OTOY Forums

Has anyone had to deal with this weird glitch when rendering specular surfaces in Octane? For some reason, a flat poly surface appears buckled, but this happens only for some camera angles. Move the camera slightly and the glitch fades away. This issue doesn't repro in other renderers I've tried, so I know it's not a geometry issue.
I'm using the Octane X beta for macOS, so it may be a Metal-specific bug.

If you can post a simple example scene that exhibits the problem, it will be easier to give you some useful feedback.

I exported an orbx file that demonstrates the issue. Let me know if you see that weird buckled reflection along the bottom surface of the tank. I don't know if this is happening across Octane in general or just the Octane X macOS beta.

It looks like you've got a small surface glitch in the actual polygons the side, not on the bottom, at least from what I can see. Switch over to Info Channels Kernel/Wireframe, and you'll see the glitch consistently aligns with a few faces on the front surface. I don't have direct access to the mesh in Octane so I can't go poke at individual faces and vertices, but as far as I can tell, the render is correct, but that big, subdivided face isn't perfectly flat and uniform, either in the actual geometry, or maybe in a face normal or vertex normal. When that surface irregularity lines up with the proper refraction boundary, you see it. Try simplifying that surface, or at least look very closely at it in your modeling app to find the guilty vertex and make it conform to its neighbors.

Edit:
Actually, I take that back. I just tried messing with it in C4D where I'm more fluent, and can't get that refraction glitch to show. I need to call it a day, but I"ll poke at it some more tomorrow.

Thanks for looking at my file and confirming that this is intended behavior. TBH I wasn't super motivated to track down this issue before because I wasn't sure if this was just a beta thing, given that 2 other renderers I used didn't exhibit the problem. And you were right, this was an issue with the geo on the side wall. Not sure why I was convinced before that it was the bottom surface. :-/

So the culprit turned out to be crazy v vectors on a few points that threw off the deformation motion blur. I guess other renderers have some logic to smooth out erratic variations in v whereas Octane takes the values more literally. Deleting the v vectors (the surface wasn't even supposed to be moving anyway) fixed the issue.

It's interesting that I wasn't seeing the problem when rendering with Octane in C4D. I wonder if Ahmet's C4D .orbx importer either fixes or completely ignores the vertex normals issue, while the Standalone just takes it in as-is. It makes a certain sort of sense that Standalone would do this, since the .orbx is native there. But it also produces possible disparity between the two renders, which could be a problem.

I don't have any experience with C4D but maybe you have to explicitly enable deformational motion blur in addition to the setting in the orbx file? If you have a way to look at velocity vectors per point in C4D, you'll see that they are swirling around in an unexpected way around the polys that are exhibiting the rendering glitch.