Removing the slowdown in Gradius III SNES

[Despatche]

Noone reads Hi Scores so let's make a thread here.

Man, that MAME thread really pisses me off. The kind of mindset MAMEDEV has is the kind of mindset that genuinely holds this hobby back in ways that people take for granted. Fuck that, right? The good news is that the SNES scene agrees.

Someone finally did what should have been done decades ago. Everyone hates the slowdown in SNES Gradius III, right? In development is a hack that adds a simulated SA1 chip to an American Gradius III ROM. This removes pretty much all slowdown from the game. It reveals a game that might actually be harder than the notorious original arcade game, as things should be.

Simply letting the attract demo play reveals differences. You can watch that here. The link to the patch is in the description. Just apply it to an American Gradius III ROM like any other patch. It's BPS instead of IPS, so you'll need something like Floating IPS (which you should have anyway). The ROM you get out of it should run on anything supporting the SA1 (that means anything that supports Super Mario RPG), so hit up the latest Snes9X or the new bsnes or something. Only issue so far is that the game will crash after the game over jingle ends, but that will likely be fixed soon.

Please check this out. It's time to play Gradius III SNES as it was meant to be.

Oh wait apparently it also crashes after clearing a loop at the moment. Good luck getting there though!

[BareKnuckleRoo]

Neat! Shame about the crashing after gameover jingle but it's definitely promising to see there's an option to play the game without the massive, nearly constant slowdown present in the game.

For people who missed it, there's also Metal Slug 2 Turbo and Guardian Force De-Lag patches which exist. Metal Slug 2 has a similar issue at release with way too much slowdown due to a bug.

[EmperorIng]

Does the patch only work with the USA SNES rom? I haven't tried yet.

All in all, I was really excited when I heard this. This is the type of thing that Gradius III's port needs to elevate itself - you have the eminent playability of Gradius III without the consistent crawl that creeps in in the games' mid-to-late stages.

[mycophobia]

the slowdown in gradius iii snes is low on my list of problems with that game...but it's cool that this exists

[Square_Air]

The slowdown in every version of Gradius III has turned me off a bit, so this is relevant to my interests.

There have been MS2 speed fixes around since the early 2000's, and I've never tried Trap's fix, but I highly recommend checking it out. MS2 is possibly the best example of why arcade accurate emulation shouldn't be a universal gold standard, since the original MS2 borders on being completely broken.

[Despatche]

Does the patch only work with the USA SNES rom? I haven't tried yet.

At the moment, yes. I should have mentioned that, sorry.

[MathU]

I have been waiting for someone to make a ROMhack like this of Gradius III SNES for a very long time. In the past I have actually considered picking up hacking it as a hobby when I ever managed to find enough time. So happy that I wasn't the only one. The tweaks and improvements made by the Gradius III team to their SNES port deserved to be freed from the shackles of its crippling slowdown.

[MarioNintendo]

Good stuff!

[soprano1]

Looks great, thanks for sharing.

[opt2not]

Hopefully the SA-1 support on the SD2SNES will be at 100% soon. I'd hate to sack a working SA-1 cart to convert this (I also try to avoid emulating if I can).

[komatik]

...why do you need to fuck around with simulating another chip or patching roms when you can just raise the emulation execution cap? Snes9X already lets you 'overclock' up to 500% which eliminates any slowdown in any game I've ever tried, including Gradius III (and I just tested it again now to confirm I'm not nuts).

I don't understand how this is different or what it's supposed to accomplish.

[pegboy]

...why do you need to fuck around with simulating another chip or patching roms when you can just raise the emulation execution cap? Snes9X already lets you 'overclock' up to 500% which eliminates any slowdown in any game I've ever tried, including Gradius III (and I just tested it again now to confirm I'm not nuts).

I don't understand how this is different or what it's supposed to accomplish.

Because this would let you play the game on a real SNES.

[Despatche]

More than that, it's also a small test for what the SA1 can do. Unlike the NES, most of the SNES chips were poorly utilized.

