Experienced Forum User, Published Author, Player
(127)
Joined: 12/3/2011
Posts: 1579
New route
Link to video
As best i can tell, this should skip killing bugs, and everything in forest temple other than boomerang. I'm not sure at this point though. Ending blow can be learned elsewhere, and bite can be used to get to kakariko. I guess i'll need to do a test run to make sure everything actually works.
Route should apparently be the usual up until twilight, then skip bugs, do EMS, go to forest temple, get boomerang, savewarp, escape twilight, rupee dive to lakebed, then BiTE, eldin twilight, etc. Still subject to change though.
Experienced Forum User, Published Author, Player
(127)
Joined: 12/3/2011
Posts: 1579
sudgy wrote:
I would also rather wait for 4 player support than make one with one player. I think it would be a lot more entertaining, and faster.
Do keep in mind, that other than otu0001 who has made all of one commit, there is not a single person working on dolphin right now who cares about tasing. There is a good chance you will be waiting quite a while for 4 player tasing.
Experienced Forum User, Published Author, Player
(127)
Joined: 12/3/2011
Posts: 1579
Definitely not. gpu hardly matters at all for dolphin, unless you want to run at very high resolutions, which i don't. I've got a 5770, which isn't the best, but will more than handle dolphin.
It's the same either way.
Experienced Forum User, Published Author, Player
(127)
Joined: 12/3/2011
Posts: 1579
Pulling out the lantern is 5 frames slower than swinging it when it's already out.
Also, my new cpu (2500k)+mobo will be here today, so i'll be restarting this tomorrow, once i've got it set up, and overclocked, and all that fun stuff.
Experienced Forum User, Published Author, Player
(127)
Joined: 12/3/2011
Posts: 1579
CoolKirby wrote:
The "TAS-input" branch includes the changes from the "more-save-fixes" branch, so while you don't need the TAS input window, those save fixes probably helped the movie sync as well as it did.
Anyway, it's great that you were able to make a Wii test TAS that synced perfectly. I'm starting to see the promise this holds too.
Right, that's exactly why i merged it.
I've been tasing nsmbw, and have yet to get a single desync (i've done the first 2 levels 3-4 times each so far, including with heavy save state usage) since merging more-save-fixes. I got one desync in the first two levels before then, which i was able to detect when playing back from save state with read only.
At this point, it seems that wii tasing is perfectly viable. I can't say yet how well it syncs, but so far from my limited testing it appears to be perfect and/or almost perfect, at least in nsmbw.
If anyone was waiting to tas a wii game, i think i can now recommend giving it a shot. Just watch for desyncs, and let us know if you get any.
Experienced Forum User, Published Author, Player
(127)
Joined: 12/3/2011
Posts: 1579
Speed is a float at offset 15e39ac.
X position is 11fdf64 (4 bytes).
Y position is 11fdf74 (4 bytes).
And at 154e282 there is a completely useless 2 byte counter that counts how many frames you have been moving for, resetting any time you stop.
Running down slops has a higher acceleration, but lower max speed than sliding down, so it's fastest to start running then go into a slide mid way down. I lied, i forgot to hold right while sliding.
It doesn't matter where you hit a flagpole, it always takes exactly the same time to get down.
Experienced Forum User, Published Author, Player
(127)
Joined: 12/3/2011
Posts: 1579
Any chance someone who actually knows wtf is going on can help with merging Source/Core/Core/Src/IPC_HLE/WII_IPC_HLE_Device_usb.cpp?
Both nitsuja and otu0001 completely rewrote something in there, and i have almost no idea what the code is even doing, so trying to merge it without breaking anything or losing changes by one or the other of them is going to be impossible for me.
edit:
http://dl.dropbox.com/u/11111638/Dolphin%20desync-fixes%20x64.exehttp://dl.dropbox.com/u/11111638/Dolphin%20desync-fixes%20x86.exe
This is the desync-fix clone i linked above (rbc61dbdf58a8bb088109839c5f25e7183830b166) merged with tas-input. I defaulted to otu0001's changes in WII_IPC_HLE_Device_usb.cpp, as mentioned above. I don't know if there's anything important missing there, but hopefully not. Until someone who knows wtf looks at it, there's nothing i can do about it. I did send otu0001 an email and suggested he merge more-save-fixes himself though, so with any luck he'll do it right.
This should not be any different for gc controller movies, but it's possible i fucked something up when merging, so i'd advise against using it for those.
If anyone wants to test it, the biggest things to look for are that playing back from save state is always the same as playing it back from the start. It still desyncs even without save states, but it seems quite rare compared to what it used to. Extremely limited initial testing suggests save states aren't causing desyncs either, though obviously it's hard to say at this point. Also, do note that for wiimote movies you must always restart dolphin every single time you start a new recording, or when playing one back. It won't work if you don't.
