Author Topic: [solved] Realistic snow in Corona  (Read 37745 times)

2018-03-29, 11:22:06
Reply #30

3dboomerang

  • Active Users
  • **
  • Posts: 214
  • Head of 3D
    • View Profile
    • 3DFLOW
following up on the "distance" thing i saw early in this thread, talking about how snow can benefit from a simple bump map starting at 2meters distance and needs displacement for detail in close ups:

Is it possible that Corona will ever have a "distance-map" to use inside material shaders? I'm having alot of problems with displacement in corona when we're looking for example at streets or pavements (flat surfaces, horizontal, like snow, pavements, streets; ...)

Displacement looks good in the foreground, when using for example 1pix-displacement. The displacement does not look good at all when looking at the same material, in medium distance for example. Simple reason being pixels are being squized into eachother, and when displacement is calculating in 1pixel-detail, this wont look good, it gives jittered edges and small points being displaced upward, instead of for example the joint of a tile.

I can upload an example later on if needed, but I think it's clear?

If I'd be able to use a "distance-map" within my material shaders, I could use a blend of 2 materials: one with displacement, and one without. This saves me memory for calculation hopefuly + no more crappy artifacts for distance-displacement. This distance-map could be a golden map here for a mask within a blend material.

Right now I'd have to augment my displacement settings to 0.5 pixels or more, which explodes my memory usage to +64Gb EASILY.

This is a very frustrating factor atm. I cant use displacement on pavements and flat horizontal surfaces. Trying to use normal maps and bump maps, but they look bad without displacement on the foreground.

Solution atm: Render crop of foreground with displacement + render full without displacement and photoshop the displacement over the non-displacement version in photoshop.



2018-03-29, 14:45:46
Reply #31

TomG

  • Administrator
  • Active Users
  • *****
  • Posts: 5466
    • View Profile
I am not sure blending materials with and without displacement would work, as probably the one with displacement would need to be fully calculated anyway. You could use the existing CoronaDistance to blend between the textures / shaders going into the displacement though.
Tom Grimes | chaos-corona.com
Product Manager | contact us

2018-03-31, 09:48:17
Reply #32

3dboomerang

  • Active Users
  • **
  • Posts: 214
  • Head of 3D
    • View Profile
    • 3DFLOW
okay tyvm, I will use this method. Atleast that way I can tell to displace using 2 pixels (which is enough or 4K resolution, in the foreground) - so the RAM doesn't go through the roof -

Grts

2018-04-12, 14:03:05
Reply #33

maru

  • Corona Team
  • Active Users
  • ****
  • Posts: 12758
  • Marcin
    • View Profile
Generally it is recommended to use displacement for large scale details (such as the shape of bricks in a brick wall) and bump for fine details (such as cracks in a brick wall). Displacement should not be used for tiny details.
There are some displacement tips and tricks here: https://forum.corona-renderer.com/index.php?topic=9728.0
Marcin Miodek | chaos-corona.com
3D Support Team Lead - Corona | contact us

2019-11-18, 17:18:19
Reply #34

Place

  • Active Users
  • **
  • Posts: 43
    • View Profile
Hi,
I don't know what is going on here: I tried the snow scene in Max 2016 and Corona 5, just open the scene
and say switch to new Displacement 2.5 and new medium resolving... (but it's the same without doing this)
This is the result, at the same resolution, different render form IR and production... both wrong.

I'm I the only one with this wrong output?

Thank you,
Man

2019-11-18, 17:27:30
Reply #35

romullus

  • Global Moderator
  • Active Users
  • ****
  • Posts: 8848
  • Let's move this topic, shall we?
    • View Profile
    • My Models
Can you right click in the VFB on one of those black pixels an see if it tells NAN somwhere? Also try to turn bloom and glare and/or denoising and see if that helps.
I'm not Corona Team member. Everything i say, is my personal opinion only.
My Models | My Videos | My Pictures

2019-11-18, 17:57:49
Reply #36

Place

  • Active Users
  • **
  • Posts: 43
    • View Profile
No NAN appears; bloom and glare and denoising didn't help.
The same happens with Max 2018.
What is really strange is the different behavior between IR and production mode...

2019-11-18, 19:36:57
Reply #37

romullus

  • Global Moderator
  • Active Users
  • ****
  • Posts: 8848
  • Let's move this topic, shall we?
    • View Profile
    • My Models
Actually i wanted to say - try with B&G and denoising turned off, but i guess you already did that. As far as i remember, the original scene is using refraction and volumetric displacement. I'm not sure about latter, but refraction has seen major change since, so it could be that scene needs to be updated to render correctly with recent Corona. Do you have possibility to test the scene with older version of Corona?
I'm not Corona Team member. Everything i say, is my personal opinion only.
My Models | My Videos | My Pictures

2019-11-18, 22:55:14
Reply #38

PROH

  • Active Users
  • **
  • Posts: 1219
    • View Profile
Hi. Check out the displacementmap settings. Set it to "high dynamic range", and/or turn the level down (maybe somewhere around 8,5).

2019-11-19, 11:24:01
Reply #39

Ink Visual

  • Active Users
  • **
  • Posts: 169
    • View Profile
I quickly checked the scene, and originally having same issue as posted by Place, it seems to be resolved after checking high dynamic range as PROH advised.

2019-11-19, 14:52:16
Reply #40

maru

  • Corona Team
  • Active Users
  • ****
  • Posts: 12758
  • Marcin
    • View Profile
Those black spots are related to the new medium resolving method and it's logged in our tracker.
(Internal ID=419535796)
Marcin Miodek | chaos-corona.com
3D Support Team Lead - Corona | contact us

2019-11-19, 15:32:49
Reply #41

Place

  • Active Users
  • **
  • Posts: 43
    • View Profile
I confirm, resolved doing as suggested by PROH and confirmed by cexec
Thank you all!
Man