Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
Saturn wrote:
You can't even start the boss fight without Ra-Dewdrop.
Oh, I guess I misremembered about this. Maybe I got a regular poison, or tried taking that item off during the fight, or just sucked at the fight without being poisoned.
Saturn wrote:
The Lizards give you 22 exp each which is the most you can get per enemy in this time. In the previous dungeon (Grecliff) you get only 17 for a stone golem and 18 for a quaker/mini quaker.
This is what I meant - the time it takes you to kill the 1 quaker and leave/re-enter seems to be shorter than 18/44 the time it takes you to kill 2 lizards and leave/re-enter. If that's true, then it should be faster to move all of the lizard leveling back into that room with the quaker, even though it'd be even more boring than usual to go up that many levels in one place. Maybe it's just an illusion or wouldn't be true in a TAS though.
Saturn wrote:
If you worry about the leveling up you should make or get a list (I think GameFaqs has one) of all enemies to see how much exp they give and just seek for a as small as possible room with as many good enemies as possible in there. If you aim for a in-game time you don't have to worry about exiting and reentering the room since it doesn't count the time during the screen-blackout. The leveling up is still the main problem IMO so it just needs to be tested.
OK. I would take the screen blackout time into account, because it won't be particularly entertaining to do otherwise if it makes a difference.
Saturn wrote:
Also do you know the time (in-game or real) you have reached in your test-TAS at the point after you defeated the spider-boss in tower 5?
I'm curious to compare it with my own non-TAS time to see if I did well in this any% run.
Wow, actually I just found a backup of the run from a search of a bunch of directories. Watching it again, I'm seeing lots of mistakes... Anyway, I went from level 2->3 during the spider boss battle, and it died on frame 90000, which is 2 minutes and 48 seconds faster if not counting the name entry screen.
BTW, what does "Text" do in the options menu? It looks like a text speed setting but I've never seen it have any effect...
EDIT: My goodness, somehow it never occurred to me how long this game really is. I saw your movie title said 4:34 minutes at the last save, but the total recording time is over 6 hours! It'll take quite a few miracles just to get the real-time under 4 and a half hours.
I wish a US version of this game had been released, that way it wouldn't be necessary to decide between PAL+English and NTSC+Japanese. There is a PAL->NTSC patch for the English version, but I don't think that would be allowed, and far too much of the movie consists of dialog for me to want to try the Japanese version.
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
Bag of Magic Food wrote:
Maybe there's an extra component to the save file that we haven't found.
There's only 1 save file, which has extension ".sav". FractalFusion, are you sure you are deleting the right one, and did so while the emulator was not open? Also, you did make sure that "Options->Emulator->Real Time Clock" is turned on (and that you're using v15 of the emulator), right? It's possibly a newly-discovered emulator problem, but I thought this had already been tested by a few people with the times being correctly loaded from the movie file.
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
Saturn wrote:
You can't skip Ra-Dewdrop. It is needed to survive the poison-gas that is in the room with the boss. Ra Dewdrop makes you immune against this poison gas.
But the poison isn't immediately fatal. Does it also make the boss invincible? When I first played, I think I didn't realize I was supposed to get the Ra Dewdrop and beat the boss anyway (although it was very hard and took lots of healing items).
Saturn wrote:
The room with the stone-heads (known as mini-quaker) is a very clever choice but hard to pull off without tools because they appear only 1 time and if you miss them you can't get them again.
Actually they reappear when you are a certain distance away from them. And avoiding their shots isn't too hard when you do a dash attack toward them. But for a non-assisted speedrun it's still probably so difficult that it would waste more time than it saves. (note also: those two lizardmen in Zu actually don't give very much experience for how long it takes to kill them - I think they give exp at a lower rate than even the leveling spots in earlier dungeons.)
Saturn wrote:
About making a TAS: Aren't you interested to make one yourself for this game? After all you are a very good TAS'er and have a good knowledge of this game. I already have 2 games (Super Metroid and Chrono Trigger) in the queue so I could only start a TAS for Terranigma after about 1 year. I would be anyway glad if someone else could do one instead.
