 My research retreat was very productive. I didn't quite finish implementing sparse multisampled path rendering on GPU, but I got close (hopefully soon). I ran into a performance snag but have a good idea how to fix it. But the real fruit is probably ideas for even higher quality path rendering - a hybrid of analytical area AA and multisampling - and also compositing. Expect to hear more before long. 3 comments
@nical I really wish I knew the answer to that. If people don't care about antialiasing quality, then I should be focusing my attention on other things. If they do, then this new idea is potentially very hot. @raph I feel I poorly worded my question. There is no reason you shouldn't strive for excellence, especially if you see can a path that leads you there. I hope I can get away with cutting a few more corners though! I have to admit that I've been using piet-gpu as an excuse to convince myself I can ignore conflation artifacts (until now that is) |
Gregory's Server
@raph how much do you think people care about higher quality antialiasing? Skia is looking like they are going all in with MSAA in graphite and that kind of swayed my confidence that better is really needed