PDA

View Full Version : Car Paint Material Node flakes in reflection



hunter
07-18-2012, 08:28 AM
Is this something known about? I did a search and came up empty. Look at the flakes in the reflection on the chrome surface. Does this look right to you guys? Using the latest 11.0.2

MSherak
07-18-2012, 10:19 AM
Wireframe would help..

and the best way to check is place a camera where you think its wrong and render in the direction of the bounce.. Sometimes it helps to see what the surface sees..

-MSherak

hunter
07-18-2012, 10:40 AM
Here's the object and scene. I suppose I wasn't really clear in what is wrong but the flakes are just way too bright and apparent in the reflection compared to the paint itself. Looks especially out of place on a chrome bumper on a car using the node.

3DGFXStudios
07-18-2012, 02:22 PM
Maybe use more samples.
Play with Light samples ray recursion limit AA. etc

hunter
07-18-2012, 02:44 PM
Thanks for the ideas. Tried them out, but still the same. Just took longer to render. :) There seems to be a disconnect between the way the flakes render and how they render in reflections. Seems like a bug to me. Here's another with a simple mirror object. The ball and it's reflection should look pretty much the same no?

brunopeixoto
07-18-2012, 09:45 PM
Looks like it's something with the flakes...

Better results with these settings:

Sensei
07-18-2012, 10:47 PM
Ray is hitting reflective surface- mirror, then bounce to ball with carpaint node, and it's white dots- so they're used by reflective surface.
Carpaint when it's hit by direct ray from camera is blending them with surrounding spots, reflective surface has no chance for it- surrounding pixels are reflecting completely different part.
I tried using Reflection Blur on reflective mirror surface, and dots disappeared- they're blended.

jwiede
07-20-2012, 12:22 AM
Ray is hitting reflective surface- mirror, then bounce to ball with carpaint node, and it's white dots- so they're used by reflective surface.
Carpaint when it's hit by direct ray from camera is blending them with surrounding spots, reflective surface has no chance for it- surrounding pixels are reflecting completely different part.
I tried using Reflection Blur on reflective mirror surface, and dots disappeared- they're blended.
Still looks like a bug, IMO. If the carpaint shader is blending in its response to direct rays, then it should also be blending in response to reflection rays -- the situation w.r.t. "adjacent pixels" is really no different for mirror than camera.

Lewis
07-20-2012, 12:28 AM
I agree it's odd behavior, not sure is it a bug or some crazy limitation but one thing is sure - Real CarPaint don't do that (reflect what's not seen in view/color) so I'd report that with scene/samples.

Sensei
07-20-2012, 01:05 AM
See what happens when Flake Density is set to 100% - reflection is just flake color...

But I noticed that Flake Distance Falloff is insane 1 meter!

After setting it 100mm, or 10mm render is fine..

I guess so reflective ring is within 1m Falloff distance, that's why they're visible.

http://forums.newtek.com/attachment.php?attachmentid=105652&stc=1&d=1342768147

http://forums.newtek.com/attachment.php?attachmentid=105653&stc=1&d=1342768147

jwiede
07-20-2012, 10:17 AM
So if the camera is moved to within the flake falloff range of the sphere, does it show the same thing as is reflected when falloff is 1m?

Seems like carpaint shader needs to look at total path length to camera for reflected rays, not just that of the segment from reflector to carpaint-surfaced object.