PS1 scaling, softening, lag on pstv, ps3 vs PS2 Best option?

[Guile]

I do not have an original PS1 and always used emulators, ps vita and pstv, or ps3 to play PS1 games. I tried popstarter on PS2 but it was too buggy for me.

I finally got around to trying the new PS2 mechapwn exploit and I was amazed at the difference when running PS1 games on a PS2 through a GBS-C. The lag reduction is extremely noticeable and the scaling and lack of a softening filter looks so much better to me than what I was used to.

After seeing that I'm thinking of using the PS2 much more for PS1 games but I was wondering if there's anything I'm missing with the other options. Has anyone tried these various methods of playing PS1 games and had better results?

[WobblingPixels]

PS1 is the best option to play PS1 games.

PS2 doesn't have a good GMU emulation for transparent sprites:

PS2:
https://i.imgur.com/7Uimhbz.jpg
PS1:
https://i.imgur.com/gjeytVD.jpg

Not all PS1 games are compatible depending on the PS2 model:

https://en.wikipedia.org/wiki/List_of_P ... yStation_2

PSTV has much higher input lag.Not worth it IMO.
PSP is OK but the output isn't that sharp compared to PS1 or PS2
PS3 output with PS1 games is to soft but not with all games. AFAIR the input lag is much higher (compared to PS1, PS2 and even PSP) when using HDMI output but not that bad as with PSTV.

[spmbx]

The mister ps1 core is coming along nicely but its not quite there yet at this time. Not to mention its pretty hard to get the hardware.

[Thomago]

The PS1 compatibility list is likely far from complete, by the way. From personal experience, I can add Metal Gear Solid PAL (reproducible crash in a late-game FMV on the infamous SCPH-75004; I had to use a PSone to complete the game).

[NormalFish]

The PS1 compatibility list is likely far from complete, by the way. From personal experience, I can add Metal Gear Solid PAL (reproducible crash in a late-game FMV on the infamous SCPH-75004; I had to use a PSone to complete the game).

As noted in the Wiki citations, that list is mostly what Sony themselves communicated about compatibility issues. There are tons of minor issues.

Thankfully, most of them aren't distracting unless you're extremely familiar with the original. If someone is willing to play on a PS3 or via POPS (which lacks scaling options that will be more destructive to some games than the minor visual bugs that are most typical), they'll probably be happy with a PS2. Rare crashing issues like the one you mention aside.

[Guile]

PS1 is the best option to play PS1 games.

PS2 doesn't have a good GMU emulation for transparent sprites:
.

That's interesting, I've never seen that before. I'll have to try more games and see if I notice any bugs.

The mister ps1 core is coming along nicely but its not quite there yet at this time. Not to mention its pretty hard to get the hardware.

I didn't think it would progress that fast, that's very impressive. It's too bad it's still rather expensive, I wish the parts would come down in price but they seem to just keep going up.
Have you tested the ps1 core and do you find it playable?

The PS1 compatibility list is likely far from complete, by the way. From personal experience, I can add Metal Gear Solid PAL (reproducible crash in a late-game FMV on the infamous SCPH-75004; I had to use a PSone to complete the game).

That's surprising that one of the biggest ps1 games would crash like that. I have a few different ps2 models so I guess I can look up known game and revision issues before playing.

[svcoco]

I finally got around to trying the new PS2 mechapwn exploit and I was amazed at the difference when running PS1 games on a PS2 through a GBS-C. The lag reduction is extremely noticeable and the scaling and lack of a softening filter looks so much better to me than what I was used to.

Once I was in similar situation like you (I have Ps3, PSP Go, PSvita, PSTv), so I got a slim PS2 SCPH-7500X and installed mechapwn to play some ps1 import discs, then I found that SCPH-7500X runs ps1 games on a PowerPC based emulator but without the tools to objectively measure accuracy and input lag I wasn't sure if it was the best way to play PS1 close to the real deal, so I finally decided to buy a PS1 and installed the x-station ODE (after several failed attempts to get the laser to work).