Experienced Forum User, Published Author, Player
(127)
Joined: 12/3/2011
Posts: 1579
Just tested it out without the more-save-fixes branch merged in, and it seemed to sync quite well with wiimote, even with heavy rerecording. I've gotten some desyncs with it, but they were rare (vs almost constantly before), and playing back from save state showed the desyncs, unlike before where it almost always synced from save state, but never from the start.
Limited testing is limited, but this is quite promising, even after just one commit. I'm going to work on merging more-save-fixes now, then i'll post builds.
In the mean time, if anyone is interesting, 64 bit only is at http://dl.dropbox.com/u/11111638/Dolphin%20desync-fix.exe
Experienced Forum User, Published Author, Player
(127)
Joined: 12/3/2011
Posts: 1579
http://code.google.com/r/otu0001-desync-fix/source/list
No description of the one commit, and there's a ton of stuff changed, so i'm not really sure what is going on here. Of note, i see wiimote save state fixes, a desync checker (no idea how that works), and something to do with undo save state (no idea what is changed here yet, though it was interesting since it always causes a desync as is).
I'm going to compile it now and see how it goes.
Also, i talked to skidau, and i believe he will be merging more-save-fixes and tas-input to master.
Experienced Forum User, Published Author, Player
(127)
Joined: 12/3/2011
Posts: 1579
The save fix doesn't matter. It only affects recording, not playback. You'd have to be an idiot to use anything else (actually i think i'm going to go ask skidau to merge it since it seems nitsuja isn't coming back any time soon..), but there's no reason not to accept it.
Experienced Forum User, Published Author, Player
(127)
Joined: 12/3/2011
Posts: 1579
Meshidaru wrote:
Sorry for asking this here, but does anyone know if there is a Speedrun or a TAS of Skyward Sword? I'm looking for it, but I don't find anything...
Thank's.
Experienced Forum User, Published Author, Player
(127)
Joined: 12/3/2011
Posts: 1579
Hutch wrote:
Thank you for complementing our run and offering your experienced opinion on whether it should be published, but please correct me if I'm wrong, but aren't the rules/guidelines in place in order to not publish movies that the author has not bothered to complete?
I'm afraid not. One of the rules is that the game must be emulated well. This is of course quite subjective, and it is possible to publish runs that have serious emulation problems. However it really doesn't get much worse than not being able to finish the game. From the rules:
The rules wrote:
Movies of games that are not emulated well (have graphical or functional glitches that do not exist on the real console) should not be submitted.
In our case, it is an unfortunate and unavoidable problem with the emulator which stops us from completing the game, which we would obviously want to finish if the problem could be fixed.
No, it's not your fault, but neither are most rejections if you don't consider terrible submissions that never should have been submitted.
From what I have learnt about this site, the existence of TASvideos is to provide entertaining and superhuman gaming movies for the public, and from what I can gather from the comments made in this thread, our TAS thankfully fulfils these requirements.
Well, it meets some of the requirements, but not all of them. Finishing the game is a big one. Futhermore, the level of optimization isn't quite up to the usual standards. On it's own, it could probably be published despite its flaws, because it really is an great and enjoyable run. However, when there are other issues, this is only going to hurt your chances.
I feel it would be a shame for it to not be published for this single reason.
Experienced Forum User, Published Author, Player
(127)
Joined: 12/3/2011
Posts: 1579
Spikestuff wrote:
rog wrote:
Finally got around to watching this. There was rather shocking number of places where very minor improvements were obviously possible. Overall, it still looked great though, and was awesome to watch, and i'd love to vote yes on it. Unfortunately though, the game was not completed, so i have no choice but to vote no.
I'm still trying to figure out why you didn't vote Meh instead due to emulator issues to this tas I too tried to see if it can still work but obviously I too got the exact same glitch with that emulator compared to my other emulator (psxFIN not JIN).
Because as awesome as it was, finishing the game is kind of important. This specifically breaks two rules.
Experienced Forum User, Published Author, Player
(127)
Joined: 12/3/2011
Posts: 1579
At 5 fps, it takes 200 ms after hitting frame advance to advance a frame. Don't tell me that doesn't matter. It's also not true that you'll always be using frame advance. There are many situations where you would actually want to speed up the game. Not to mention what comicalflop said.
I think the even bigger problem with GC/Wii TASing is getting a hold of the ISO's and Wii game files, since those aren't readily available on the internet as much as ROMs for other game systems.
Experienced Forum User, Published Author, Player
(127)
Joined: 12/3/2011
Posts: 1579
And yes, GameCube and Wii games are mostly the longest and most complex of all the console games that have runs on this site. Not to mention, some of the runners have really annoying real life commitments that prevent them from working on their long, complex runs. Please just be patient.
So you're not going to drop out of school to tas ttyd full time? :(
