You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We talked about this a bit on Discord, but I thought I'd make an official report here. Load the attached scene and compare the motion blur in Arnold and Cycles. Arnold looks good, but Cycles creates "pointy" blur. It might be related to this issue: https://projects.blender.org/blender/blender/issues/103918 Thanks!
This does seem to be because we're not packing the radius in to the vec4 for the motion positions.
I think there is some opportunity to simplify the motion blur implementation in GafferCycles. Currently every primitive conversion function has its own duplicate logic conforming the samples to Cycles' constraints, and there is also additional logic in CyclesRenderer to determine if a sample coincides with the frame. I think this could probably all be consolidated into a single place.
We talked about this a bit on Discord, but I thought I'd make an official report here. Load the attached scene and compare the motion blur in Arnold and Cycles. Arnold looks good, but Cycles creates "pointy" blur. It might be related to this issue: https://projects.blender.org/blender/blender/issues/103918 Thanks!
cycles_particle_blur_bug.zip
The text was updated successfully, but these errors were encountered: