      Trouble Stirrer in RnD @ Chaos
                      ----------------------
                      emanuele.lecchi@chaos.com

                      Disclaimer:
                      The views and opinions expressed here are my own and do not represent those of Chaos Group, unless otherwise stated.

                      Comment


                      • Originally posted by AlexP View Post
                        Seems to have worked, thanks
                        Can't say i am proud of the fix, but i am surely glad it sorted it out.
                        Lele
                        Trouble Stirrer in RnD @ Chaos
                        ----------------------
                        emanuele.lecchi@chaos.com

                        Disclaimer:
                        The views and opinions expressed here are my own and do not represent those of Chaos Group, unless otherwise stated.

                        Comment


                        • Originally posted by eyepiz View Post
                          Here's the RT test of the last scene file you uploaded...It seems that beyond 8 min It doesn't clean up much more...Vlado has LC coming to RT! it should do a much better job with cleaning up the extra noise.


                          [ATTACH=CONFIG]20559[/ATTACH]
                          Speaking of which, with the hardware you're sporting you may as well render the very clean (ie. 1.5 hrs on my amd) scene: i have a feeling it would render in the same amount of time as the GPU one, for substantially better results.
                          Lele
                          Trouble Stirrer in RnD @ Chaos
                          ----------------------
                          emanuele.lecchi@chaos.com

                          Disclaimer:
                          The views and opinions expressed here are my own and do not represent those of Chaos Group, unless otherwise stated.

                          Comment


                          • Well, I cant say I understand your script, at least from the math point of view, but having now played a lot with it, and experiment myself without the scripts help, I have got my BF/LC 6k still renders out in just under half the time, 2 hours down to 1:05. Very pleased! Fair bit of planting and large building, CG in about 60% of images, rest BG photo.

                            I'm not using all your suggestions such as untick the magic divide shading subdivs, but I have a workflow where the 1h (cleaner then before as well!) render can be a 6 min full size draft with just two number changes.

                            Time will tell if my "standardized" settings will work out, but they are better then before and I have a good understanding of the reasons behind the speed up, so many thanks. Now holiday time for 11 days as been solid for the last 4 weeks!

                            M

                            Comment


                            • Lele's a witch! Burn the witch

                              Comment


                              • Glad the concepts helped, and ever more glad you got to a deeper understanding of V-Ray's inner workings, Mike!
                                Burning me is going to be easy John: start from the hair! XD
                                Lele
                                Trouble Stirrer in RnD @ Chaos
                                ----------------------
                                emanuele.lecchi@chaos.com

                                Disclaimer:
                                The views and opinions expressed here are my own and do not represent those of Chaos Group, unless otherwise stated.

                                Comment

                                Working...
                                X