Hello!
Some of you know me already. I am the guy who submitted a run of Ocarina of Time that got highly criticized for being only sub-optimized.
I really want to get some respect on this site because I love Tool Assisted Speed runs and everything this site does.
I would like some help. What game should I do?
Also I have some questions.
1. Frame Advance. I understand how to use it but barely. I know how to hit a certain thing on a certain frame but I don't know how you guys always know what frame what needs to be hit.
2. It seems like everyone here always knows what to do. Like if I chose a game I wasn't already used to speed running where I knew a good amount of stuff (like I did with Oot) I wouldn't know what to do. I think someone on here said something like "you don't need to be good or familiar with a game to TAS it." How can that be. If I had some guidlines to follow or a route I might be able to produce something nice but I want to be like the other people here.
I have a LOT!!! of free time. I will work hard. I want to be like the big name people here but as of know I am looked at as a noob.
When I was looking at the submission so many people had like 100% yes votes and I have like 60 something THAT NEEDS TO CHANGE.
So please all you pro TASers help me... help me PWN!!!!!
Frame advance lets you use the frame counter as your bitch.
If you are able to accomplish some timed button press in 5 frames, go back and use frame advance to see if it can be done in only 4 instead. Then go back and try 3. If it fails, 4 is the best you can hope for so you know and can always nail it right on the spot.
Most games have animations, events and so forth that run like clockwork. Frame-advance lets you plan what you'll do on each frame.
Just to be annoying though, OoT only appears to read input once in every 3 frames. Expect the game to have a delayed reaction sometimes, or to just downright miss a button if you didn't press it for that 1 frame in 3 where it's actually paying attention.
That said, I think everyone would very much like to see the new and improved version and will forgive you for the old one.
I agree with everything DeHackEd said.
Also, I would suggest trying to TAS a more simple game (platformer like SMB would suffice) — for experience, not for submission — and train on it. Once you understand all the generic techniques, you will be able to apply this knowledge to any game including OoT.
See also:
http://tasvideos.org/CommonTricks.htmlhttp://tasvideos.org/GameResources.html
The first game I TASed and submitted to this site was Castlevania III. I did the Grant path because no one else was into to doing it and I always wanted to see it done. I really didn't know a whole lot about the game (I sucked at it) but I did my best and have redone it a few times now.
Basically, all I'm saying is that you should consider a game that you either like or want to see a TAS of. I highly recommend that you learn Frame Advance because it's much more useful than than just slowing things down.
Just practice and you'll learn good habits in no time. Most importantly, have fun because otherwise what would be the point.
Try to TAS another game that you enjoy, and/or play around with the emulator functions and get more familiar with them and make another OoT run if you so desire.
When I was first learning, I would try a simple game that I knew well on NES and attempt to match the current run on this site using frame advance and frame counts. You can always go back frame by frame on the currently published run along with your test run to see exactly where you lost/gained time. It should be easy to spot errors through juxtaposition.
Use this technique to check your work and find out where you may have made an error to lose time. It should help you make insights.
And moozooh brings up an excellent point: read the FAQ--it's a great resource.
There are two ways to know what frame something needs to be hit:
1) Exhaustive testing (as follows)
Make a save state you know is before the first possible success.
advance one frame
Save state
Try it and see if it works
If not, load the state and repeat from step 2
2) Finding a Target Window (if the success of a maneuver depends on the position of something similarly timing-dependant), you can do the following
Save state
Measure the number of frames it will take to get into a proper position, and perform the maneuver. Write this number down somewhere (I will call it "Len")
Find the first frame that whatever you need to wait on will be in a usable state. Write this frame number down somewhere (I will call it "Target")
Load state
Advance to "Target" - "Len"
Save state (in a different slot than your previous one)
Perform your maneuver
If it fails, load your most recent state, advance 1 frame, save state, and try again until the first frame it succeeds
If it succeeds, load the prior state, and advance to one frame earlier, to test if you got it on the first possible frame. Repeat until it fails because you started too early.
Well, moozooh already linked you to the list of general tricks, which is a good place to start. Some games (super mario world, megaman, etc) also have specific lists of tricks for them.
Remember, also, that it is easier to fix something early than it is to fix it late, as many games are very easy to desync when hex editing, and most emulators won't allow you to record from a certain point without erasing everything after it.
