[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/vr/ - Retro Games

View post   
File: 73 KB, 800x616, MiSTer.jpg [View same] [iqdb] [saucenao] [google]
8722687 No.8722687 [Reply] [Original]

How accurate is MiSTer really? For all the talk about how FPGA allows for transistor-to-transistor logic how many of the cores actually do this? The upcoming PS1 core is literally just the Duckstation emulator isn't it?

>> No.8722828

Dunno, feels pretty accurate to me though, certainly good enough to where I don't care. To me, the biggest appeal is that it's such a wonderful emulation device for the living room, whether you're using your big 4K TV or a CRT. It blows everything else I've used out of the water in terms of user experience, accuracy and just general ease of use. With all that said though, I don't think it has much practical use if you're intending to use it at your desk next to a computer. At that point, just use BSNES, Duckstation etc.

If you're interested I recommend just buying a pre-configured one this year before China does some funny shit and invades Taiwan and buttfucks Terasic. Shame about the current price of the thing too.

>> No.8722907

>>8722687
I think it’s really accurate. I think emulation is really accurate too. The MiSTer is not for connecting next to your computer, because there are limited benefits.

I wanted to play emulation games on my living room TV, and my options were a raspberry pi, building a new fairly-bulky PC, or a MiSTer. Because of the chip shortage, pis are de10-nanos cost about the same, and the Pi will be emulation on top of Linux so it’ll have some more input lag, so I went with the MiSTer.

For what it’s worth, I noticed the MiSTer has less input lag than my PC. Not enough to be worth the cost by itself, but it is definitely more responsive.

>> No.8722917

>>8722907
>building a new fairly-bulky PC
SFF exists you dumb liar

>> No.8722919

I can't weigh in on this, OP. I still have a full head of hair

>> No.8722925

>>8722917
Not him but even a SFF build is a pain in the arse to use on a TV in comparison to something like the Mister, Nvidia Shield, Raspberry Pi etc. PC to a TV is the autists answer to everything. Their living room always looks like a complete fucking mess of wires and stupid shit for using KB+M on your lap. Plus it's so unintuitive.

>> No.8722959

>>8722828
I think the main issue is that if the cores aren't actually transistor accurate then that effectively eliminates the main benefit of using a comparatively more expensive FPGA in the first place as opposed to a cheaper raspberry pi or something. There are still some fringe benefits like lightguns over SNAC but if the accuracy stops at "good enough" then it's not really doing anything that isn't already happening in software emulation.

>> No.8722973

>>8722959
Light guns don’t need SNAC, there’s a new implementation for GunCon 2 over USB with CRTs with all cores. There’s even GunCon 3 support for flatscreens.

Even without transistor accuracy, there’s still less input lag than Raspberry Pi, because there’s no software interface for interpreting USB, the only lag is USB polling frequency and core-defined polling frequency.

>> No.8723210
File: 42 KB, 878x493, Hidden_Mounting.jpg [View same] [iqdb] [saucenao] [google]
8723210

>>8722925
There's a wide variety of mini PCs that are small and designed to mount on the back of a TV.

>> No.8723245

>>8722925
it's literally the same anon. power and display cable. if anything mister setups seem way more messier than a sff pc.

>> No.8723442

>>8722907
>Because of the chip shortage, pis are de10-nanos cost about the same
lolwut?

>> No.8723508

>>8723442
Maybe you haven’t checked but raspis are selling for a couple hundred dollars these days. unless you’re just making fun of that ESLpost.

>> No.8723534

>>8723508
I bought a 4gb 4B couple of weeks ago for 100 ausbucks, got it to play arcade games on my CRT
No noticeable lag honestly

>> No.8723546

>>8723534
What are you playing on it?

>> No.8723549

>>8723546
Arcade games?
Didn’t I just say?

>> No.8723553

>>8723549
There’s a lot of genres of arcade games spanning decades, anon, im just wondering what you’re playing.

>> No.8723559

>>8723553
Ah, at the moment just played Pulstar and a bit of SF3 and Final Fight with a friend

>> No.8723579

>>8723559
I’ll probably try some tHotD and vitrtua cop later

>> No.8723591

>>8723559
Nice anon, Pulstar is a great choice. You can play lightgun games on the Pi?