Yes, actually I started one quite a while ago. I put it on hold when I accidentally erased most of it, but mostly because I wasn't sure of a leveling strategy (not because I'd gotten very far; I hadn't). In that test run, I was at the minimum possible level (2, or maybe 3) when I beat the last-tower guardian boss (and it was a much faster fight than the one here where you're at level 5(?) - manipulating luck goes a long way in this game, although I'm still confused as to how the criticals work and would need to analyze them more before making a run).
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
Oh, right. I was only talking about Snes9x (since that is what forum this is in). I find it way more convenient to turn a knob than to open a menu, go to a dialog, and drag a slider bar, anyway. (I usually change the volume many times even during the game depending on how annoying the music is getting.)
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
I've been meaning to watch this sometime but haven't seen much of it yet. I noticed some things that could be done faster from what I watched so far, though (things to keep in mind if/whenever you make a TAS of this game):
- In the intro, it's faster to say "No" twice and let the other guys try to open the door themselves (there's less text to scroll through that way).
- You don't have to throw the pots straight at the front of the door. You can throw them through the wall without even moving much from where you pick them up.
- When exiting any area of a dungeon (going in a door/stairs to the next level), it saves some time to do the run+attack move into it.
- When climbing up the chains on the side of a tower, you can jump then press up to grab onto the chain higher up to skip a little climbing.
- You don't need to equip the Crystal Thread to give it to Elle.
Other things for a TAS:
- I believe you now that levelling is necessary, but I think more efficient locations could be chosen for some of them (that would be harder to do normally, like the room with the 7 little stone heads that come up and shoot at you).
- I think skipping the Ra Dewdrop would be a good idea.
- Does resurrecting Mu really help that much?
- When you hit a switch or break a door or talk to someone, or anything like that that pauses you temporarily, you can almost always make it so you slide toward where you want to go during the pause. It might require jumping and fast-turning sometimes, and definitely requires running.
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
Hmm... I think there must be something wrong with the way the "In the Talons of the Big Bird" star is being done here, unless either Curtis Bright's record time on the level isn't right or he used the wing cap - he got 0:17.x and this only gets 0:19.7 on that star. (Fastest Star Times)
edit: And yes, I believe those times are fully compareable for runs of this game, and most of the other (applicable) records listed are at least a couple of seconds slower than in this. The timing used there starts when the screen starts fading out from white to the level, and ends when mario starts dancing with the star.
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
Bag of Magic Food wrote:
That happened to Nrg too, but it worked for Dehacked and me.
I can only assume that some people are playing the movie incorrectly, from the basically-zero details that I've seen so far about the 0-frame thing.
First of all, something like this should appear in the play movie dialog: play_mario_dialog.gif
If it doesn't (if it says 0 (0) frames, for instance), that means something's already wrong - possible causes that come to mind are: not actually specifying the file's real location, trying to play the file without unzipping it first, trying to play a corrupted download, or using a several-versions-old mupen64-rerecording.
About the movie, it's looking good. Breaking the wall when you completely missed it was neat (it would have been better if it somehow saved you from having to do it a second time, but of course that's impossible since the level resets for each star). One question: why not do something while waiting for Bower's key to become pick-up-able?
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
Bisqwit wrote:
The formula used to calculate the confidence is explained in the bottom of the page.
I think you misunderstood; I was not talking about the confidence value at all. I know it factors into that.
edit: But that's neat, I never realized that was MathML - I thought it was simply written as a very confusing plain text formula.
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
Bisqwit wrote:
nitsuja wrote:
Although I think it's kind of weird that 'meh' votes contribute to the "votes" (assumedly 'yes' votes) percentage shown in the submissions list.
It contributes to uncertainty.
Yes, but it doesn't seem like it should contribute positively to the other column - for instance, if 2 people vote yes, 6 people vote meh, and 0 people vote no, I would expect to see either "100% (2)" or "25% (8)", but right now that example would display as "100% (8)".
(And when all of the votes are 'meh' it doesn't show any percentage at all there, which is odd given that the votes column doesn't distinguish between 'yes' and 'meh' in any other situation.)