What I always do when TASing a game I'm not totally familiar with, is do a sort of sloppy test run first, to get comfortable with the engine, and find glitches. Then I start over from the beginning with a more serious run. I understand that this isn't always feasible with longer games (like Ocarina of Time), but you can still do something like a sloppy run through an area to find a route, and then redo that area with tighter timing and such.
How fleeting are all human passions compared with the massive continuity of ducks.
Emulator Coder, Site Developer, Site Owner, Expert player
(3569)
Joined: 11/3/2004
Posts: 4754
Location: Tennessee
I couldn't disagree more with the suggestions to do another game.
Redo oot. The best way to learn how to TAS is to redo the same game several times and attempt to keep improving it.
Use the frame counter. Take a room or small segment and find a definite end point (like the black screen of a fade out) and write down the frame. Then redo the level until you succeed in doing it faster (even if by only 1 frame). Make a savestate but then redo it again. If you were doing it at full speed, then cut down to say 70% and try again, then an even lower speed, and so on. Try different strategies and see which one is fastest.
Most of my published TASes are not my first version of the game. By improving your previous work, you learn so much more about the game, and about how to make a good TAS.
Joined: 4/25/2004
Posts: 615
Location: The Netherlands
Start with games you played on a normal console. Games you finished over and over again. Those are the games you probably know every trick of and that will help you along the way. But I wouldn't exactly go into areas like Zelda or Mario. Those games have been run so many times it's really difficult to improve them.
Yeah, I think it's very true that you don't need to be good or familiar with a game to TAS it. When picking a game, it is more important that you like the game (since you will be spending lots of time with it) and that it's suitable for TASing (a game is generally suitable for TASing if you can make the TAS look really different than a normal speedrun of that game). When TASing a game, you look way more analytically at a game then you would normally do. You check what exactly is needed to reach the goal of your TAS, what exactly is needed for that goal, and what is not needed. You check what input causes what effect, and try to understand how the game works. You think of all things that could possibly be the fastest to do something, and try them all out. If you have decided which strategy is fastest, try to optimize and perfect it. When you are done with something, check out how it looks in realtime (this is what people will eventually see, and can possibly motivate you). It is also good to upload your progress, since others at the forum might have some good ideas, or see a mistake somewhere.
Well, he should indeed redo it, but maybe doing another game in between might not be a bad idea. I can imagine N64 input can be annoying (way more buttons, and doesn't always respond), you have to watch desyncs, and a ~2 hour TAS is quite long... Maybe doing a ~10 minute NES or SNES TAS to get better known with general TASing strategies and tools, such as frame advance, might be a good idea before redoing OoT
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
Adelikat wins... Redo OoT is the first thing I thought! The only thing that people complained about was accuracy? Well, do that!
About knowing other games: you'll have to read about them on the net, make some test runs (rough TAS just to test the route) and then see what needs to be done, plan, tests, and then make the final thing. For a really long game like OoT I agree that it is easier to know the game before hand, but most games here are 20 minutes long movies, so you can TAS a completely unknown game if that's how long it takes to finish and still come up with a good result. But knowledge IS a key factor, so use it to your advantage.
But understanding the ram watch tool is a good way to understand better how games work, so you'll want to try that one on some game (does mupen64 has it? I've been idle for too long)
It's not as easy as "do that". First he'll need to learn to use the emulator tools to their full potential. Fooling around with a different and known game is a good way to do so.
Joined: 11/18/2006
Posts: 2426
Location: Back where I belong
I know I've learned more about TASing in general by working with several different emulators and trying different games, so if you're concerned more about future TASing, I'd start out with that. Tricks that you learn about in one game are sometimes usable in others.
But since you know a lot about OOT, it may be to your benefit (for now) to stick with Mupen and make another run of OOT that you can practice on. Sometimes totally taking apart a game is a great way to learn also. It just depends on your learning style and what you intend to do in the future.
You also ask about how you can make a TAS w/o knowing anything about the game. With frame advance, you take a level and tear it apart step by step, so you can chose different routes without much hassle, and try different things easily. It may be a good idea to do a fullspeed run through of each level before you try and TAS it, but other than that, you'll find a lot of cool tricks just by taking things slowly. However, one of the risks you run is finding out a general movement strategy or trick that speeds up everything, but you happen to find it near the end of the run. Having to redo an entire run is frustrating, but also often improves other areas you thought were perfect. Anyway, hope some of this helps, and let us know how you're continuing.
I am thinking about giving up TASing.