[SuperDeadite]

As a Gradius III PCB owner, I can say that it gets mad slowdown too...

[MathU]

I think you would need to physically change some bosses and level design to make the arcade version play well without its slowdown.

[Despatche]

The PS2/PSP version seems to disagree.

[Tatsuya79]

Even without slowdown it's not as hard as the arcade version which has some crazy moments that were removed/altered on the console port (flying ice blocks, craziness of the fire stage and more I forgot).
You could remove the slowdown with the overclock option in snes9x set to "max" in the libretro core, now in stand-alone since 1.60.

[Shepardus]

The PS2/PSP version seems to disagree.

It's surely playable with wait turned to 0, as pegboy demonstrated, but certain parts (most notably stage 7, the fire stage) get really ridiculous and practically mandate alternative strategies to get through.

It reveals a game that might actually be harder than the notorious original arcade game, as things should be.

I'm not sure why that's "as things should be," but pretty cool if this works on the actual hardware. I didn't catch that from the original post and was wondering the same thing as komatik.

[opt2not]

Hopefully the SA-1 support on the SD2SNES will be at 100% soon. I'd hate to sack a working SA-1 cart to convert this (I also try to avoid emulating if I can).

Self quote. But I just tried this out on my SD2SNES with the latest firmware 1.10.3, and it works! Man, this is a game changer, it runs so smooth on real hardware!

[MathU]

It's surely playable with wait turned to 0, as pegboy demonstrated, but certain parts (most notably stage 7, the fire stage) get really ridiculous and practically mandate alternative strategies to get through.

My strategy has always been to let an option hunter steal my options for the fire stage and fly through it barely shooting. Once you learn how to do that you can effectively recover anywhere on that stage. I'm thinking more about a few bosses whose strategies boil down to "memorize this very specific strategy to win, otherwise fail" when slowdown is removed. While the whole game is surely full of this stuff anyway, it just makes it even more unwelcoming to casual replays.

[Despatche]

Just got an update! You want "patch.bps". Seems game overs and loop transitions now work. Haven't tested it yet.

Something interesting I haven't mentioned yet: the patch throws in SRAM support, so the game saves whatever your scoreboard looks like. Of course, this game doesn't split for difficulty or anything, but it's still nice.
----------

Even without slowdown it's not as hard as the arcade version which has some crazy moments that were removed/altered on the console port (flying ice blocks, craziness of the fire stage and more I forgot).
You could remove the slowdown with the overclock option in snes9x set to "max" in the libretro core, now in stand-alone since 1.60.

The cube stage is kind of a gimmick. Most of the difficulty is in the weird pixel perfect/randomness aspect, not so much actually trying to figure the cubes out. Top players routinely get bodied by a cube not behaving like it should. This is because the developers intended for players to dodge all of the cubes raw, and they tested the stage to make sure that was possible. Good luck convincing anyone to actually do that.

Yes you could, and still can! However, this hack is significantly better because it doesn't rely on an emulator-specific feature and because it furthers SA1 research. You can't just overclock a real SNES like it's no big deal. In fact, overclocking like that is a huge waste compared to just offloading to another chip. That's why devs had to be really careful with action games on the SNES, and most of them didn't use special chips like the NES did.

It's surely playable with wait turned to 0, as pegboy demonstrated, but certain parts (most notably stage 7, the fire stage) get really ridiculous and practically mandate alternative strategies to get through.

It's strange that it seems as if "alternative strategies" are required, except... I think it's being caused by something noone's been talking about but is very relevant with the Gradius III interview from the Portable guide: the devs were playing a somewhat different game than the players were. Combine this with being able to choose between multiple slowdown settings in the port, something basically no port does. Removing the slowdown from SNES III is "how it should be" because this version of the game has spent almost its entire life being seen as deficient. Now that there's a practical way to actually remove the slowdown from the game (using an emulator feature to overclock virtual parts is not as "practical" as people think it is), we can compare this to the arcade version with various slowdown settings and get a better picture of the game's actual design.

