In part, it's because I never bothered to optimize it. It's likely possible to get sub15 with good momentum control, I just had to push something out fast to prove it's possible.
@@Selicre doesn't run too bad especially considering most of the game is going to be frame by frame anyway. try using fast forward mode during loading/cutscenes, it helps a lot
@@lexikiq Yeah, true, but stepping a frame and having to wait a whole second for it to render is a bit silly. Resolution changes help, but only slightly because then you don't actually see what you're doing.
![[WIP] [TAS] Sega Atlas: Sonic 3 & Knuckles "ring attack"](/img/n.gif)
Wow, my record's really close to the tas!
In part, it's because I never bothered to optimize it. It's likely possible to get sub15 with good momentum control, I just had to push something out fast to prove it's possible.
Holy cow
glad to see libTAS has already been patched for MM support :)
It hasn't - this is my own fork which allows for replays. libTAS is not very usable due to just how slow it is to render due to the usage of llvmpipe.
@@Selicre doesn't run too bad especially considering most of the game is going to be frame by frame anyway. try using fast forward mode during loading/cutscenes, it helps a lot
@@lexikiq Yeah, true, but stepping a frame and having to wait a whole second for it to render is a bit silly. Resolution changes help, but only slightly because then you don't actually see what you're doing.
@@Selicre ah yeah, guess weaker graphics cards would have some difficulties running this at any reasonable speed ;p
@@lexikiq The GPU isn't even involved in this, it uses CPU rendering via llvmpipe to achieve determinism.
Damn.