Re: A pain in the Arnold…

2018-06-01 Thread Matt Lind
I distinctly remember looking at cityscape scenes in XSI a number of years ago when another user had similar problems. If memory serves, the geometry is horribly corrupt and each material was duplicated per polygon or polygon cluster. There were winged edges with more than 3 polygons sharing

RE: A pain in the Arnold...

2018-06-01 Thread Ponthieux, Joseph G. (LARC-E1A)[LITES II]
Stephen, I’m fairly sure it has nothing to do with UVs. But I’ll try to get the log. What I’ve discovered is that if I extract pieces of the mesh away from the outer perimeter, paring it down till it reaches about 250x250 units and no bigger, the problem disappears. Breaking it up into

Friday Flashback #340

2018-06-01 Thread Stephen Blair
1999 SOFTIMAGE|3D screenshot…with caustics!

Re: A pain in the Arnold...

2018-06-01 Thread Stephen Blair
Hi You don't mention whether you checked the Arnold log for warnings. If it was tessellation errors, then you would see it with or without a texture. You can try using the Utility shader (Primitive ID with polywire) to check that. The Utility shader also has UV modes for debugging. On Fri,

RE: A pain in the Arnold...

2018-06-01 Thread Ponthieux, Joseph G. (LARC-E1A)[LITES II]
Just one shader. Its using a custom UV set for a single texture that I don’t dare touch since the UVs are generated automatically by CityEngine to repeat sidewalk segments along the sidewalk artery. Since other renders have shown they can handle them, I don’t think it’s the UVs. There are

Re: A pain in the Arnold...

2018-06-01 Thread Steven Caron
How are the different shaders applied to this single object? Is it using face sets? There is a cap of 256 shaders per object if that is the case. Although, what you initally described doesn't sounds like this is the case... Talk to the support guys, it's Stephen Blair BTW... He will help get you

RE: A pain in the Arnold...

2018-06-01 Thread Ponthieux, Joseph G. (LARC-E1A)[LITES II]
I have found one workaround so far that seems to force the render to behave. The geometry is procedurally created in CityEngine. The offending geometry in the scene is the “sidewalk” object. But the way it is generated is that the entire sidewalk is like an arterial system created as one object

RE: A pain in the Arnold...

2018-06-01 Thread Ponthieux, Joseph G. (LARC-E1A)[LITES II]
Nope. First thing I checked. CV counts at each vertex are 1. There is no non-manifold geometry that I can find related to the objects with the anomaly. Besides, if it were the geometry, I would expect that removing the texture would not remove the anomaly. If there is no texture, the bug is not

Re: A pain in the Arnold...

2018-06-01 Thread Jordi Bares
I would suggest you double check the geometry, is it possible you have double triangles? jb > On 1 Jun 2018, at 14:32, Ponthieux, Joseph G. (LARC-E1A)[LITES II] > wrote: > > The original shader was a Lambert material with an image and a texture > placement. > > Replacing the shader

RE: A pain in the Arnold...

2018-06-01 Thread Ponthieux, Joseph G. (LARC-E1A)[LITES II]
The original shader was a Lambert material with an image and a texture placement. Replacing the shader elements with like for like makes no difference. Replacing the file and Lambert nodes with aiImage and aiStandardSurface also made no difference. I also replaced the SG and deleted the

RE: A pain in the Arnold...

2018-06-01 Thread Ponthieux, Joseph G. (LARC-E1A)[LITES II]
No. In part because the scene has over 900 shaders. The thought of having to edit every one to make Arnold happy is not a pleasant one. I’ll look into though. Stay tuned. Thanks Joey From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of