Achieving My Goals

Although my problems with rendering artifacts over the past two semesters were eventually corrected with hacks in the compositing stage, I wanted to avoid them altogether and invest a bit more time in the modeling instead of spending much more in the lighting, rendering and comping stages. It seems that the culprit was polygonal edges that were not correctly beveled, and triangular polygons.

I had noticed that Subdivision surfaces and NURBS objects tended to render very cleanly and without any issue in both Renderman for Maya (RmF) and Mental Ray (MR), so I ended up creating my models exclusively in polygons that were then converted into SubD’s. Using this strategy, I had a great level of control in molding the general shape as well as tweaking details. Additionally, UV layout was rather straightforward and less tedious than it had been with beveled polygons. I modeled all of my poly’s completely with quads (which subdivide the best), and would then use the Split Edge Ring tool to “shore up” the edges of the models that I wanted to be sharper or harder upon subdivision. In other words, the closer two edges were, the sharper the resulting polygon between them would become.

I had no rendering artifacts or crashes with this method of using nothing but SubD’s for geometry. Renders were predictable and also efficient, leading me to focus more on the lighting strategies (instead of bug fixes/hacks), and also adopt this method for the remainder of my modeling for my thesis project.