Talk:Fast approximate anti-aliasing

Edge-detection
The section about edge-detection - specifically, the claim about it being "a depth aware search" - doesn't seem to be supported by any of the (surviving) references given, and doesn't appear to be correct.

When the author says "Finding edges is typically a depth-aware search, so that pixels which are close in depth are not affected", does the word "typically" refer to common real-world implementations of FXAA? (In which case (a) those would appear to be enhancements to the FXAA, rather than part of the technique itself, and (b) the claim should be supported by additional references regarding those specific implementations); Or does it refer to how things are done in *other* anti-alias techniques? (In which case, the missing sentence is presumably something like "FXAA, however, omits this step").

- Paul Bristow — Preceding unsigned comment added by 80.195.5.244 (talk) 15:52, 3 September 2015 (UTC)

Uncited trash removal
I removed a section about how FXAA works since it claimed that FXAA is depth aware, when it is definitely not at all since it has no access to the Z-buffer or any depth buffer. It only has access to the allegedly aliased rendered image. Jesse Viviano (talk) 08:55, 15 June 2020 (UTC)