I think if you are looking for something close to the real thing (I might be wrong on this) the options are:

1-PS1 (obviously)
2-PS2 Fat (uses ps1 hardware)
3-PS2 Slim (up to SCPH-700XX model uses ps1 hardware)
4-PS2 Slim (from SCPH-7500x to latest SCPH-9000x model uses emulator)
5-PSP Go (also emulation but cool features like portable gaming, analog video out and Dualshock 3 support)

[AaronSR]

There is actually a much more complete list (the one for Japan is still up but US is not, archive to the rescue though)

https://www.jp.playstation.com/info/not ... 75000.html
https://web.archive.org/web/20100309031 ... /index.htm

Of course starting with the 75K it is purely emulation AFAIK. You can't really beat using an actual PS1.

That said as a speedrunner, the 75K+ is a must for a lot of games and being from a PAL region MechaPwn has saved me quite a lot of money lol

[Guspaz]

All versions of the PS2 use some degree of software emulation. The launch PS2 had to emulate the entire PS1 GPU in software.

[Ed Oscuro]

Your *best* options are PS1, for complete accuracy, or a cheap PC running an emulator like Duckstation or Retroarch Beetle PSX HW. Many original PlayStation games have low or inconsistent framerates and even the cheapest and smallest PCs today (for Windows or Linux) can run the PSX with an overclocked GPU - technically, the GTE coprocessor. On top of that, with a PC emulator you can also easily use the reportedly improved PS1 firmware images that Sony shipped with the PSP (and the later versions shipped on the PS3, etc.), which may improve things even further. Of course, there are some drawbacks still like laggier sound than the original on most sound hardware, which may be a deal breaker but no worse than the other emulator options. I'm no expert on PS1 emulation on PSTV, RasPi, etc., though.

[H6rdc0re]

Your *best* options are PS1, for complete accuracy, or a cheap PC running an emulator like Duckstation or Retroarch Beetle PSX HW. Many original PlayStation games have low or inconsistent framerates and even the cheapest and smallest PCs today (for Windows or Linux) can run the PSX with an overclocked GPU - technically, the GTE coprocessor. On top of that, with a PC emulator you can also easily use the reportedly improved PS1 firmware images that Sony shipped with the PSP (and the later versions shipped on the PS3, etc.), which may improve things even further. Of course, there are some drawbacks still like laggier sound than the original on most sound hardware, which may be a deal breaker but no worse than the other emulator options. I'm no expert on PS1 emulation on PSTV, RasPi, etc., though.

You could run in run ahead mode but for the best performance I would say play on the original PS1.

[Ed Oscuro]

Input lag is now almost a non-issue with modern operating systems and input lag tweaks. Run-ahead seems less stable and less important than baseline input lag fixes that have been rolling out the last couple years.

[ZellSF]

Input lag is now almost a non-issue with modern operating systems and input lag tweaks. Run-ahead seems less stable and less important than baseline input lag fixes that have been rolling out the last couple years.

Source? Because every time I've tried an emulator lately it's been an issue.

[NormalFish]

Input lag is now almost a non-issue with modern operating systems and input lag tweaks. Run-ahead seems less stable and less important than baseline input lag fixes that have been rolling out the last couple years.

Source? Because every time I've tried an emulator lately it's been an issue.

It obviously depends on the emulator. Some emulators will be less accurate and less laggy, some will be cycle accurate and exactly as laggy as a proper console, and some will be less accurate and more laggy. And everything in between. There's a lot of public discussion of emulators and their lag, so thankfully you can probably find tweaks online to tune the lag to your liking.

[Rulumi]

The DECKARD IOP is supposed to have a MIPS APU connected to the PPC where the actual MIPS core instructions are "emulated" and run there.
I imagine the PS1 MIPS instructions are run there too instead of directly on the PPC.

[Ed Oscuro]

Source? Because every time I've tried an emulator lately it's been an issue.

I'd love to hear your experiences so please share more details. 2D games may present a worst case for lag perception.

If I'm emulating a PlayStation game which benefits from the GTE fix - which is most of the 3D ones - responsiveness is improved by more consistent frame pacing and faster frame delivery. Jumping from 30-50 fps to even slightly less than 60 fps locked is quite a boost which I personally notice most. Of course, for a well-tuned game which already ran a stable 60fps framerate, I'd expect that latency could be an issue. Granted, none of this is "original" behavior.