>> No.8723610

>>8723591
Yes
https://www.youtube.com/watch?v=V7fPdX8VxW0

>> No.8723617

>>8722687
It depends on the specific core. One thing that MiSTer does do better than software emulation ever could is audio-video synchronization, this is important, far more than emulation accuracy. Keeping audio and video in sync is literal hell in software and results in excessive audio latency.

>> No.8723621

>>8722973
>USB
Nobody who would ever consider using usb cares about input lag or emulation accuracy. USB is heresy its very existence makes me want to start beheading people.

>> No.8723629

>>8723621
SNAC is a meme and it’s nothing but a snake oil comfort blanket for schizoid autismos. it’s the retro gaming equivalent of thick diamond-coated audio cables.

>> No.8723648

>>8722687
>The upcoming PS1 core is literally just the Duckstation emulator isn't it?

No, FPGA is a completely different approach as it uses configurable logic gates to transmutate into the real hardware, as thus, it works as real hardware (as close as the configuration[core] is to real hardware).
This particular configuration is limited to 100k logic gates, so it won't be able to recreate more modern hardware, specially GPUs

>> No.8723853

>>8723508
Ah, I see that a lot of retailers have sold out of them. So I'm guessing those are secondhand prices?

>> No.8723867

>>8723648
Wrong.
https://youtu.be/EMZVp48SMps?t=266
Gate level accuracy isn't even attempted for the Playstation.

>> No.8723883

>>8723648
No he's right. Its just Duckstation converted to HDL. That's why the core progressed so quickly. That's why the "its just emulation" argument rages, since while it may be accurate as far as game compatibility goes, it's not an accurate recreation of the PS1 hardware.

>> No.8723967

>>8723867
That's why I said "as close as the configuration[core] is to real hardware". If the developer wants to cut corners and use the ARM to emulate features, well... It's not even an FPGA approach anymore.

>> No.8723969

>>8723967
That's not what's happening either. Like >>8723883 said it's still synthesising HDL to run on the FPGA, but the HDL is ported from software emulators as a best guess effort for what the actual chip does.

>> No.8723996

>>8723969
I understand your point, as I said, it depends on how close the core is to real hardware. In this particular case, this core is pure bullshit, it's indeed just emulation with extra steps. It defeats the purpose of FPGA hardware recreation.

>> No.8724029

>>8723996
Which is OP's point. How many cores are more accurate than software emulators? If its not replicating the transistor-to-transistor logic then its not taking advantage of FPGA's potential. And if most cores are just effectively MAME converted from C to HDL then its hard to argue against the people saying its just an expensive emulator.

>> No.8724063

You would need the schematics for the chip or a decap of it to find exactly how it works, and only then could a proper FPGA reconstruction be possible. For some things like the 6502 the entire chip is completely documented and we know what every transistor in it does.

>> No.8724074

>>8723648
>This particular configuration is limited to 100k logic gates, so it won't be able to recreate more modern hardware
That FPGA can't even recreate a 286 CPU (that had 134,000 transistors)

>> No.8724084

>>8724063
but it's more than that, for some things such as Commodore ICs the chips relied a lot on quirks in the process used that can't really be recreated with more modern stuff. The 6502 for instance had those infamous illegal opcodes, but only on NMOS variants--newer CMOS ones didn't have them.

>> No.8724095

The SNES core is able to render Air Strike Patrol correctly?

>> No.8724113

>>8724084
Can you elaborate on this? Why can't those quirks be recreated?

>>8724095
I think it does. This actually adds another layer, too. There are so many hardware variants with these consoles, some with very different performance. It'd be nice if MiSTer would implement the different revisions and let you pick.

>> No.8724123

>>8724074
Transistors are not the same thing as logic gates.

>> No.8724178
File: 115 KB, 1280x803, file.png [View same] [iqdb] [saucenao] [google]
8724178

>>8724113
https://www.masswerk.at/nowgobang/2021/6502-illegal-opcodes

I'm kind of a retard when it comes to assembly-level stuff but I suspected it would be something along pic related's lines. There's some other stuff in here that I don't understand, but I'd think that the MiSTer could reproduce these sorts of effects.

>> No.8724191

>>8724113
You could probably imitate it it but at the cost of using many more gates than the original chip.
All else being the same, an NMOS logic gate will drive a signal high when a CMOS logic gate leaves it floating - and in the case of a conflict, NMOS logic will drive low much more strongly than high. That latter aspect is why illegal opcodes often perform an AND operation "for free".
Even though it's technically out of spec, the underlying technology usually does this reliably.