I know this sounds like I'm about to go into "Gradius III is for galaxy brain 300 IQ people" territory, but I'm not. If anything, I'm saying that Gradius III is a pretty normal example of escalating design for escalating developers and players, and neither group expected what was going to happen. Fortunately, the arcade Gradius III was seen as legendary instead of being maligned, because Japan has (or at least had) a better appreciation for this sort of thing. Westerners, on the other hand, dismissed Gaplus almost immediately, and that game gets nowhere near Gradius III.

By the way, this interview also disproves the common myth that Gradius III SNES was supposed to be a "better" version of the game. Other than bug fixes the developers wanted to do, the SNES game is really more of an arrange version. Just like any other arrange version, some might like one and some might like the other.

[Shepardus]

It's surely playable with wait turned to 0, as pegboy demonstrated, but certain parts (most notably stage 7, the fire stage) get really ridiculous and practically mandate alternative strategies to get through.

It's strange that it seems as if "alternative strategies" are required, except... I think it's being caused by something noone's been talking about but is very relevant with the Gradius III interview from the Portable guide: the devs were playing a somewhat different game than the players were.

Are you referring to a particular part of the interview or the whole thing? I'm not really catching your drift here.

Removing the slowdown from SNES III is "how it should be" because this version of the game has spent almost its entire life being seen as deficient. Now that there's a practical way to actually remove the slowdown from the game (using an emulator feature to overclock virtual parts is not as "practical" as people think it is), we can compare this to the arcade version with various slowdown settings and get a better picture of the game's actual design.

Oh, by "how it should be" I thought you were saying that SNES Gradius III "should be" harder than arcade Gradius III. But if you're referring to removing the slowdown then I have no objections.

If anything, I'm saying that Gradius III is a pretty normal example of escalating design for escalating developers and players, and neither group expected what was going to happen.

I think I've said this before but to me arcade Gradius III is the most "Gradius" of the Gradius games, realizing the vision suggested in Gradius 1 and more fully fleshed out in Gradius II. Then between that and Gaiden they experimented with a different path in the Parodius series, which also turned out very nicely.

Confession time: I realize I'm somewhat derailing this thread and that's probably because I've never actually played SNES Gradius III...

[Tatsuya79]

I feel Gradius III snes on easy without slowdown is pretty accessible.
Like a decent arranged version kind of "fixed" for a wider public.

[DJ Incompetent]

How bad does the high speed stage get?

[FRO]

How bad does the high speed stage get?

This may be just my perception, but as someone who has been playing the SNES version the last week or so, and having reached Stage 7 a few times, it's weird that it doesn't feel very slow, compared to the rest of the game. Maybe that's because it's just a handful of turrets to shoot at, no bullet spam, and very few moving sprites to cause slowdown, but this level feels sufficiently fast on SNES to me.

[pegboy]

Damn I need to try this, need to update my SD2SNES.

[opt2not]

I feel that this conversion makes the game, A. more difficult, especially if you're familiar with when the slowdown happens and are used to exploiting it...you can no longer do that. And B. it feels like this is what the game developers intended.

[Despatche]

Maybe not 100% the developer's intent, but definitely closer than the game we originally got.

Game overs seem to work fine now, by the way. I saw someone get to the second loop too.

This should work fine in Snes9X and bsnes. Might work in SNESGT or uosnes too. Doesn't work in ZSNES because it doesn't have great SA1 support.

[Mortificator]

I feel Gradius III snes on easy without slowdown is pretty accessible.
Like a decent arranged version kind of "fixed" for a wider public.

In the same vein, I bump down the difficulty of the arcade version to easy (setting 3) when I play the PS2 port to compensate for turning off the slowdown.

[Mark_MSX]

This is the type of thing I've wanted to happen for a long time. I remember a year or so ago I was looking to find ways to overclock snes cores in retroarch to try and remove slowdown from the snes shmups. Huge props to the guys who did this!