- Fri Jan 03, 2014 11:09 pm #376074
First, I would say that for industrial design, I would use nothing else. The whole idea to work with a "real" camera and to use "real" (IES) lights as emitters is ever so easy to understand. But then, as you say, the problems start for the regular user who does not use MR daily. Where do these oddly shaded triangles come from, that look like yours http://bayimg.com/JAFofaaFm in a scene with 3 emitters and the default material? Where does the glare come from here http://bayimg.com/naFiCaafL although the single emitter in the scene is hidden from view? What is inducing all that noise after 3 hours with three nodes? As you point out, it is not always intuitive to know what to look for, even if one has the entire MR documentation and knowledgebase printed and reads it constantly. If some of that counterintuitivity could be reduced by way of better example, there would be no complaints : )Polyxo wrote:If you create an SSS material, apply it to a cube and start rendering... does the material look clear right away on youreric nixon wrote:What's wrong here? - your settings/geometry problems, hence the noise, so why let it render? The noise is saying , hey theres something wrong.
machine? Were you able to judge the look of a tiny surface structure? I can say that this isn't possible where I sit.
My point wasn't complaining - but I would agree with Martin that predictions on certain subtle material properties
are really hard to make currently. That is not to say that persons who do little else professionally than rendering might
be able do develop certain instincts but one certainly can not see what's wrong :).
The mesh in question btw is not greatly more complex than a cube, it's all quads and closed and nice, the sss material is
a basic preset one - the only thing I did that I added the nor_foam.jpg from the default material library at a
high tiling rate. Actually I still don't know what went wrong... here's a downsized version of the raw render and I swear
that there was no chance to see this triangulation artifact at Fire res/low sampling levels.
My personal guess is that MW treated that Normal map with the same criteria it uses on Bump maps where high intensity
multipliers lead to shading defects too - but this should actually not be the case with Normal maps: They have a baked in depth
and typically should be used at 100%. Or MW doesn't like the Normal map at such small tiles...
Whatever the cause is - I could see the defect far too late for my liking and Fire didn't help either.