[Thomas An.]

Yes, but if people haven't suddenly changed their modelling habits, it means that these geometry glitches were there before, but the Maxwell alpha was more lenient. No ?

[Tyrone Marshall]

Yes, but if people haven't suddenly changed their modelling habits, it means that these geometry glitches were there before, but the Maxwell alpha was more lenient. No ?

In the case of the black dots that I got, it would be through the use of a hypernurb applied to object geometry.

I am sure it is a result of some logic expression in the re-writing of the render engine that will be soon corrected. Makes the most sense as we did not see it in the alpha version.

Good work Adam T!

[Tyrone Marshall]

Do you mean to tell me that the black dots are SPILLING out of vertices because they are touching each other? I think the problem is from when emitters and physical sky share the same SPACE. This could be from sunlight, or skylight, I don't know which. What does everyone else think?

I have found another wild source of noise from the glass texture. Has anyone else seen this too? Reversing the normal or flipping them will cause noise to be out of glass in seconds, otherwise, there is a lot.

This is not true as I got the dots of vacuum with just emitters only.

ref:

[rivoli]

in my case geometry wasn't the cause of black dots. different materials applied to the very same object could or could not give black dots.

[rivoli]

i will, my modelling is not the cleanest, so there might be imperfections (more than 4 edges polies, not perfect edge loops and so on), i will check that. even though i think i got black dots even with plain, unsmoothed, unchamfered 6 faces boxes.

[Tyrone Marshall]

Do you mean to tell me that the black dots are SPILLING out of vertices because they are touching each other? I think the problem is from when emitters and physical sky share the same SPACE.

No, I think the engine doesn't know to handle certain kinds of unorthodox geometry, and as a result something is getting screwed up in the calculations. Well, I don't just think it--I KNOW it.

There may well be other causes of the problem, but if you have a theory try to back it up with some kind of evidence.

I agree!

[Becco_UK]

Adam: I'm not quite convinced about the theory, probably because I cannot make out what you mean by four poly's joined at a single edge but the image text says three poly's. Is it three or four? Either way I still wouldn't grasp this because the geometry appears to converge to point rather than an edge, much like a sphere does at its poles.

Also, I rendered a scene yesterday and loads of big black dots appeared. In previous test renders of the same scene there was no problem. So I stopped the render, and restarted with all the same materials etc and everything rendered as it should.

Is your theory the same sort of thing and you think the geometry of the object was at fault. Or did the geometry you refer to keep producing black dots?

[sidenimjay]

i am wondering if using SSS is at the core of the problem?


does anyone get black dots when using glass or metals without diffuse or pastic materials in the same scene? or at least with the scatter and absorb values at 0?


i notice that the more diffuse/plastic materials with SSS the more dots i get . . .

[rivoli]

does anyone get black dots when using glass or metals without diffuse or pastic materials in the same scene? or at least with the scatter and absorb values at 0?

i do, see here:

http://www.maxwellrender.com/forum/viewtopic.php?t=3901

[rivoli]

i was just testing it adam, even with the most unhortodox geometry i couldn't get any dot.

can't really see a pattern here. does anybody?

[sidenimjay]

what are the default values for absorb and scatter in your package of choice?

is absorb and scatter greater than 0?


rivoli i see for your scene you have diffuse materials, do they have any absorb or scatter?

i have a scene that was rendering fine when i used a metal shader, but i changed to a diffuse with some absorb and scatter and when i rendered that out i have the wonderfully large black holes

will test more when there is time

[Becco_UK]

Adam: Thank you for the explanation. I tried to replicate your black dot theory using Cinema4D and its Maxwell Plugin.



I created a singe 100 metre poly', mirrored it and moved it 50 metres so the two edges shared the same space. Duplicated the mirror and rotated it 90 degrees. Made the mirrors editable and connected the result. Used Cinema's optimise. Checked to make sure the polygons were all connected.

Used a Maxwell Diffuse material on the object. Used 1 emmitter and a blue skydome for the lighting.

Rendered but got no black dots @ sample level 8 with this test.

[rivoli]

rivoli i see for your scene you have diffuse materials, do they have any absorb or scatter?

nope, absorption and scattering are both at 0.

[Becco_UK]

Using the same simple scene above I applied Absorbtion & SSScattering values of 0.3 to the plain diffuse material and done another render.

The scene rendered correctly with no black dots.

With that I'm off to bed - goodnight.

[n1tram]

I rendered a white plastic plane with a RGB bump map on it. On top of it a metal/gold torus knot with v roughness set to 1... Physical sky on + sunlight. No emitters.

And I get dots.

Ill start messing around wih the settings to see if I can find something in common

Edit: Removed sunlight and rendered again. The dots are gonne.

But I found something odd... I have one metal object (torus knot) + a plastic object (white plane). Under the render - Scene data tab, It says: 1 metalic object, 1 plastic object, 0 emitters, 0 dielectrics... 1 Diffuse object?!?!?

Is physical sky taken as a diffuse map object?