Hey Guys I would like to mention something that I consider very important to the speedrunning community of SMB1 and SMB2J.

It all started off right here.

If you notice keenly, you will see that FCEUX is lagging behind Nestopia despite starting at the same time. ( I enabled Background Input in FCEUX and mapped all the keys to be same in both the emulators.) So I got Curious and used the built in lag counter in FCEUX to see what was the issue. And This is what I found.

Everytime a level is started or a pipe is entered or exited or a vine is climbed a single frame of lag is added. If this is some other game this would be insufficient. But in SMB1 and SMB2J due to the framerule system this is a pretty big deal. The issue is that this adds up over the course of a run. You lose around 0.1 seconds for MWE, 0.2 seconds for Any% and more than half a second for Warpless.(These calculations are done assuming that you don't die.) This can affect runs that get calculated based on milliseconds. So there should be a system of calculation that is capable of eliminating this lag or people should avoid using FCEUX altogether.

The versions of emulators are FCEUX : 2.2.3. Nestopia : 1.40

I notice you're using New PPU in FCEUX. What if you switch that to Old... and maybe switch everything else to default settings while you're at it, for consistency. You should also display FPS and Frames in FCEUX, to give more clues. This is interesting... ruling out settings would help us figure it out.

Well I was experiencing the same issues with the old ppu as well. As far as settings go I haven't changed anything

iirc there's always a lag frame in level/pipe transitions no matter if you're on console or emu, plus lag frames are outside of framerules anyways, so this isn't really a big deal

Several points I want to make about this:

• We can't see the framerate FCEUX.
• This comparison isn't done on the original game, it's on the Pellsson ROM.
• How well can your computer handle 2 emulators and encoding video at the same time?

I don't think we have an actual issue here. We time high level runs using patterns, not the time of the video for any% runs. We do time the video, yes, but in most cases for legitimate runs, the the difference between the video time and the bowser pattern is a couple frames at most. If all FCEUX runs were off by 0.2 seconds as you claim would be possible, this would have been noticed years ago as it kinda throws a red flag to us if there's a pretty significant difference between the timing of the video and the matching bowser pattern.

Hey Guys Sorry for wasting your time. I am just dumb.