So this game currently has no rules set, and there has been some discussion as to what this game's rules should include, as well as timing methods.
As far as timing goes, there are more nuances to this than there first seems. First of all, you need to determine where the start/end point of timing is, which sounds easy, but actually isn't very simple, for a few reasons.
First of all, I believe that the RTA should correlate to the Ingame timer as closely as reasonably possible, for consistency purposes. To this end, this would set the start/end point for RTA as the start/end point for IGT. For the end, this is easy enough; the ingame timer stops counting once you enter the Spirit Stone Chamber after defeating Evil Gaia, so this should presumably be RTA's endpoint. Currently, Elmagus is the only one with an RTA run, and he uses last hit on Evil Gaia as his endpoint, but he also wastes no time between there and entering the Spirit Stone Chamber, so that's easy to retime.
The problem is determining the start time. I've just spent some time running some various emu tests, and the ingame timer does NOT start counting when you select New Game, which is what people have been using as their start point for RTA.
When you select New Game, the game begins with an FMV cutscene that lasts for a couple of minutes, then the first ingame cutscene featuring General Baal begins. As far as I can tell, this is when the ingame timer starts counting. Not even specifically when the SFX of the cutscene starts, it appears that the cutscene actually begins several seconds before the sound starts playing, because that was the only way I could match up timing with Livesplit against what the ingame timer read when I reached the first save point in Parm Port.
So that sounds simple enough, but then I encountered another problem; emulation accuracy. I ran these tests on both Bizhawk and PSXfin, and got entirely different results. Turns out, PSXfin is a very inaccurate emulator. From selecting New Game, to gaining control in front of Gantz:
PSXfin: Gain control at 9:30, save point reads 7:16 Bizhawk: Gain control at 9:45, save point reads 7:32
PSXfin saves 15~ seconds over Bizhawk just in the game intro purely because it's not emulating the load times correctly. What's more, the ingame timer reads an entirely different value! This would suggest that the ingame timer actually takes loading times into account, so shorter loads due to poor emulation = lower ingame time.
Bizhawk is definitely the accurate one; it gains control at 9:45, which is exactly the same as Elmagus' RTA run, which was performed on a PS1. If emulator runs are to be accepted for this game, Bizhawk is for sure the emulator we should be permitting, and only that, unless it can be demonstrated that another Playstation emulator has similar accuracy. Just from the intro, PSXfin saves 15 seconds due to shoddy emulation.
If we compare start points of 9:30 to 9:45, this implies that PSXfin consistently saves approximately 3% extra time over Bizhawk/original hardware.
To put this into perspective, Elmagus' RTA WR is 13:35:03. Under these parameters, it would have been approximately 13:10:36 on PSXfin. That's just ridiculous to allow.
tl;dr I would like to make RTA/IGT comparable but it's a little silly to work out and measure, and it brought to my attention that you should really run on original hardware if you can help it, and use Bizhawk if you absolutely must use an emulator.
I have never really heard of PSXfin. However, from what I've heard, ePSXe 1.9 was supposed to be decently accurate. What are your views on that emulator? It's also fairly widely used as far as I know (then again, my knowledge could be a bit outdated).
For starting time, as a runner of other games, I feel it's easiest to start timing on "New Game", personally, instead of hoping to accurately get the start on a specific in-game queue. Then again, that could just be me.
Jimmy33 was the first one to start routing the RTA run, and Elmagus used some of his notes as far as I can remember, which included the timing as well, so I think that's where the end point was taken from.
In regards to IGT vs RTA, I'm not sure what is best. Should PS1/PSX and PS2 with fast load speed be judged together or as separate categories?
As far as establishing a consistent start point for if you're trying to start RTA timing at the same time as IGT, I'm about to do some testing on Bizhawk. Gonna establish roughly how long the intro FMV cutscene is, then start splits with an Offset of that length, then change it incrementally and try the game intro a few times. I'm expecting to see something along the lines of starting with an offset of -2:35, but I'll see how it goes. I'll be streaming the process just now. Whatever findings I get, it would be really helpful if someone is able to replicate what I do on original hardware and confirm that the timings are the same, since I don't have any hardware set up right now to try it with.
I have updated the rules with the new RTA timing information (start game to entering Spirit Stone Chamber). Should we do some testing with ePSXe as well? I know ePSXe is generally allowed on SRL, but it wouldn't hurt to test different versions of it to see which ones we might allow.
There was a run on YT under 12 hours , made on real hardware, so if its true, 11:40 in segmented is possible on emulator. We should test more and i think there should be a category for emulators and for real hardwares(consoles)
I don't necessarily agree with a console/emulator split, since the idea behind this proposal is "Consoles and emulators are not equal", yet I've demonstrated that even different emulators aren't equal. Because of this, I'm inclined to lean towards having consoles and the most accurate emulator on the same board, and then just discourage usage of other emulators.
One could go further and argue that the consoles themselves should be standardised on this basis, as it has been shown that the various platforms this game can be ran on have significantly different loading times, but this is a separate issue to console/emulator as variable loads by console are just a natural consequence, whereas console/emulator is due to the emulator itself simply not emulating the game accurately, and this can be due to more things than just loading times; emulators can run at a higher framerate, not emulate lag correctly, etc.
tl;dr at this point it's an issue that is still just to be discussed, but I think making a console/emulator split on the boards would be premature at best.
So are EPSXE runs not allowed to be used for Grandia? but then debate about faster versions of the game or consoles thats best to run on? kinda confuses me a little. also, is their any version of EPSXE that is allowed like 1.6, 1.7 or whatever.