There are some recent fixes which make things better: Most top emulators have specifically adopted frame pacing tech and especially Dolphin team have promoted this with graphs. Retroarch has a slew of input lag-reducing controls (although it is hard to measure the effect of some). Low-latency audio backends are cropping up - WASAPI, ASIO, and whatever the Mozilla API is called show up in a few emulators nowadays. Vulkan renderers are showing up and allowing some nice gains especially for 3D systems that struggle to hit original framerates.

I'm not sure if there are any "modern" end-to-end latency comparisons for many emulators. MAME may be the odd one out, but I think I've seen some steps made there towards combating latency too, although it's not strictly ever been a part of their design document (which is a shame).

[gray117]

I've no real idea - but appreciate the trend. I've no idea the lengths that are gone to create cycle accurate emulators, and similarly the various fpga approaches, but I think both of these things have been a bit of challenge, and prompted renewed interest in reducing some of these problems across all types of emulators.


Really I suspect perfecting v-sync or frame pacing is more an icing on the cake compared to the myriad of pitfalls that go into creating these emulators and solving timing/sychronising issues that would have just been managed by hardware and could have been assumed by the games ... although it is none-the-less very admirable and apprecitated when emulator developers do dedicate efforts towards the experience of playing these games.

I think the problem for many emulators is that there maybe certain functions that simply rely on a dependancy that cannot be scheduled or is difficult to shedule without the hardware; so a timing may simply become dependant on the next frame call of the emulator and assumption that the game loop was completed. Any delay on the loop in that you can't keep up with the target fps, and potentially you end up with a bit of a metal slug 2 cascade effect where lag is compounded by performance.

[Ed Oscuro]

I believe you're talking about different things. With the usual disclaimer that I'm not a programmer or emulator coder, these are all different issues: cycle accuracy, framerate, frame pacing, and frame latency.

In cycle accurate emulation, a component will act during one of its emulated clock cycles as it would in a real system. It'll stall waiting for emulated memory only when and if it would on the real device. This should help those creeping errors from showing up like the famous Speedy Gonzales example fixed in bsnes. Every 3D era system emulator I've read about gets realtime performance because it doesn't strive for this level of accuracy when dealing with the many simple components like the individual processors working on pixel data in parallel with many other processors on a GPU. (Newer systems seem to be designed with more attention to making the system's operating cycles a bit easier to predict and less likely to trip up if one component falls behind.) Dolphin and Flycast deal with the special cases where the approximation isn't good enough for simulating render-to-texture effects, but this still doesn't seem to be "cycle accurate emulation." Many of the other routine elements of 3D image creation are being carefully approximated on modern hardware. Of course, there's also an interesting possibility that removing cycle accurate behavior - like a memory stall - will let the rest of the system perform better than it could in the real system, without making it more likely fatal errors are going to creep in.

One of the seemingly simplest things for any emulator to get right is emulating the basic framerate, right? Even many poor emulators emulate the correct framerate without proper behavior. This doesn't tell us if the host is fast enough to emulate in real time, or how fast the host system presents each emulated frame to the display. Borrowing from modern PC and console gaming, frame pacing ensures that frames are delivered at an even pace to the monitor, so none are live too much longer or shorter than others. Frame latency is related to frame pacing, measuring instead how long it takes for a frame to get to the monitor. You can normally have low latency but still have really bad frame pacing - for example, a spurt of frames output quickly, each lasting less time than they should, followed by the emulator hitting a bottleneck and taking too long to get out the next frame. Of course some simulated systems can have poor frame pacing themselves, but this is separate from the host system's (and emulator's) frame pacing. Whew!

Finally, variable refresh rate does help with latency (waiting for v-sync isn't as necessary) and it can accommodate bad frame pacing by just displaying each frame as they come - but it doesn't do the job of frame pacing.

[NewSchoolBoxer]

All these responses telling you to play on PS1, I think are silly. There is more than one correct answer. There is also the question on how well PS1's were built to last. I anecdotally think PS2's were made better. Regardless, every optical disc drive will have to be replaced in the near future. OP is suggesting they don't need to play off real disc or backup disc at least.

