C4D Anti-aliasing settings; Geometry vs. Best

#1

The documentation here https://help.maxon.net/us/index.html#DRENDERSETTINGS-RDATA_GROUP_ANTIALIASING states that the “Geometry” anti-aliasing method uses 16x16 sub-pixels to to smooth all objects. However “Best” gives you a choice of how many sub-pixels you want to use, up to 16x16. If this is the case, why is using “Best” with min and max levels of 16x16 so much slower than using “Geometry”? Should they not be equivalent?

I’m just curious to know what’s so technically different about the algorithms that are being used to make “Best” so much slower (and not equivalent). With a min and max level of 16x16 and a threshold of 0, why is it not equivalent to “Geometry?” What is it doing differently?

#2

Look under the section called Filter in your link and it mentions the different base methods for antialiasing: Geometry (object edge smoothing) and Best (color edge smoothing), so Best is calculating more than just edge data by including color data as well, explaining the difference in time calculation.

#3

In simple terms, the geometry AA mode uses 16x16 but only on the edges of 3D models. This means you’ll get perfectly smooth edges, but nothing else which can cause jagged edges or noise will be considered. So a hard raytraced shadow edge will be jagged. Reflections in mirrors or refractions through glass will be jagged. Small detailed textures with hard lines or fine grain will flicker and shimmer.

Best AA mode goes a step further by also smoothing out all of the above scenarios, hence it is slower. One thing to note, Best AA mode does also still do the 16x16 AA on geometry edges. So in theory setting Best AA mode to 1x1 min + 1x1 max will act identically to Geometry mode.