Page 2 of 2 FirstFirst 12
Results 16 to 24 of 24

Thread: Random SSS Node

  1. #16
    Not so newbie member lardbros's Avatar
    Join Date
    Apr 2003
    Location
    England
    Posts
    6,045
    Have you tried baking the irradiance cache?

    Irradiance caching takes a different sample each frame, so it will be different each time. Baking the irradiance cache should stop it, unless this is a bug with the SSS?

    - - - Updated - - -

    The difference in your two frames is the noise though, where LW's renderer (as most renderers do) takes a Russian roulette approach to firing samples... you have to increase the samples in order to remove the flickering from each frame.
    Maybe I'm misunderstanding the issue, but brute force, reflections, shadows all have this Russian roulette sampling... it is always different per frame.
    Last edited by lardbros; 09-24-2020 at 05:40 AM.
    LairdSquared | 3D Design & Animation

    Desk Work:
    HP Z840, Dual Xeon E5-2690 v2, 32GB RAM, Quadro K5000 4GB
    Desk Home:
    HP Z620, Dual Xeon E5-2680, 80GB RAM, Geforce 1050 Ti 4GB

  2. #17
    Registered User
    Join Date
    Aug 2017
    Location
    Slo
    Posts
    151
    Quote Originally Posted by lardbros View Post
    Have you tried baking the irradiance cache?

    Irradiance caching takes a different sample each frame, so it will be different each time. Baking the irradiance cache should stop it, unless this is a bug with the SSS?

    - - - Updated - - -

    The difference in your two frames is the noise though, where LW's renderer (as most renderers do) takes a Russian roulette approach to firing samples... you have to increase the samples in order to remove the flickering from each frame.
    Maybe I'm misunderstanding the issue, but brute force, reflections, shadows all have this Russian roulette sampling... it is always different per frame.
    I will comment on Irradiance caching later, but yes I have baked frame and there is an issue at least for me. It seems that "Clear cache" button does not work as expected.

    If LW have a Russian roulette approach, why there is NO difference if you render twice a single frame?

    My last renders that I showed are all WITHOUT GI as well as Johns.

  3. #18
    Not so newbie member lardbros's Avatar
    Join Date
    Apr 2003
    Location
    England
    Posts
    6,045
    Quote Originally Posted by next_n00b View Post
    I will comment on Irradiance caching later, but yes I have baked frame and there is an issue at least for me. It seems that "Clear cache" button does not work as expected.

    If LW have a Russian roulette approach, why there is NO difference if you render twice a single frame?

    My last renders that I showed are all WITHOUT GI as well as Johns.
    Yeah, there does seem to be a problem with clearing the cache, but not entirely sure what it is.

    I tend to go an manually delete them. If you overwrite them it will cause major issues if you move things, or change the lighting.

    Russian roulette sampling generates a random seed per frame, so if you render the same frame twice, it looks exactly the same. This is what most render engines do.
    LairdSquared | 3D Design & Animation

    Desk Work:
    HP Z840, Dual Xeon E5-2690 v2, 32GB RAM, Quadro K5000 4GB
    Desk Home:
    HP Z620, Dual Xeon E5-2680, 80GB RAM, Geforce 1050 Ti 4GB

  4. #19
    Registered User
    Join Date
    Aug 2017
    Location
    Slo
    Posts
    151
    Quote Originally Posted by lardbros View Post
    Yeah, there does seem to be a problem with clearing the cache, but not entirely sure what it is.

    I tend to go an manually delete them. If you overwrite them it will cause major issues if you move things, or change the lighting.
    Exactly that! Even if you change samples, it seems that the baking procedure expand the cache file in an inappropriate way.

    Quote Originally Posted by lardbros View Post
    Russian roulette sampling generates a random seed per frame, so if you render the same frame twice, it looks exactly the same. This is what most render engines do.
    If this is the case, it is new to me, and I wonder WHY? Why the engine would work that way?

  5. #20
    Not so newbie member lardbros's Avatar
    Join Date
    Apr 2003
    Location
    England
    Posts
    6,045
    Quote Originally Posted by next_n00b View Post
    Exactly that! Even if you change samples, it seems that the baking procedure expand the cache file in an inappropriate way.



    If this is the case, it is new to me, and I wonder WHY? Why the engine would work that way?
    It's this way because it removes the weird effect of static noise.
    I've tested a static noise renderer before (this could be done in LW's renderer as an option maybe) and when you move through an environment, or objects move around it looks like they are moving through a cloud of static dots. It generally looks really strange.
    The random noise method is used in most renderers these days as it enables them (like in Arnold) to ditch 50% of the samples through glass or do clever things with the samples, that don't cause horrendous artefacting.
    LairdSquared | 3D Design & Animation

    Desk Work:
    HP Z840, Dual Xeon E5-2690 v2, 32GB RAM, Quadro K5000 4GB
    Desk Home:
    HP Z620, Dual Xeon E5-2680, 80GB RAM, Geforce 1050 Ti 4GB

  6. #21
    Registered User
    Join Date
    Aug 2017
    Location
    Slo
    Posts
    151
    Yes, I think I answered to myself in the meantime, that must have something to do with camera movements and blending frames together to deceive the viewer.

    But what about Sensei point of view, he said, if frames differ, it is a bug. It seem to me as a knowledge man.

  7. #22
    Not so newbie member lardbros's Avatar
    Join Date
    Apr 2003
    Location
    England
    Posts
    6,045
    Quote Originally Posted by next_n00b View Post
    Yes, I think I answered to myself in the meantime, that must have something to do with camera movements and blending frames together to deceive the viewer.

    But what about Sensei point of view, he said, if frames differ, it is a bug. It seem to me as a knowledge man.
    Yeah, not sure what Sensei means... he's a very clever guy, so I value his views.
    LairdSquared | 3D Design & Animation

    Desk Work:
    HP Z840, Dual Xeon E5-2690 v2, 32GB RAM, Quadro K5000 4GB
    Desk Home:
    HP Z620, Dual Xeon E5-2680, 80GB RAM, Geforce 1050 Ti 4GB

  8. #23
    Electron wrangler jwiede's Avatar
    Join Date
    Aug 2007
    Location
    San Jose, CA
    Posts
    7,113
    Quote Originally Posted by next_n00b View Post
    Now I am lost. I have tried the same thing in 2018 version and the behavior is the same. Renders are different from frame to frame with Sigma node or Principled BSDF with Subsurface enabled. What part I don't understand I ask myself?
    The various "translucency" shaders have been a mess* since LW2018. That an "SSS rewrite" occurred with LW2020, yet the shader(s) are still fundamentally broken in similar manners, highlights the problem.


    *: Unstable shader results, broken with basic light types, etc. Bugs filed in LW2019 by myself and others.
    Last edited by jwiede; 09-25-2020 at 02:22 PM.
    John W.
    LW2015.3UB/2019.1.5 on MacPro(12C/24T/10.13.6),64GB RAM, NV 980ti

  9. #24
    the other part of of SSS is Caustics, did any of you up the max recursion limit and samples for Caustics or even up the light samples for your back light to get a smoother look, there's a lot of noise in theses images, besides I always use BSDF, never bother with standalone SS Shader

Page 2 of 2 FirstFirst 12

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •