/vr/ - Retro Games

>>4573681
does this mean kirby superstar co-op is going to be even better?

>>4573683

Try it. Who knows. It's experimental as hell. Hasn't caused any issues I've seen so far.

That's pretty cool, although it would've been nice if hacking roms to improve on real hardware became a thing after that great SGnG hack

>>4573687

Early Konami SNES games were notorious for their slowdown,and they just weren't programmed well. The issues with SGnG are likely the same issues with Gradius III, Super Castlevania. Someone just has to spend the time to go through it. Additionally, if you use a romhack to speedup the game, and OC the SNES, that should produce perfection. Even with OC, the bubble section in Gradius III still has some slowdown.

This post goes through why SGnG is slow:

https://desuarchive.org/vr/thread/4554504/#4554512

>SNES9x aims to have compatibility with all games and avoid using game-specific speed hacks. If you want to use a emulator with a different focus, then please consider using a different emulator such as UOSNES.

http://www.snes9x.com/phpBB3/viewtopic.php?f=6&t=5728

>>4573681
wow this was immediately noticeable in mmx1
no slowdown when the giant bee enemy dies

This is pretty amazing. It's like Gradius III is an actual video game or something now.
Still has a slight bit of slowdown here and there, mainly against those laserspam guys in the boss rush, but it's barely noticeable.

>>4573681
Is it possible to play NES games in any of the emulators with no slowdown?

Would this work with SuperFX?

>>4573942
Yes.

http://emulation.gametechwiki.com/index.php/Overclocking#NES

>>4573681
I'm dying to play Gradius III this way, one of the fondest memories I have of a made for console shooter (it's so different from the arcade I might as well consider it that), and that's with the atrocious slowdown.

>libretro

>libretro

And zero fucks are given. libretro cores induce shitty levels of input lag. Overclocking the core will make fuck all difference for shoot-'em-ups and fightan games that rely on precision input timing.

I thought you couldn't overclock SNES games for some reason, like I was under the impression you couldnt just speed up the CPU like you do with a genesis cause it causes all sorts of glitches

>>4574419
You don't know what you're talking about.

>>4574493

Shut up and learn up
https://forums.libretro.com/t/an-input-lag-investigation/4407

>>4574497
Next-frame response times have been achieved too. There's nothing inherent about retroarch that causes lag.
https://forums.libretro.com/t/an-input-lag-investigation/4407/534

So besides SGnG and Gradius III, what else was unplayably slow on SNES?

>>4573942

Yes, but some NES games depend on the CPU being a certain clock and will not run OC'd. Experiment, anon. If a game won't load that used to looad before, turn OC off.

>>4574820
Super R-Type is not only significantly harder than the first game, but has awful slowdown to boot

a true disgrace

>TFW actually liked when games got so intense that they lagged
Am I the only one?

>>4574843
yeah it seemed to make them easier i found, like I thought NES lag was a great advantage gave you sometime to slow down and think more

>>4574820
the snes port of Out Of This World had insane lag

>>4574843
It's cool, but Gradius III is the kind of game that was in slowdown 90% of the time. Once you're fully powered up the game just stops moving.

>>4574419
They don't. Are you fucking retarded?

>>4574493
This. If anything it probably has the least lag possible in an emulator when you factor in options like frame delay or hard gpu sync

>>4574843
You may like non-retro Cave games since they feature a dynamic slowdown feature depending on all the stuff on the screen. What I mean it's that afaik they chose to put that it, it's not born out of limitations or bad programming, and in fact has made those games harder to emulate.
>>4574835
Play the arcade version (R-Type II), then. Why are you comparing a semi-port to the first arcade?

https://www.patreon.com/posts/beetle-psx-cpu-16891128

>>4575286
I think that's the anon that hates Retroarch for some autistic reason. Don't pay attention to him.

>>4575458
>What I mean it's that afaik they chose to put that it, it's not born out of limitations or bad programming
hahaha i have bad news for you, all the sh-3 slowdown is the result of peaking that shit hardware and is the actual reason it's so hard to emulate or port accurately, furthermore all the sh-3 games have such poorly implemented and lazy design that a majority of deaths will come from the game slowing to 10 fpbs then rocketing to 60 in a second fucking your judgement, this is something all high level players have mentioned. This isn't all though, they pushed that shit hardware so far in a few games especially dfk you can make the fucking sprites disappear as that shit overloads, cave were completely shit during the sh-3 period they had no ideas left and no thoughtful design, a good dev will stick within limitations and make great games not push it beyond capacity to the point of poor function.

>>4575458
Super R-Type has (shitty) brand new levels. But yeah I agree R-type 2 is better.

>>4574481
Pretty much this. Both CPU and PPU run off the same clock so overclocking it will severely fuck up the timing of the video signal.

This sort of thing is fun for five minutes or until you run into one of the many games that gets completely broken by it.

>>4575298
true, BUT ALSO, afaik the retroarch versions of bsnes, snes9x, nestopia, and fceumm actually have slightly less input delay than the regular versions, so if you combine that with hard gpu sync on and some frame delay it's 1 frame slower than real hardware on CRT if that.

ITT: retroarch/libretro ""devs"" desperately trying to stay relevant so they can shill their patreon

>>4573681
finally! Faster load times :D

>>4575832
Well, how many games truly need it? It works perfectly with the worst offenders, at least.

>>4575837

The code is open source, and there's nothing stopping these features being added to the original or even new emulators.

>>4575928
See >>4573696. Shit like this is why the whole dev community hates you.

This is a miniscule change at best with no real substance that any code monkey could make in minutes.

>>4576047

>Completely optional features are bad

>>4574082
Snes9x already has SuperFX overclocking.

>>4574481
>I thought you couldn't overclock SNES games for some reason
That's because unlike the Genesis overclocking a real SNES is completely unviable, even mild ones just break everything
There's that one video of a guy running pilotwings with the CPU @3.88Mhz (stock is 3.58Mhz) and the game just shits random sprite all over the screen randomly, and that's a very mild increase,
If you come close to what that hack does on hardware (200% clockspeed) your SNES isn't gonna boot shit and you might even damage it without appropriate cooling and even that big of an increase isn't enough to erase all slowdown in the worst cases

>>4576139

How does this work then?

>>4573681
>Previously OC the main SNES CPU was only available in MAME
Believe it or not there is a command line in zsnes that already did it in the 90s. You can chose anywhere between -50% to +150%, enough to make a difference.
>>4574419
>libretro cores induce shitty levels of input lag
wat
>>4574497
Standalone emulators without hard gpu sync are ten times worse.
>>4574843
Yeah it never bothered me that much. Although some examples are nuts like double dragon 2 (arcade) or metal slug 2. Overclocking in those games is almost mandatory.
>>4575818
>they had no ideas left
Hence the waifus

>>4575818
Thank you for this information. It seems I was misinformed; I was never too much into post-PGS Cave games so I didn't know all the details, but I had heard what I mentioned. I guess it was a bad source.

>>4576141
I'm not 100% but I'm pretty sure the reason is that on hardware you can't fully control how each chip will sync (not to mention the way you overclock a SNES is rather with each other if you start running some of them at non-standard clockspeed, on an emulator you can have complete control over that.

Also here's how you do it on real hardware.
http://web.archive.org/web/20070519170453/http://www.undergroundcm.com:80/index.php?option=com_content&task=view&id=15&Itemid=38

>>4576154
meant to say PGM

>Business Entity Detail
>000942380: Limited Liability Company - Domestic
>Name: Libretro LLC
https://tnbear.tn.gov/Ecommerce/FilingDetail.aspx?CN=129081049022108006163108130004225038122089084111

>>4576047
>Shit like this is why the whole dev community hates you.

> Being rejected by 'people' like this
> such peer pressure

>>4576164
This. Who gives a fuck about being rejected by a bunch of faggots in the gaming community? If you seriously care about such things you are a human embarrassment.

>>4576164
That's the new soy flavored Cola

>>4576047
>This is a miniscule change at best with no real substance that any code monkey could make in minutes.

Then don't use it? No one's forcing you. And it's not in any way against greater game compabitility, so you lose nothing. It's not like development of SNES9x has to stop now.

>>4573696
>game-specific speed hacks
Well funny enough, they managed to do just that it seems. A general purpose overclock that works in several games.
It is only game specific if you want its use limited to, and automatically enabled on, specific games.

Well technically, it's quite a bit better than overclocking for compatibility. Seems it's actually just cycle delay reduction.
Kind of like beetle-psx's GTE speedhack, which just makes GTE actions take less time and have zero emulated latency to/from the CPU.
https://github.com/libretro/snes9x/commit/8b180496ac64e206da1b09ede2617f634788ff59

>>4576659
>Seems it's actually just cycle delay reduction.

What does that mean?

>>4576659
>Kind of like beetle-psx's GTE speedhack,

It has a proper CPU OC now.
https://www.patreon.com/posts/beetle-psx-cpu-16891128

>>4576665
>What does that mean?
It reduces how many cycles actions take to complete, in emulated context. Something that would take 8 cycles would only take 4.
Kind of like increasing the IPC of a processor without changing its clock.
Super optimizing the capabilities of the emulated processor without actually changing its clock. Just by fucking with timings.
>It has a proper CPU OC now
I know. Just saying that this is not one. The old GTE hack is closer in function to this, sort of.

>>4575837
>>4576047

Why so salty though?

>>4576157
pgm was caves peak for sure, sh-3 has some ok games but just mostly held back by such shoddy hardware and poor use of said hardware, ibaras probably the only game thats actually done really well with little by way of slowdown nonsense. Your source just had everything the wrong way round.

>>4576962
I like their older games a lot as well (I'm a Toaplan fag after all), but it would be foolish of me to not respect DOJ and Ketsui, I've had my fun with them (though I've only managed a 1-ALL on both). Dangun Feveron and DonPachi are a bit more up my alley, with DDP as a nice bridge between these. I like the rest of their retro games, though I've barely played Progear.