Super Mario World Any% comparison v3

There is a new Super Mario World any% TAS record by bahamete, kaizoman666 and PangeaPanga, so here is a new version of the Super Mario World comparison video, comparing the new world record with the previous records. The previous version can be found here (http://www.youtube.com/watch?v=yEJhrU1EAhk).

The videos compared are, in order of speed, those of ISM&Mister, Mister, Fabian, JXQ, Flagitious, Neuhaus, Volkov and Bladegash, where the screen itself follows bahamete&co's run.

To avoid having the slowest runs left behind after the first course, the runs are re-syncronized at the beginning of each level; the number of frames (60ths of a second) they had to be shifted for them to sync up that way is written at the top of the screen, in colored boxes in the same order as the author list above. These numbers thus represent how far each run is behind the current record at that point (note that these numbers might temporarily show nonsensical values during transitions).

This video was made with an improved version of snes9x called snes9x-rerecording, which can be found here: http://code.google.com/p/snes9x-rr/ This version not only supports the LUA-scripts used for making this video, but is also well suited for making the Tool-Assisted Speedruns (TASes) this video shows off.

The script itself can be found here: the script itself can be found here: http://tasvideos.org/forum/viewtopic.php?t=6539&postdays=0&postorder=...

Category:

Gaming

Tags:

• Super Mario World
• TAS
• comparison
• lua

[wuschelbeutel]

this video has been obsoleted again. hopefully, you will find some time to do an update. these are very enjoyable and interesting videos. also, perhaps you can do one for mario kart on the snes...there have been two big ones: 23:10.88 by Huffers.and a 20:29.600　time from some Japanese TAS'er.

[amaureaLua]

@wuschelbeutel The v4 video has been out since december 25th, so you don't have to wait to see it. That includes the newest not-totally-glitched super mario world run. There is an even newer totally-glitched run which is not included, since it basically skips the whole game and goes straight to the ending. That makes for an interesting movie, but not an interesting comparison. Anyway, you can see the v4 video on my youtube channel.

[wuschelbeutel]

i think the best video is over 1 min better than the worst video, so why does it say 2236 frames on the boss fight? 2236/60 << 60

[amaureaLua]

@wuschelbeutel The numbers at the top are the delays when entering the room, so part of it could be due to finishing bowser off more slowly. But that isn't enough to explain the whole difference. My guess is that the slowest one ends input much later, perhaps after the celebration scene.

[neurario]

So, description in mind, what emulator is used in relation to recording romhack runs of this engine? I seem to have some issues recording with either Zsnes (desync) or snes9x (the rom crashes?).

[amaureaLua]

@neurario It should work fine with the snes9x version I linked in the description. It definitely shouldn't crash. For example, TASes of Super Demo World are routinely made using snes9x-rr. What do you mean by "this engine"?

[neurario]

@Zeldara109 That would probably be why! The music would cut out usually by the second level, and then after dying/completing that level the ROM just seems to freeze. I'll try 1.51! Thanks! :)

[Zeldara109]

@neurario I think I might know what your issue is: Most SMW hacks made in 2007-2010 that use custom music will crash in Snes9x 1.52+ (the older tools for inserting custom music exploited an emulator glitch), so you'll need to use Snes9x 1.51 or earlier.

[neurario]

@amaureaLua The Super Mario World engine, sorry about that. :) Ah okay then, I'll have to look into that then, I've been messing around with the Vip And Wall Mix hacks, and want to try and run those, for fun :p

[amaureaLua]

The slowest run does not press a button as early as possible during the title screen and intro screen. Especially the intro screen is sloppy. It first presses buttons too early, when the game is not ready for input, and then tries again too late.