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
The current confidence system seems to be doing an adequate job as it is. Although I think it's kind of weird that 'meh' votes contribute to the "votes" (assumedly 'yes' votes) percentage shown in the submissions list.
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
kooz wrote:
I would like to obtain optimal settings from one of our encoding members as I'd like to make this movie as well as others in Nesvideos-friendly format. This had been brought up in the General forum, however that topic went nowhere.
If you still haven't gotten a response, maybe you should ask them on IRC. It could be that a response to your question would require more of a back-and-forth than they wanted to deal with on the forums.
Oh, and I voted 'yes' for this movie. I think it starts out somewhat boring (because of the way the game starts), but it gets pretty fun once you have the amulet. I didn't particularly like the vehicle sections at the end, but I can't tell if there's any way to make those better, and the boss fights looked good (although too complicated to be sure that they're optimal).
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
Maza wrote:
I guess I'll have to start again the run since I had RawData turned on.. Well, I wasn't that far so it really doesn't matter.
You don't have to start anything over unless it desyncs on playback when you turn it off.
Maza wrote:
There are some minor graphical bugs with each plugin. Some are a bit different than other, but they are there.
Like what? (I haven't tried running this game yet.) Do you know of specific plugins that work better for it? Also, most graphics plugins have a lot of options to fiddle with; I bet some of the graphical glitches would go away by changing the right options, for some of the plugins.
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
OK, the EXE can be made separate. I mostly just put v7 up to see if the people who were having desync and frame advance problems could verify that it fixed them. The read-only toggle hotkey isn't finished in that it's not yet in the dialog so you have to edit it at the bottom of your mupen64.cfg file, and apparently there's an alignment bug with the cfg file saving that affects it.
Also, you should turn off "Raw Data" for now. It shouldn't matter, but apparently there is some code elsewhere that is still affected by it, so until that's fixed you might get desyncs if you leave it on.
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
flagitious wrote:
The blublu's movie desynched in the exact same fashion as before though :(
Try turning off "Raw Data" in your Input Settings, maybe that is causing the problem.
flagitious wrote:
There are two frame counts in the bottom left, one seems to take many steps with 1 frame advance sometimes, is this because the computer cannot emulate the n64 fast enough, or is this the way the n64 works, it stops accepting input for a short spurts?
The N64 doesn't check input every frame, so frame advance fast-forwards through frames where it doesn't check input. The leftmost frame count is "real-time frames" (each is 1/60 of a second in an NTSC game), and the frame count to the right of that is "input frames" (number of input checks made).
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
The best games to make runs of are the ones where speed and entertainment aren't mutually exclusive, where doing extra tricks also saves time. But, for a game where doing something that's not as fast as possible would be more entertaining, I think it's important to define and follow a definite goal that allows for it to be more entertaining, instead of just trying to mix speed and entertainment in some undefined way. Aiming for highest score in Star Fox 64 with fastest time that gets that score is a good example. Even made-up limitations are fine, I think, if they are simple and allow for the movie to be more entertaining. No damage, highest score, no failing bonus rounds, 100%/0% items/kills, no running, etc. have all been used before. None of this "As fast as possible except when it looks cooler", there should be some way to formalize that or else it's impossible for anyone but the original author to improve (or evaluate) the movie.
(edit: and this hasn't been done much but I think that "shows as many bugs as possible" would also be a good one for certain games even though it's a little vague, as long as it's defined clearly enough.)
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
flagitious wrote:
ok I will make a frame perfect short vid sometime soon to help test desync issues with, frame advance doesn't work but I'll just use really slow mo.
Oh yeah, I forgot you were having both problems.
EDIT2: Try v7 of the emulator. Frame advance should work for you now. And I'm not sure about this but it might also fix your desync problems.
