Yes, you're absolutely right, but then it ceases to be anything resembling a TAS as we know it.
And another problem with that is that it's physically impossible under any theory of time travel (that I know of). You couldn't possibly know what had happened during a timeline that never existed, and even if you did, it would mean you'd have actually experienced it. That would mean you'd have to be physically altered during every save state so as to retain the information from previous save states. Doing this too often would result in aging (part of your life will be "lost" to everything else in the universe except you, so you'd suddenly jump to a state where you're older).
Of course, you could just say that the time traveler is separate from the timeline itself and can freely go to whenever he pleases. That's not what happens to characters in a TAS, however.
Derakon wrote:
We can assume that's one of the tools used to make this a valid thought experiment.
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
The prospect of using TAS tools "in real life" doesn't really make sense. That would imply that the character in the game itself (the character being you, and the game being life) would have control.
In a TAS, the controlling entity is completely separate from the game itself, and the characters in the game are completely oblivious to him or her. Even though the characters operate based on input, and it's possible for them to break the fourth wall, it's impossible for them to discern between normal playing and TASing.
So for this to work even as a thought experiment, you'd have to be god, and, at the same time, you would have to exist as a character in the game who is oblivious to god. Since you cannot be two entities at the same time, the premise itself is impossible.
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
Sounds very good :) Keep us posted!
Too bad it doesn't work in PSXjin yet, hopefully it will be fixed by the time you start working on your final run. Although it doesn't matter that much, PCSX-rr is still an accepted emulator of course.
PSXjin is currently actively being worked on, so maybe you could make a bug report detailing some of the things that could be improved. I'm sure they'll get to it eventually.
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
BadPotato wrote:
Dada wrote:
So, it's not a simple matter of giving everything a 4:3 aspect ratio. It would be nice if the encoding gods could give their thoughts on this.
There some thoughts about this, in this thread.
Also humbly requesting for the pjm file, at least once in a while.
Question to DarkKobold: are you using the NTSC or PAL version?
Request: can anyone make a screenshot of the first dialog box in the game ("How are you feeling?") as it appears on an actual PSX connected to a TV, and also one of the menu? I want to rule out the possibility of my capturing software (my PSX is connected to my computer) improperly resizing the image. Even just a photo taken from the front of the TV would be great, since I want to know what the aspect ratio roughly is. Preferably an old CRT TV since I don't know if 16:9 LCD TVs would make a difference in this regard.
I don't know if we can figure out a nice, scientific way to encode both these resolutions (since aside from being a different size, the menu also appears to be slightly differently positioned; at least on my capturing software). I think we might be able to resize the regular gameplay to 1280x896 using nearest neighbor (x4 original), then crop the menu to 357x216 (take off 11 px from the left side), upscale it to 1428x864 using nearest neighbor (x4 original), then downscale it to 1261x864 using regular bicubic resampling, and then letterbox it to 1280x896, positioning it in the top left with a black background.
The menu is then positioned at roughly the same place as in the screenshots I took from a real PSX using capture hardware. The resulting movie can then simply be given the correct aspect ratio in its meta information. I don't know if there's a better way to determine these sizes, but it's a start. I'll give it a try later.
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
I was thinking just now about how we should go about encoding this movie.
The problem: FFVIII switches resolutions a lot. Normal gameplay is 320x224, whereas any full-screen menu is 368x216.
I've done some tests with a real PSX hooked up to my PC via composite cable (sorry, but my S-video is broken):
Gameplay on a genuine PSX.
Menu on a genuine PSX. (Both images have overscan cropping turned off.)
Gameplay in the emulator. Sorry, I only have a ripped ISO of the French version available on my computer right now, but since they're both PAL I assume this isn't a problem.
Menu in the emulator.
I loaded all four images in Photoshop and resized the emulator images until they matched the "genuine" screenshots. What I found is, unsurprisingly, that they don't have square pixels.
The regular gameplay picture had to be resized to about 216.79% in width and 200.42% in height. The menu went to 190.34% in width and 200.42% in height. Obviously, these are only approximations. Nothing scientific here. And of course the sizes of my PSX screenshots are arbitrary. It's only the aspect ratio that matters.