>> No.8724265

>>8724113
Commodore used an ancient 1970s fabrication process that was akin to bashing rocks with a mallet. It was very "dirty" compared to more modern processes and resulted in the chips having numerous exploitable bugs you could use. While an FPGA could recreate the basic functions of a VIC-II, it wouldn't necessarily reproduce the chip's glitchy behaviors that games and demos quite often make use of.

>> No.8724279

>>8724178
The 6502's instruction decoder has space for 256 opcodes, but about half are not used and so interesting things happen when you attempt to execute them. All NMOS-based 6502s have the illegal opcodes but they disappeared from CMOS ones.

>> No.8724293

>>8724191
Ah so hypothetically a "100% accurate" FPGA implementation would have to choose between being faithful to the actual circuit and accepting that there will be differences in the performance or be faithful to the performance at the cost of a less than authentic circuit.

>> No.8724375

>>8724293
It's behaviour rather than performance that is the issue. Functionality which is not part of the design can be lost when switching from NMOS to CMOS. To restore this accidental functionality, you'd need to add it to the design explicitly.

>> No.8724428

The better way to explain it is the stuff was closer to analog electronics than the clean architectures we have today.

>> No.8724443

>>8724074
Transistors are in pairs of three, therefore 134,000 transistors translates into 44,666 logic gates.

>> No.8724498

>>8724178
>>8724293
>>8724375

How does this affect software compatibility on real hardware?

>> No.8724502

>>8724443
>pairs of three
Something isn't adding up here.

>> No.8724513

>>8724443
something like that. if the FPGA has 100k logic gates that translates to 300k transistors, which would cover a 386 (275,000 transistors). However that's plenty for any 4th gen and earlier hardware which is anything that actually matters.

>> No.8724514

>>8724513
>However that's plenty for any 4th gen and earlier hardware which is anything that actually matters
This is the same selection of hardware which has really good software emulation.

>> No.8724529

>>8724498
I don't understand the question. Original software runs the way it's intended to on original hardware. What are you comparing it to?

>> No.8724537

>>8722687
Also if you do an FPGA and it outputs HDMI or something instead of the original composite signal, we're done. Get the fuck out.

>> No.8724552

>>8724513
man a whole lot of people's childhoods were PS1 games not Atari shit

>> No.8724554

Relevant

https://hackaday.com/2014/12/18/counting-transistors-in-the-playstation/

>> No.8724557

>>8724443
The 6502 has 4,528 transistors or 1,509 logic gates.

>> No.8724569

>>8724554
>300,000 transistors
So it's a bit more than the 386 but a lot less than a 486 which cracked the 1 million mark.

>> No.8724584

>>8724529
Is there software that rely on these instructions that do not work on revisions without it?

>> No.8724589

>>8724584
If you mean undocumented 6502 opcodes, yes C64 and Atari 8-bit stuff does use them sometimes. It's not as common on Apple II software, a few programs from the II/II+ era had them and those broke when the IIc switched to a CMOS 6502.

>> No.8724595

>>8724569
If so and it's 300k than an FPGA could just about pull off a PS1 CPU.

>> No.8724606

>>8724554
it mentions the NES CPU. for all practical purposes only the APU matters as far as FPGA recreation, the CPU portion of the chip is just a 6502 with the decimal mode disabled. the sprite DMA doesn't matter since it was bugged anyway and doesn't work so nothing actually used it.

>> No.8724617 [DELETED] 

>>8724606
>>8724554
The Russians already had plenty of experience with the NES chipset from Famiclones 25 years ago.

>> No.8724632

>>8724595
Well the DE10 in the MiSTer can just about pull it off. There are much better FPGAs but the price jump is dramatic. I'm assuming that over the next few years if costs start to come down then the project with transition to something more powerful and N64 and maybe Dreamcast will become viable.

>> No.8724639

>>8724584
Something similar happened with the Genesis. I thing Gargoyles doesn't work right on a Genesis 3 because it relies on behavior that changed with that revision. Though the Genesis 3 is more like an official clone than a true revision.

>> No.8724649
File: 2 KB, 288x175, kool-aid-man.png [View same] [iqdb] [saucenao] [google]
8724649

Kool Aid Man glitches on Atari 2600 Jrs as they have a newer revision TIA that breaks the game.

>> No.8724676

>>8724632
The jump is indeed very dramatic. I don't have a mister and only got a few glimpses here and there mostly due to youtubers praising it.
This thread made me less interested in it as work on the cores seem to be more in the way of making games work than actual hardware reproduction and preservation

>> No.8724718

>>8724676
It's frustrating because there are a lot of competing interests. On one hand, people just want the games to work and MiSTer is behind on that front simply because it's newer and the cores have to be developed. But that means that in the interest of speed they have to be built from existing emulators, which means the cores aren't really any better or more accurate than those emulators. Ideally it's just a stop gap so that the games can run while the core developers focus on accuracy in the background. A few cores have had major overhauls like this in the past and so long as people can actually play the games the work behind the scenes is more invisible and less time sensitive. The trouble though is that a ton of proprietary chips are total black boxes and decapping them costs thousands of dollars. MAME had done a lot of that work in the past but unfortunately the emulation community is very fractured so there isn't a coherent system in place to collaborate. It means there's a lot of reinventing of the wheel going on.

>> No.8724740

>>8724718
Yeah, the amount of work for accurate arcade PCB recreation is immense. Taito hardware alone would give tremendous work

>> No.8724782

>>8724718
Honestly, I get the point of accuracy autism from a preservation standpoint, but past a certain point it legitimately turns into the same kind of shit that makes people upload "rare" revisions of some shovelware Famicom game nobody gives a shit about and nobody has ever played. Sure, I guess it was important for the sake of completeness, but is anyone really going to play it?

All the consoles the MiSTer handles already have excellent software emulation. I don't give a shit if the PSX core is 90% duckstation. What do I care about? If I can get SNAC support for all my fucking PS1 controllers. Bonus points if it lets me feed in my 5V PSU and transform it into the 7.2V the console put out for rumble.

I like retroarch, it works fine, I just played AC1 and it was great. But if I wanted to play Air Combat instead? Well, I'd want to use my NeGcon, which means fighting with the fucking settings menu for an hour and wiping my settings for my regular PS1 controller. And if I want to use a Guncon, well, I'm shit out of luck for obvious reasons. A plug and play option that isn't going to cost me 100 bucks for an ODE is pretty useful, I don't care if there are some minor glitches. If I cared that much about "authenticity", I'd still be using all my old consoles, but they play just as well on my MiSTer to my eyes and it's more convenient to use.

>> No.8724789

>>8724782
>Honestly, I get the point of accuracy autism from a preservation standpoint, but past a certain point it legitimately turns into the same kind of shit that makes people upload "rare" revisions of some shovelware Famicom game nobody gives a shit about and nobody has ever played. Sure, I guess it was important for the sake of completeness, but is anyone really going to play it?
There's a certain case for what you say, for example nobody would really need to recreate the first shitty VIC-II that had the ceramic casing and video output so bad it looked like vasoline was smeared on the screen.

>> No.8724793

>>8724782
There are two reasons its important. First, bugs are a pain in the ass when you have to go hunting. And then it becomes a matter of ongoing support. Transistor level accuracy theoretically bypasses the problem because if you clone the hardware then the software should "just work" barring the rare quirk like the above mentioned 6502 illegal opcodes. Second, if the job is done right then it doesn't have to be done again. It can be ported indefinitely and the quality of emulation will be universally excellent.

>> No.8724794

>>8724782
thing is, you never know when a game exploits a specific hardware glitch

>> No.8724796

>>8724793
Actually, transistor level accuracy would handle illegal opcodes happily. It's gate level accuracy which doesn't.

>> No.8724802

>>8724794
For most PS1 games it's not a huge deal as they use the API functions but the Crash Bandicoot games (which are major ones everyone wants to play) do write to the bare metal and accuracy is very important.

>> No.8724814

>>8722687
About as accurate as any downloaded emulator don't fall for this $300+500 for all the attachments you'll need. Its cheaper to just buy the original hardware with a flashcart for $250
Saturn and ps1 are in a completely unplayable state and other cores suffer the same instability as early emulators
The product is only interesting in itself as it seems to have garnered an Ouya-like cult

>> No.8724815

>>8724814
>Its cheaper to just buy the original hardware with a flashcart for $250
Not for all the platforms.

>> No.8724826

>>8724796
What's the difference there?

>> No.8724829

>>8724814
>Its cheaper to just buy the original hardware with a flashcart for $250
That is slowly but surely changing. Even a Super Nintendo is upwards $80 today.

>> No.8724868

>>8724826
It takes a few transistors to build a logic gate. CMOS uses pull-up and pull-down transistors, but NMOS uses resistors in place of pull-up transistors.

>> No.8724875

Oh well, the only custom part in a ZX Spectrum anyway is the ULA and it's just some integrated switching logic that's been completely recreated and you can get new FPGA ones.

>> No.8725075

>>8724794
>>8724793
for stuff like C64 it's very important because those hardware glitches are essential to a lot of games and demos

>> No.8725127

For all that emulation has evolved since the 90's, it's still a very complex thing when you get to the edge of accuracy.
We still don't have an emulator with 100% game compatibility and ability to run all scene demos for a platform as basic and popular as the Sega Genesis

>> No.8725150

>>8724829
FPGA replacement for a SNES CPU would be nice for all the ones that got fried from Nintendo's shitty power buffering. but then again only the DMA and multiplier/divider actually matter as the rest of the chip is a basic 65816 core which WDC still makes new ones of.

>> No.8725163

stuff like SID that has analog parts can also be a bit of a pain

>> No.8725282

>>8725127
It's because it takes a certain type of person who is both capable and cares enough. Let's be real. SNES emulation wouldn't be where it is today if not for byuu/near. Unfortunately there's nobody who's independently wealthy and willing to hire people to do this kind of work so you're relying on there being a semi-obsessive fan with the necessary skillset to do the work as a hobby. And frankly the way the emulation community is it's an extremely thankless job.

>> No.8725396

>>8723210
Yeah, such as a MiSTer. Retard.

>> No.8725725

https://youtu.be/4Ckzt7uTLmY

This kind of thing is an example of something that is neat and all but has a cart before the horse vibe. They're implementing all these extra things but still just adapting an existing emulator.

>> No.8725934

>>8724814
>$300+500 for all the attachments
only if you're so incompetent you can't google alternatives. MiSTer + RAM + cheap USB hub is all you need.

>> No.8726225

>>8725725
Widescreen hacks are no different from any other form of internal resolution scaling, soulless garbage for idiots.

I don't mind eschewing accuracy for the sake of playability, as in having an option for playing on a souped up PS1 that doesn't have framedrops when there are 20 enemies on the screen, but when you start messing with the picture you've lost the whole point. You weren't meant to see more of the game.

>> No.8726475

>>8724802
>Crash Bandicoot games (which are major ones everyone wants to play)
i don't. Bring it on the API

>> No.8726814

>>8724875
Speccy chads win again.

>> No.8726919

>>8725396
I’m not sure why all these contrarians spazz out at the idea of a MiSTer and start saying “j-j-just build a PC!!!” as of a major benefit of the MiSTer isn’t that it’s small, self-contained and low-power.

>>8725725
>>8726225
This widescreen hack is hot garbage and I’m honestly bothered they’ve even bothered releasing this. The warped FOV looks like shit, rendering glitches, and it’s not at all in the spirit of “preservation”. Total shit.

>> No.8726954

>>8722925
Just use RetroArch. The whole UX is easily navigated and customized with a controller.

>> No.8726960

>>8723617
RetroArch already does this and lets you set the exact delay in ms and use any audio driver you want. Personally I just leave it at 40ms and I don't even notice.

>>8723883
DuckStation and Beetle both have 100% compatibility. Beetle is basically perfect in terms of accuracy, so can you even give a single concrete example of where the higher accuracy comes into play, or where Beetle is inaccurate with the proper settings on high-end hardware?

>> No.8726964

>>8724782
Can't you configure one controller as port two and the other as port one? Or at the very least, just set custom rebindings for games you know you will play with a unique controller. You can even load different remap files so you can swap back and forth based on your controller.

>> No.8726967

>>8723629
eh, sort of. There's not hellish much benefit over some 1000hz polled USB device but the reality is that lots of USB controllers are not polling anything like that. They are sampling their buttons, debouncing and feeding some internal state at some unknown rate while the usb controller talks to the host at a high rate pointlessly. Sure, you could go research good USB controllers that don't introduce unnecessary latency, or you can just use SNAC and remove all doubt.