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
Hmm, well the sound settings didn't make any difference. And I found an even more striking example: Try making a save state on frame 29239. Played normally through there, Dixie swims upward, but if you load the state, the "swim up" command is apparently lost and Dixie dives downward instead. Even if you pause, save, and immediately load the same thing you just saved, and unpause, that will cause desync. I tried comparing the memory states using the cheat finder, but it just said that the memory is exactly identical, until the next frame of each where many things are different in both. So it seems to me that loading a save state causes the input for that frame to get ignored for 1 frame when it shouldn't be, but only in certain cases, and I'm not sure what could be causing it... These aren't lag frames or anything like that. The game stays synchronized just fine as long as you don't save on the wrong frames.
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
RT-55J wrote:
I remember being able to consistantly "manipulate luck" in real time so I just walk through a room full of Wizzarobe thingies... so... the randomness generator sucks (and I think my spelling is).
Isn't that just you getting lucky? Or, possibly it's not all that lucky and in fact quite likely to happen. The random generator doesn't have to suck for either of those to be the case. Unless you were counting steps or timing (or something like that) and using that as a reference to make the Wizzrobes behave a certain way, which I doubt.
edit/note: BTW, I think this whole idea (of this thread's original topic) is possible, but not worth the effort that would be required for the results it would get compared to what people can already do with this game.
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
Try downloading it now. (The links got screwed up somehow. Note that the mirror download is faster, if you can tolerate rapidshare.) When I first uploaded v6 it was missing a .dll file, but I fixed that later, so you will probably still get the missing DLL error unless you re-download v6. I don't think it has to do with your version of windows.
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
JXQ wrote:
What kind of data might be on it, and how does an emulator circumvent this problem? Does this mean that the emulation may not be true to the console in some way related to the eeprom?
Because the eeprom (or SRAM etc.) is just a separate file or chunk of memory to the emulator, all it has to do is clear it out to all 0's or something like that, and it will effectively be back in from-factory condition. On a real cartridge, it's not so easy to reset that memory, as it is intended to be permanently saved and there isn't necessarily any mechanism in the game for resetting it.
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
Sabikage wrote:
I would think that doing an entire room successfully is beyond the scope of Bisqbot as of current.
In a way, it definitely is: Unless I missed some development about this, Bisqbot's current state is that it must be modified to do anything at all besides the part of the game it currently plays.
But, saying that "Bisqbot can handle it" is basically saying that it is possible for us to come up with some ad-hoc formulas for generating input and deciding failure/success that will be able to reach the goal with a reasonably high probability. The question is not whether such a method can work (it is quite general, if not all-encompassingly so), but whether it will be easy enough for us to find the right formulas/algorithms. The algorithms it has used so far have been highly randomized (although still with some intelligent guesses built-in), because it's really easy to implement that way compared to almost anything else, but that doesn't mean these other possibilities being discussed are something different altogether.
One modification that should be made (if it hasn't already) is support for not starting the entire trial over when it makes a mistake. This could be considered part of the current 'framework' if you allow for input generator to remember something of previous states, but it's probably useful for it (partial restarts) to be a separate feature.
Question: Link can only turn to follow a grid with intersections at every 8(?) pixels, so is there ever any benefit to turning when not at a grid intersection? If not, or if usually not, then that information should definitely be taken advantage of to reduce the number of stupid moves considered.
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
NintenDan wrote:
Those videos on the TSC would be MY videos.
(They're not quite all yours, right?) Anyway, I knew that, in fact that's why I asked in this topic you made - it seemed that if anyone would know about them, you would. I wasn't sure if it was an oversight on my part that I couldn't access them, or if nobody noticed they weren't up, or something like that. Apparently that wasn't the case, so... never mind, it can wait. (EDIT: Some of them are up now, finally.)
Emulator Coder, Experienced Forum User, Published Author, Skilled player
(1301)
Joined: 12/21/2004
Posts: 2687
flagitious: It occurred to me that it might help to isolate what the emulator problem is if you post or send an .m64 file (of the first level or whatever you want, or even just getting into any level, if the menu manipulation is frame-perfect). That way I could compare the start of movie with what happens when I try to record the same thing, to see if the problem is there. The rest, if anything, would just be to determine if movies you make desync for other people. (I've been too busy to look at this or basically anything for the past few days and probably will be for a while longer... just letting you know it might help sometime.)