If you want to be a streamer, it's hard to beat a PS3 that has HDMI out that carries lossless video and audio and forces everything into IN SPEC consistent resolution and framerate. Easy to criticize how the emulation and scaling work on certain games though. Scanlines gone but no one going to see them anyway in a Twitch stream or YouTube upload and I never noticed scanlines in 3D games myself.

I go the PS2 route with Fast Disc Speed option. The greatly reduced loading times over PS1, no going back for me. I bought a slim model 10 years ago. Never had a game issue. Thick models cost more these days because they can use an HDD and that may be relevant. I respect the argument that thin and late thick (75000+) models lack the PS1 hardware and the resulting emulation has game-specific issues. Flagged for issues Twisted Metal 2 and Final Fantasy Anthology played perfectly fine on thin model for me, however. Wouldn't surprise anyone if PAL is worse off but just saying X models emulate and emulation = worse, let's see proof on NTSC games people actually play.

PS2 also has YPbPr/Component which is relevant outside Europe. I'm trying to be more scientific and compare against Japanese CRT and different PS2 models but every PS1 game I try in Component, the coloring is wrong. Extremely noticeable on Final Fantasy VII. My theory is PS1 games expect 7.5 IRE black level but Component encodes with 0 IRE black level. PS1 over PS2 game colors look correct in composite, s-video and RGB.

WobblingPixels transparency comparison on non-obscure game is fair point. No mention of model number or video format and more than one way to code transparency. Fighting game health bars like to use dithering that wouldn't be affected. I think OP should play on emulator or PS2 model of choice and wait to see if they ever notice an issue. I think PS1 that can only play PS1 games is the worst possible option.

I tried (free) DuckStation Windows 10 route too. Made my own ISO's to compare and every game looks and plays great except interlaced video on the < 1% of games that use it outside of menus. In terms of non-obscure games, this is Ridge Racer Turbo, Tekken 3 and Dead or Alive. DuckStation option to disable interlacing solves most of the problem and its 4K (9x) upscaling does the rest even on my 1080p capped display.

I haven't tried PS1 on PSP/PSTV/PS Vita.

[Ed Oscuro]

Sony had a learning curve on reliability with each of their consoles up through the PS3: Early models all have their weak points. On the PS1, parts of the sled are plastic which deform under heat. The early PS2s were notorious for failure, and same for the first model of PS3. Late "fat" PS2s are widely considered reliable and I think the PS3 Super Slim is very reliable (no experience using it to play PS1 titles though).

Standalone Duckstation seems to be making good progress on the video front. It doesn't explicitly offer variable refresh rate awareness, but its v-sync setting can be combined with forcing it on in a driver. The Vulkan renderer only seems to offer borderless fullscreen at 60Hz. Ideally, it would have an exclusive fullscreen mode with variable refresh rate and high frame rate for the best latency performance. There's a D3D12 backend which does offer very high refresh rates and exclusive fullscreen but I don't know if the performance is as good as Vulkan.

[Kyusha]

That said as a speedrunner, the 75K+ is a must for a lot of games

why?

[Ed Oscuro]

He's referring to this: https://www.reddit.com/r/speedrun/comme ... s_because/

There is a comment in that thread stating that the situation isn't as simple as "scph7500x is always faster than a real PS1," though; YMMV. Test on a per-game basis for best results.

[azmun]

I think PS1 that can only play PS1 games is the worst possible option.

I get there are various choices to run PS1 games (via emulators, PS2 and PS3) but to say running PS1 games on original PS1 hardware is the worst option, would you care to expound on this please? Apart from the wearing of optical drives (which PS2 and PS3 are not exempt) and possibly slower load times, am I missing some other point(s)?

[Einzelherz]

PS2 also has YPbPr/Component which is relevant outside Europe. I'm trying to be more scientific and compare against Japanese CRT and different PS2 models but every PS1 game I try in Component, the coloring is wrong. Extremely noticeable on Final Fantasy VII. My theory is PS1 games expect 7.5 IRE black level but Component encodes with 0 IRE black level. PS1 over PS2 game colors look correct in composite, s-video and RGB.

How are you comparing these? I've yet to find a monitor that lets you calibrate RGB separately from component.