Help with ambient occlusion in solaris

   1984   6   5
User Avatar
Member
361 posts
Joined: Nov. 2015
Offline
Hey guys, anybody had any luck using AO maps in solaris using either VEX or MTLx? I haven't been able to get either the VEX nor the materialX occlusion nodes to work. Any help is much appreciated. Thanks.
hou.f*ckatdskmaya().forever()
User Avatar
Member
1766 posts
Joined: May 2006
Online
Yes with CPU, no with XPU.

I've attached both vex and matx setups that work with CPU. I understand and expect the vex one to not work with XPU, but I don't understand why the matx one fails with XPU.

Attachments:
lops_ao.hip (142.7 KB)

http://www.tokeru.com/cgwiki [www.tokeru.com]
https://www.patreon.com/mattestela [www.patreon.com]
User Avatar
Member
361 posts
Joined: Nov. 2015
Offline
mestela
Yes with CPU, no with XPU.

I've attached both vex and matx setups that work with CPU. I understand and expect the vex one to not work with XPU, but I don't understand why the matx one fails with XPU.
Thanks a lot Matt. Hopefully it gets implemented into XPU soon!
hou.f*ckatdskmaya().forever()
User Avatar
Member
117 posts
Joined: Nov. 2017
Offline
No, it hasn't been implemented yet. Luckily I came across this post. Would have never suspected that as the solution.
User Avatar
Member
149 posts
Joined: Aug. 2015
Offline
hope it will be supported in 20.5...
User Avatar
Member
53 posts
Joined: April 2013
Offline
The keynote says it will be supported in 20.5

Attachments:
AO.JPG (35.7 KB)

User Avatar
Member
117 posts
Joined: Nov. 2017
Offline
Anybody else prefering the ao from RS?

I have to set the max ao distance on the MtrlX node with a fit expression taking the frame into account. E.g. on frame 1 > distance 1m, frame 2 > distance 2m, frame 3 > distance 4m, frame 4 > distance 8m and so on. In Nuke I layer these renderings. Otherwise the falloff look really off. No chance to emphasize the areas close to cavities otherwise. This works for stills, but for anims not that much. Usually I'd use a pow expression to achieve this effect in post, but results don't look nice, since the falloff of the ao is kind of odd to what I'm used to.
  • Quick Links