Here's how the emulated screenshots stack up when resized to suit the genuine screenshot exactly:
So, it's not a simple matter of giving everything a 4:3 aspect ratio. It would be nice if the encoding gods could give their thoughts on this.
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
When pirohiko released his FFIII run, he included comments in the movie file. I extracted them from the FM2 and converted them to subtitles, then added that to the movie on Youtube. The result is very nice.
So if DarkKobold doesn't write subtitles for his run, I think I might as a test. :)
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
zeromus wrote:
Loading an archived iso is analogous to loading archived roms in other emulators and is not in principle objectionable, however it is hard to believe that anyone uses it given the substantial decompression time and RAM requirements, so I recommend against holding your breath while waiting for it.
Yeah, this isn't like a 1 MB SMC file. These are gigantic, full-CD ISOs. You could always simply use Windows' NTFS compression attribute if you want to cut down on filesize, if you're okay with the subsequent performance hit.
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
I loved your WIP, you really ought to continue working on this. :) Seeing you drive at top speed and practically launching yourself out of your car upon exiting is hilarious. And it's pretty hard to do in real time.
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
Very nice so far!
One thing that I was thinking about: how difficult would it be to manipulate Selphie's The End limit break? Even if it takes, say, 10 seconds of waiting to manipulate, it would instantly end even lengthy boss battles and be worth it.
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
DarkKobold wrote:
Dada wrote:
Compare the sound Seifer's sword makes in your video to the real thing, for example.
Oh wow. That is pretty bad. At least the music works in this version. The previous SPU had horrible music issues.
To be honest, I find Seifer's new weapon sound effect to be downright hilarious.
I don't know exactly how the PSX's sound system works but I suppose there's something similar to the NES's "noise" channel that isn't being emulated properly. That's not impossible to fix in the future considering the rest of the sound seems to be quite good. The music sounds pretty good (as far as I can tell in this low-quality Youtube video; the reverb might be a little off).
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
DarkKobold wrote:
Someone did make a new SPU.... PSXjin?
Aren't these latest videos you posted recorded using PSXjin?
The SPU still needs some work, though. It's not perfect yet. The sounds that enemies make when they die are especially different.
Compare the sound Seifer's sword makes in your video to the real thing, for example.
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
Shadow Byrn wrote:
Patashu wrote:
I don't think there are any precedents for speedruns that stop and wait for a separate game, actually.
Exactly. In the end, though it is totally up to the opinion of maker of the run whether or not to use it.
Not exactly. When using data from "outside" of the game (i.e. not obtained through input from a power on with clean RAM) you really need to check with the mods/judges to ensure that it is legal. The only way to get the data from that Chocobo game back into FFVIII is by loading it. The data will have changed as a result of something other than user input. In a way, if I understand this correctly, it's really no different from starting from a save game.
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
DarkKobold wrote:
This run is turning into more of a test run.
That's probably not a bad thing. I can guarantee you you'll have some regrets by the time you reach the end. And hey, maybe when you start working on the final version someone will have made a better SPU. :)
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
Shadow Byrn wrote:
So after doing some quick research into how Chocobo World works on the PS, depending on how the emulator handles the import of the Chocobo, it could turn into a very interesting discussion.
Essentially the Chocobo World is a memory card with a display on it. So at the first available point to send the little chocobo to Chocobo World and back again 100% loaded would take very little time in game time. A speed-runner could very easily get to that point, save, and then spend the next few weeks/months playing Chocobo's world to get a fully loaded chocobo, come back to the game and suddenly have all kinds of really cool stuff and have spend almost zero time in-game to get it.
That's a fun thought, but I think it's really no different from starting with a save state or with preloaded memory instead of a hard reset. In that you're allowing something other than controller input to influence the game.
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
Warp wrote:
But the major question is whether gravity is pointing up or down at point A (from A's perspective). It's not at all clear to me which way it would (and whether changing the dimensions or the mass of the torus would invert the direction).
Like Scepheo mentioned, A's gravity would keep him grounded. The hollow center of the torus does not have any mass, so it is not part of the gravity well. Thus, gravity points towards the ground at A's position.