>> No.8726972

>>8726960
>Personally I just leave it at 40ms and I don't even notice.
Most real consoles from the 8/16/32 bit era didn't have audio latency. There's no (significant) buffer that's being fed. The CPU tells the sound chip to play a sound and within microseconds the sound is being produced and sent to the DAC. The N64 is the first console to buck this trend where it operates like a dumb soundcard at its core albeit with some hardware assistance for mixing and such.
While we're on the subject the fact that the (windows) PC can't reliably operate with 1ms audio buffering is an embarrassment.

>> No.8726974

>>8722973
Raspberry Pi has a ton more lag than RetroArch on PC as far as I know and the tests I've seen. That might have something to do with the display connected and V-Sync (which is necessary to prevent tearing without VRR).

A high refresh-rate VRR monitor is going to be even lower lag than CRT though, since frames take less than 17ms to output. 240Hz monitor outputting at 60Hz (or 54, 55, 59, 60.1, etc. depending on the game's native) is outputting the full frame in 4ms instead of 17ms of a CRT

The main thing to avoid is V-Sync though because that's a massive amount of lag if not pair with high refreh-rate VRR (V-Sync adds no extra lag when combined with high refresh-rate VRR). If you have a fixed refresh-rate display or no VRR support, your best bet is RTSS scanline sync which has input lag roughly in-between the two

>> No.8726987

>>8726972
This is all a bit nonsense though since each person has an individual video and audio reaction time, usually from 150ms to 250ms. If you want the video/audio to appear synced to you, you don't want them at the exact same time. You have to consider both the video latency (anywhere from 16ms to 100ms depending on game/setup, could be much higher if you want actual animation latency for when you should perceive sound), and also your individualized reaction time to video vs. audio (audio usually has faster reaction time, but not always).

This all seems like pure placebo from a few dedicated autist because it's something that would be very difficult to test. It's probably from the same school of thought as retards who claimed retro consoles had "zero input lag" but now that everyone has slow-mo cameras (or hell, even just frame advance in an emulator), we know that's bullshit.

>> No.8726994
File: 185 KB, 611x856, x.jpg [View same] [iqdb] [saucenao] [google]
8726994

This is the first "is mister just emulation" thread I've seen on this board that didn't immediately devolve to childish shit-flinging. Cheers everyone.

>> No.8727080

>>8726954
RetroArch is hated by every emulator author. The MAME folks especially.

>> No.8727084

>>8726960
>so can you even give a single concrete example of where the higher accuracy comes into play, or where Beetle is inaccurate with the proper settings on high-end hardware?
Yes. It's not transistor accurate. That's the answer. There doesn't have to be a single game where it matters for that to be a relevant point when the entire marketing behind MiSTer and FPGA in general is "accuracy and preservation, man!" Either they need to live up to the hype or stop lying about it.

>> No.8727114

>>8727080
And that's my problem how? The MAME team are autistic faggots and no one cares about their opinion.

>> No.8727197

>>8726987
No wonder I've never been good at vidya. I don't think I've ever had the best reaction times.

>> No.8727212

>>8724649
There were 14 different TIA revisions. Choosing which one you FPGA recreates is a problem in of itself.

>> No.8727250

>>8727114
The MAME team has done more for emulation than anyone else. The RA team has just stolen everything wholesale. Even disregarding the inside baseball, RA is worse than just using the emulators individually because it becomes a bitch to configure and organize. It's just a cheap frontend, which have been in existence since the early days of Windows 3.1.

>> No.8727271

>>8726987
It doesn't really matter if retro consoles had zero input lag or not. What matters is whether or not the things trying to imitate them match. In that case you don't need to rely on human senses. Those are concrete measurements. It's true that most emulation is "perfect" as far as the end user experience is concerned but that's to be expected. We're now at a point where you get into the weeds since most of the big stuff is done already. I'll be the first to admit I would fail horribly at a double blind test if you sat me in front of a real SNES and an emulator and made me pick. But that shouldn't be the end of the process where everyone just packs up and goes home.

>> No.8727290

>>8727212
If you're using an FPGA you shouldn't need to make that decision in the first place. You can literally recreate all of them. That's one of my beefs with MiSTer. I should have the option to choose to synthesize not just "Sega Genesis" but any one of the variants of the Sega Genesis.

