That’s the idea, yeah. I want to have an optional path to do everything on GPU, but only if it’s proven to be a win (i.e. if the CPU would otherwise be idle it’s not a win). There’s more overhead when doing everything on GPU in compute shader.
The problem with SDF is that it’s inherently hacky — there are always cases in which it fails because every texel can only represent the distance to one line (in the case of mSDF, 2 lines). So the question is how to make good enough heuristics. msdfgen is really complicated :\
-
-
(This is why we can’t ship SDF on the web, fonts and SVGs would break)
-
So atleast SDF is predictable, MSDF glitches out thats bad. Atleast SDF is really just resolution dependent. Right or is SDF also lossy, i thought that with enough resolution you can always SDF something.
- 4 more replies
New conversation -
-
-
No i know msdfgen is one of those 'great i'll never fix a single bug in it' kinda things :)
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.