The other side of the torus (the part which is "above" A), however, does have mass. Since it's further away, it does not affect a person standing at point A all that much, but it does mean he would be able to jump higher (this phenomenon is exacerbated the thicker the torus becomes).
It would be pretty hard to calculate the exact direction of gravity for point B. That depends on a number of variables (such as the mass of the torus and the exact size of the minor radius). Someone who actually knows physics should answer that one.
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
Warp wrote:
Dada wrote:
It also depends on how large the planet is. The smaller the torus, the stronger the relative difference between the inside and the outside of the torus would be. An extremely large and thin torus would behave more like a rectangle, whereas you would be nearly weightless in the inside of a small and thick torus.
These things sometimes work in surprising ways. For example, one could easily think that the gravity outside of a sphere (ie. at some distance x from the center of the sphere) might depend on the radius of the sphere. After all, if the radius is larger (and hence the surface of the sphere closer to the point where we are measuring) but with the same mass, the mass of the sphere will be more spread to the sides, spreading the pull of gravity to a larger area. Likewise if the sphere is smaller (but with the same mass), the mass is concentrated on a smaller space and hence the gravity (when measured from distance x) is likewise more concentrated on one direction.
However, when one does the math, it results that the radius of the sphere has no effect on the strength of gravity as measured from distance x (as long as the radius is smaller or equal to x). It doesn't matter if the radius is x, x/1000, or even 0, if the mass of the sphere is the same, the force of gravity as measured from x will be identical.
That's true, but when we stand atop the surface of a sphere, our distance to the center depends on the radius. That's (partly) why there are minuscule gravity differences at various altitudes on Earth. If the Earth were a lot smaller, but had the same mass as it has now, gravity wouldn't be stronger but it would affect us much more because of how much closer to the center we would be. But you are right that the net force itself would be the same.
This is why I mentioned the difference between a small and thick torus and a large and thin torus. A person at point A should be able to jump a lot higher than a person at the point C I mentioned earlier if the torus is small and thick because he is relatively quite close to the other side of the torus, which has its own gravity pointing towards its surface. In the case of a large and thin torus, this effect wouldn't be as strong because one is further away from the other side.
Joined: 11/22/2004
Posts: 1468
Location: Rotterdam, The Netherlands
Warp wrote:
This got me thinking: If by some magic a planet was indeed torus-shaped, how would gravity work on its surface? More specifically, where would gravity point to at different points on the surface of this planet?
Consider this cross-section of a torus:
What would be the direction of gravity at points A and B, for example? (If this direction depends on the major and minor radii of the torus, take them into account in the answer.)
I am not capable by far to give a proper answer to this question but it would appear to me that gravity should be less strong in the "inside" of the torus (point A). On the "outside" of the torus (let's imagine a point C existing on 270 degrees of the circle where point B is at 180 degrees; to the left, that is), gravity points exclusively inwards, whereas at point A gravity would be affected by the portion of the torus that is "above" A.
It also depends on how large the planet is. The smaller the torus, the stronger the relative difference between the inside and the outside of the torus would be. An extremely large and thin torus would behave more like a rectangle, whereas you would be nearly weightless in the inside of a small and thick torus.
At least, that's how I envision it, and I don't know if any of it is correct. I'm pretty sure there was something interesting going on outside when the teacher was explaining about gravity.
Gameplay on a genuine PSX.
Menu on a genuine PSX. (Both images have overscan cropping turned off.)
Gameplay in the emulator. Sorry, I only have a ripped ISO of the French version available on my computer right now, but since they're both PAL I assume this isn't a problem.
Menu in the emulator.
I loaded all four images in Photoshop and resized the emulator images until they matched the "genuine" screenshots. What I found is, unsurprisingly, that they don't have square pixels.
The regular gameplay picture had to be resized to about 216.79% in width and 200.42% in height. The menu went to 190.34% in width and 200.42% in height. Obviously, these are only approximations. Nothing scientific here. And of course the sizes of my PSX screenshots are arbitrary. It's only the aspect ratio that matters.
Here's how the emulated screenshots stack up when resized to suit the genuine screenshot exactly:
So, it's not a simple matter of giving everything a 4:3 aspect ratio. It would be nice if the encoding gods could give their thoughts on this.
