Announcement

Collapse
No announcement yet.

Do these bugs from 1.49.73 still exist in 1.5?

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

  • Do these bugs from 1.49.73 still exist in 1.5?

    hi,

    sitting here waiting for the dongles, i'd like to know in advance if those bugs present in 1.49.73 still persist to 1.5 release version.

    - motionblur with heavy motion drops some corner buckets. It does not depend on motion blur geometry samples or AA samples/thresholds.
    movie: http://www.sushidelic.com/tmp/mb_bug.mov

    - sss settings scatter, fwd/bck, multiplier don't have any effect, neither in hard nor soft model.


    - when using the Vray VFB to output render elements, openEXR files are saved as black files with correct alpha.

    I have sample files lying around here documeting this issues, but however i will only make them available on personal demand.

    So anyone back from siggraph already playing with the new build could probably test it out.... or if vlado got over his jetleg....

    thanks a lot,
    michael
    This signature is only a temporary solution

  • #2
    Re: Do these bugs from 1.49.73 still exist in 1.5?

    Originally posted by Sushidelic
    - motionblur with heavy motion drops some corner buckets. It does not depend on motion blur geometry samples or AA samples/thresholds.
    movie: http://www.sushidelic.com/tmp/mb_bug.mov
    This is not a bug. The reason for this to happen is that none of the initial samples that V-Ray takes actually hits the object, therefore V-Ray decides that there is no need for further sampling. To correct this problem, you need to increase the minimum samples per pixel (min. subdivs, if you are using QMC AA) or use more motion blur subdivs. The issue is the exact same one where V-Ray may miss thin horizontal or vertical lines with the adaptive subdivision AA. In that regard, the new build is not different from the older ones. The same issue can be observed in other renderers too (e.g. mental ray) with similar settings.

    - sss settings scatter, fwd/bck, multiplier don't have any effect, neither in hard nor soft model.
    Those seem to work fine here when I'm looking at them now, but then again, they have always worked, as far as I remember.

    - when using the Vray VFB to output render elements, openEXR files are saved as black files with correct alpha.
    This will work if you UN-check the "Use RealPixel RGB data" option for the openEXR writer. In that regard, the new build is not different from the older ones. Since in 3dsmax version 7 and below, floating-point images were not supported by 3dsmax internally, the RealPixel channel had to be used to obtain the high-dynamic range image. However, this is not necessary in 3dsmax 8 and newer versions, where this is handled in the proper way.

    Best regards,
    Vlado
    I only act like I know everything, Rogers.

    Comment


    • #3
      thanks for your reply vlado! you're absolutely correct for the motionblur, i just tried adjusting it by the noise threshold, but increasing the min subdivs helps, although the geometry sampling doesn't in all cases. as i set up the scene from scratch, things got better anyway, looks like there were some problems as the original scene before originated from a 1.47.03 version.

      still can't get sss to work properly here, dunno why, but if 1.5 does it correctly than that's obsolete anyway.

      my last concern is now the elements data... how ist the workflow to save unclamped elements layers directly into exr files? do i always have to output a .vrimg file and convert it?

      thanks a lot,
      michael
      This signature is only a temporary solution

      Comment


      • #4
        ok again... dunno what has happened... i tried it several times, just disabling max framebuffer, enabling vray's one, adding elements, checking save separate render channels and unchecking real pixel data in the exr saver... it was always outputting clamped elements. Now, after completely resetting max and with a new scene it is working! really some metaphysical stuff going on here, but anyway it it working.

        thanks again, michael
        This signature is only a temporary solution

        Comment


        • #5
          I'm having the same problem as Sushidelic with translucency in RC2. Scatter and fwd/bck coefficients seem to have no effect.


          Scene here http://ftp.tskgroup.co.uk/~tsk3d/SSS_Test.max

          Has anyone else tried using translucency in RC2?

          Thanks,

          Dan
          Dan Brew

          Comment

          Working...
          X