Last night I was messing around with the emulators trying stuff and I just suck so bad. I got mad and deleted everything on my computer related to TASing. It takes so long and I still don't really understand how to do it.
And when I got SOOOOO much hate for my OoT run that really depressed me. I know you guys were being honest I don't want you to lie and tell me it is good when it sucks but I just like TASes so much but I can't make them to
save my life.
Joined: 7/4/2005
Posts: 714
Location: Albuqueruqe, New Mexico
Dude, it just takes a lot of time and patients. They would have published your run for OoT if there were not so many visible mistakes. You missed a lot and thats from not reading the 170 pages of useless junk to find out what you missed. Start on NES games. They are the easiest to get familiar with. Something that can be improved and i think has a lot of potential is SMB2. I feel like there is a big ladder glitch that can lead somewhere. I have found one but when I tried to re-create it, it didnt work again :(. Keep trying duuuuuuuude
Change my sig. again, and I will murder your pet fish.
It wasn't hate. It was criticism revolving around the site's standards. I still believe it's a good thing you submitted the run, it served many good purposes. No-one hates you for doing it.
I love SMB2 and I speed run it unassisted a lot I know all the tricks.
Another problem I have is finding working or the correct version rom (I am not asking for help finding them I know it is against rules). I am just saying like I can't find a lot of games I would be into doing.
Also on Mupen (before I deleted it) I had TAS input plugin.
I would be patient if the product that came out was good. When I try to make it 100% optimized it takes way to long like I am doing something wrong long.
And when I do it in a way that I make reasonable (still long) progress it get HIGHLY criticized (as it should).
This might just be something I can't do (it wouldn't be the first)
Also is it possible with TAS movie editor to copy some frames from one of my .m64's and pasting in another? It let me do it but it wouldn't let me save it said "file possibly locked"
i thought you have so much freetime, a 100% of optimizing something sure takes loooooooooong! longer then you might think off but even then the result is awesome!
you just need some time to get more into it then you will be likely get better results a bit faster!
If you're having fun, keep doing it! If not, find something else that is fun. Just remember that whether or not your run is published doesn't have to change how much you enjoy TASing.
Since you have a lot of free time, and some interest, you may find that using this time to persevere through the initial learning stages of TASing can make it more fun later.
As to which game you should do, I would recommend OoT again. If you do another game first, the initial TASing experience will be helpful, but you will still have to get the hang of this game, instead of having knocked both of them out at the same time.
Also, if you try another emulator first, you may not want to return to Mupen. <_<
<Swordless> Go hug a tree, you vegetarian (I bet you really are one)
Joined: 11/11/2006
Posts: 1235
Location: United Kingdom
Perhaps your best bet would be to make a loosely optimized run of a game, then when thats done, break it up into little sections and try to optimize each section. This will probably make things much easier for you. Of course you'll have to experiment also, seeing if you can push the game's limits to the edge and over.
For the record, your TAS of OoT is better than anything I would be able to produce.
<adelikat> I am annoyed at my irc statements ending up in forums & sigs
A lot of you are telling me to redo OoT? What is the point?
By the time I finished wouldn't GuanoBowl's be complete? He has a head start now. And he is a pro mine would not be able to stand up.
No-one is 100% sure whether GuanoBowl will even finish it, let alone when. If you like the game, why not redo the run? You already pulled it off once, you are technically able to do it again, just with better precision (much easier than starting from scratch, if you ask me).
This is your first submission, right? In my experience, first submissions have a tendency to become gruefood. N64 games are especially rough. However, resubmissions can be much better, and closer to perfection. I say: Go ahead. You may surprise yourself with how big an improvement you can make.
You didn't. There was a lot of arguing about the quality of your run, but nobody hates you for it. The criticism was directed at the run, not at you!
Most people who submit their first TASes face a rejection. Yours even got a lot of yes votes, and many of us enjoyed watching it. For your first TAS, that's quite a good result.
Frustration cannot be defied by logic, but it'd be a shame if you gave up because your first attempt couldn't compete with the site's standards - standards set by people who have been TASing for years. Keep on learning and you'll get there. As long as you have fun doing it, there's no reason to stop.
no, he isn't (sorry guano :p). His first attempts at OoT-TASing weren't better than yours. He's been learning fast because he's been releasing WIPs and got a lot of valuable hints and feedback from this community. But nobody is keeping you from doing the same. You've got the potential and this community is usually helpful when asked nicely.