[m-Que]

So, here's the deal:
I have a fillet (cube) with some genetic_frosty_material applied to it and the same fillet (cube) resized object (about 0.8x of the original size) with an emitter_material inside.
And here's what I get


The thing is, that the emitter_object that is not placed inside the genetic_frosty fillet works just fine.

And BTW, the object looks like as it has been triangulated. But in fact it's not.

[JorisMX]

did you check the normals?

[m-Que]

did you check the normals?

Yep, nothing's wrong with normals.
It only appeared in "fillet_in_fillet" so far. The other possibilities, for example "cube_in_fillet", seems to be rendered normally.

[JDHill]

Since the material is a dielectric, I think you should model the outer cube with an inner cavity, then size the emitter so it fits completely inside (no co-planar faces). I would also try to simplify the geometry of the emitter to avoid long render times - a simple cube would be best, but if necessary, you could add some chamfer.

[msantana]

I don't know if you are using Maya, but if you are, check if this happens if the object is at 1,1,1 scale.

There was a bug (I don't know if it is fixed yet) if you scale an object, the normals will not display correctly in Maxwell, check this thread:

http://maxwellrender.com/forum/viewtopic.php?t=27030

[tom]

I can't make it fail here for some reason so, I'd be curious to see your MXS.

[m-Que]

Hello everyone,
Sorry that I'm getting back to this thread after a big break, but I've been really busy recently, so I couldn't have a closer look on that problem...

Now, back to business:
I did some test recently, and more details popped out:

-Tried what JDHill recommended:

Since the material is a dielectric, I think you should model the outer cube with an inner cavity, then size the emitter so it fits completely inside (no co-planar faces).

It gave me exactly the same result as on the pic from the first post.

-Since JDHill mentioned dielectric material, I tried applying basic dielectric material (from wizard) and surprisingly the image came out clean. So I started playing around with it's settings and came to a point, that it's actually roughness value that effects those black spots:
The higher the roughness value is, the more visible they are.
Here's an example: Same geometry cloned 4 times - emitter inside, dielectric with different roughness value on the outside.
Roughness values: 0, 20, 50, 99.

So, to make it clear: the strange-looking effect seems to appear when using some certain inner geometry with emitter material (I'm saying 'some certain', because it appeared only on some spots, not the whole geometry) together with ANY outer geometry with dielectric material with roughness value =>0.

BTW, I also used 'untriangulate' option to make sure non of the objects is triangulated.

And last night I came along to a VERY strange thing:
It turned out that the further the camera is to geometry, the the more visible the effect is.
Check out the picture: Again same geometry cloned 3 times, same size, same material settings. The only thing that differs is the distance from the camera.



PS: I've PM-ed you the MXS of the scene, tom

[tom]

I confirm this bug.