Now I know the response. There's an accepted "best" revision of these consoles so you shouldn't ever want to use a shittier one. But MiSTer also implements computer cores up to 486. I would love the option to essentially build a rig right in FPGA.

>> No.8727292

>>8727250
MAME team is faggots and FB Neo is much better for the games that support it (like 99% of pre-2000 arcade titles).

There is literally NO alternative for the dozens of features RetroArch offers, and you will be missing out on them all if you get filtered by being too stupid to figure it out (you only need to learn RetroArch once and it's the same for dozens of platforms and consoles)

People use it for the
>consolidated UI and settings for hundreds of different emulators
>run-ahead
>full rebinding
>VRR, BFI, and sync to exact framerate
>shaders, filters, and overlays
>audio customization and level
>unlimited save states, pause, rewind, fast-forward, frame advance
>cheat support
>network play
>50+ different hotkeys with full rebinding
>Retro achievements that are uncheatable with a hardcore mode to disable save states, cheats, pause, rewind, and fast forward
>updated forks
>being able to set any aspect ratio or refresh rate
>Rotation/TATE support
>recording, screenshots, and streaming
and being able to save ALL OF THESE, as well as specific core options and global settings, on a per-game, per-core, per-directory, and global basis. That means each and every game can have its own rebindings, shaders, volume and audio settings, run-ahead level, core and core options, rotation, aspect ratio, etc.

There is literally no reason to use anything else to play gen 5 and earlier games

>>8727271
I'm pretty sure certain audio drivers offer very low latency that should be just as good as original hardware or extremely damn close to the point where no one can tell even comparing side-by-side.

That's not really a fault of the emulator though, but rather the audio drivers available today. Again, I'm not even actually sure if it is a problem or to what extent because I've never actually seen it thoroughly tested, and just like with input lag, placebo "muh feels" means basically nothing.

>> No.8727303

>>8727292
You don't need RetroArch to do any of that. It may make it more convenient if you switch emulators a lot but it's not necessary.

>> No.8727307

>>8727303
You do need RetroArch for like half the list, and it's significantly more convenient to use one consistent UI that allows you to so easily save all settings on a per-game or per-core basis with the press of a button.

Genuinely the reason so many people hate it is an IQ issue.

>> No.8727313

>>8725282
>if not for byuu/near
He truly was a giant in the emulation scene, controversies aside

>> No.8727460

>>8725725
>cart before the horse vibe
Perfectly described

>> No.8727895
File: 557 KB, 1702x1024, PokeyMAX.jpg [View same] [iqdb] [saucenao] [google]
8727895

>> No.8727920
File: 115 KB, 1280x720, image.jpg [View same] [iqdb] [saucenao] [google]
8727920

>>8726919
>MiSter
>it’s small, self-contained and low-power.
So are modern PCs. They also can do a bit more than run ps1 with debatable code.

>> No.8727997

I think it depends. Something like the Motorola 68000, used in a hundred architectures, is very well understood so that's far more likely to be transistor accurate even if the rest of the rest of the system isn't.

>> No.8728457

>>8726987
>If you want the video/audio to appear synced to you, you don't want them at the exact same time
Actually, you do.
Take a hammer and smash something. The sound, the feeling, and what you see will be simultaneous. Lag differences in human senses are corrected within the brain.

>> No.8728503

>>8726994
I love this guy's art but the PVM triggers me.

>> No.8728504

>>8722687
Why would I want to use this thing, instead of a custom built system tailored to a particular console experience?

>> No.8728558

>>8727920
>MicroChang MemeBox
At that point you’re just actively avoiding the MiSTer for no good reason.

>> No.8728602

>>8728558
nta, but thats an actual mini pc with a proper x64 CPU, not an ARM console.

https://www.youtube.com/watch?v=fLATODi7KlU

and yeah, im sure a MiSTer machine might be good, but even a low-tier ARM chinkcrap console can emulate more than a MiSTer.

>> No.8729191

>>8726994
The question is a bit more pointed so that helps. Really though its very hard to argue it isn't emulation when at least some of the cores are verilog conversions of actual emulators. It's frustrating the FPGA scene is so unambitious.

>> No.8729230

>>8728457
Except this is something that you literally can't notice directly side-by-side and each game has different programming for how the sound works in conjunction with the animation.

>> No.8729239

>>8728602
i love the NUC form factor, i use one for a low-spec media pc