Joined: 4/2/2008
Posts: 103
Location: The Netherlands
SmashManiac wrote:
Huh, I didn't know that! Learning something every day...
It only works if you approach them with the powered up pendant (otherwise it doesn't even ask the question of course) - I guess the idea is that the interaction with Lavos' energy makes the item evolve over the next 400 years. (or is that edging too close to Chrono Cross lore?)
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
Sarcasm is more snarky than that quote. This was just joking. It would be ironic that the improvement was found if, say, the TASers had said "out of all the levels that we've optimized, this one is the least likely to be improved". And one of the TASers could sarcastically have said something like "Well, this really makes my day" if the improvement really had meant starting over.
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
partyboy1a wrote:
It might have something to do with the 2^16 colors limitation of the SNES.
Should be 2^15 - if I'm not mistaken, the SNES uses BGR555 internally. There are several ways to convert that to RGB888, the most mathematically correct is probably
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
My word, that was a dinosaur on a mission! Perhaps he was looking for whoever stole his feathers?
Looks polished, and the length keeps it from getting boring even to the untrained eye - yes vote.
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
Instead of calling it 'TAS mode' you could always call it 'desync proof mode' and put a note in the manual about it being slightly less accurate. Then TASers can make the choice on their own.
Another possibility would be to put 'resync' events into the TAS files, so bsnes knows when a savestate was made and loaded and can synchronize its cores accordingly. Depending on the type of game and/or dedication of the TASer, this could involve a lot less actual synchronization (and it gives you a definite number of reloads-actually-used, if you wanted to know).
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
I'm pretty sure you should be able to luck manipulate getting Renzokuken when you want it. It's easier when Squall is the only party member alive, but if you just keep cancelling his attack (triangle button) he'll get it eventually even with reasonably high HP (possibly not with full HP, I dunno), and you should be able to do that very quickly in a TAS. It's annoying when you're doing it by hand because you'll be cancelling so fast you're likely to miss it when it appears :P
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
If playing cards for any length of time is viable through luck manipulating, I'd say get some Abyss Worms and Mesmerizes (I believe that's what they're called) for Tornado and Regen respectively. Tornado is good for basically anything and Regen is very nice for HP; their cards also have a fairly high chance of showing up. Using them you'll be well prepared for the spider bot, which you can 'kill' (that is, enough to get the 50 AP - actually killing it might still be faster than the cutscene though) in a single renzokuken with Tornado junctioned to strength. (followed by a normal attack to stun it again). I don't know if the 200 AP you can get from him is useful with only the first four GFs though.
If playing cards is out of the question, getting some fish scales from the coastal area for some Waters might be a possible second best, assuming you can manipulate them coming out of the ground right away.
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
scwizard wrote:
Also I'm against any option that can't be called "as fast as possible", so #2 is out of the question.
I think option 3 fits the definition of 'as fast as possible' the least of all options. At least with 2 you can say it's as fast as possible barring one glitch - with 3 it's up to the runner's discretion (which I'm sure is fine, but not objectively definable).
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
Baxter wrote:
Ver Greeneyes wrote:
2) Using tools, the trick isn't hard to do - it's like faster text scrolling using the Japanese version of a game: it's faster, but it isn't a testament to the skill of the TASer.
NxCy could probably write an essay on why this isn't true. I'll only say, give it a try and see for yourself.
Sorry, I guess that was a bad assumption to make. I figured that since it's something used throughout much of the game it'd either be simple or you'd be able to script it where appropriate. Well, I do still stand by my other reasons :P
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
Hmm, tough one. Well, I look at it this way:
1) We already have a run that completes the game as fast as possible that demonstrates this trick. (the any% run)
2) Using tools, the trick isn't hard to do - it's like faster text scrolling using the Japanese version of a game: it's faster, but it isn't a testament to the skill of the TASer.
3) For such a long run, entertainment and variation are very important.
4) A constantly wobbling Yoshi goes against the cute feel of the game by messing up the animations.
For these reasons, I choose option 2.
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
Tonski wrote:
Also, what is bruteforcing?
Tonski wrote:
systematically try different patterns with subpixel positioning and shinespark
That ;) i.e. using the brute force of modern computing capabilities to check every possibility (although humans can probably help out by making it only check reasonable ranges of values)
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
Yeah, there's a difference between codes input from some external source like a game genie, and built in codes (including level select codes, in theory, but those don't tend to make for very good runs ;) )
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
Hehe, that was mad. I am a bit worried that timing inaccuracies are coming into play here though (I'm hopeful that we'll see a version of bsnes with movie playback someday soon; savestate support is the major hurdle for rerecording and without it movie playback has little priority).
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
This run should get its own category (any% with savegame abuse / savegame corruption abuse) but I think it should obsolete the old run because their goals are the same. I think there's room for a 'best ending' run or a 100% run and an any% run that explicitly avoids savegame abuse. (but would still be a lot shorter than the current run considering findings made since that run's publication)
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
I think categories for this game should now be based on either 'uses savegame corruption' or 'doesn't use the save system'. Either way I hope this run is published soon; it's a real gem.
ShadowWraith wrote:
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
DarkKobold wrote:
I was able to play the video back fine. Hooray for SD getting PSX encoding going.
ShinyDoofy wrote:
It plays back fine for me with VLC 0.9.8a and mplayer.
Thanks for testing guys. It seems to be some kind of strange initialization issue: if I begin playback of the encode for the Banjo Kazooie TAS, then load this one while it's playing, playback begins without error. I updated to rev 2874 but that made no difference.
Joined: 4/2/2008
Posts: 103
Location: The Netherlands
Is anybody else having trouble playing back the encode using ffdshow? Both libavcodec and ffmpeg-mt instantly crash media player classic. MPC's internal filters seem to work alright. I'm using the latest 64-bit build from www.xvidvideo.ru (2872)