Tech & stress tests

Forums Gallery Tech & stress tests

This topic contains 21 replies, has 5 voices, and was last updated by Tim Ellis 5 months, 3 weeks ago.

Viewing 7 posts - 16 through 22 (of 22 total)
  • Author
    Posts
  • #2040
    Tim Ellis
    Tim Ellis

    Single wavelength cornell test:-

    Apollo Level 85 after 23hrs 36mins Bench 2132 (Bench updated at render finish!)
    Raw output:-

    Boosted levels in Photoshop:-

    Denoised version:-

    Boosted levels in Photoshop:-

    Not 100% sure about the Denoised versions, as I added this as an after thought after 13hrs of rendering.

    Tim.

    #2042
    Jeremy Hill
    Jeremy Hill

    Something seems off, but on the other hand I did render mine to level 98. However at your level it was definitely an image, just noisy. Let me go through my stuff and find my bsx.

    #2043
    Jeremy Hill
    Jeremy Hill

    Here is a bsz: https://bellarender.com/dl/TST-SingleWavelengthCornell.bsz

    The main trick with this was, if you don’t get the IOR & abbe exactly right, then the dispersion won’t hit the target right, so you may see nothing.

    #2044
    Jeremy Hill
    Jeremy Hill

    For others following along, here’s my level 98 result:

    TST-SingleWavelengthCornell, level98, iso400

    (and for anyone who wonders what we’re on about, posting grainy images, see here the description for “Single Wavelength Indirect Cornell Box”)

    #2056
    Tim Ellis
    Tim Ellis

    I noticed differences in using a black collimator enclosure verses a white enclosure. You get much better results with a white lambert than black Lambert material, which is to be expected, however it is black in Thomas’ original scene.

    I also had focus issues in my test and possible other material problems. That was version 0.0.16 as well so need to check your bsz with version 19.0.

    I’ve left the TIR validator scene running on two rendernodes, so will combine those in the morning and post the render.

    Tim.

    #2105
    Avatar
    Thomas An.

    I noticed differences in using a black collimator enclosure verses a white enclosure. You get much better results with a white lambert than black Lambert material

    If you use a white collimator, the white sidewalls at the very tip will spread a light-glow sideways, defeating the parallel-ray purpose of the collimator. So, once you go white, it would be similar to deleting the collimator housing and leaving the thin rectangular emitter on its own (which is mush easier on the engine but not so good for isolating dispersion effects).

    #2107
    Tim Ellis
    Tim Ellis

    Thanks for the feedback Thomas. Yes I found the same with Mxs too, I did revert it back to black, but definitely need to double check everything in the scene is working correctly before running another render.

    Without the complex IOR file for the second dielectric and an internal ior limit of 20 for the perfect mirror, I dont think we can yet exactly match your intended setup in the original scene. Still promising results so far.

    Tim.

Viewing 7 posts - 16 through 22 (of 22 total)

You must be logged in to reply to this topic.