Aftermarket SNES Cartridge Expansion Hardware

[SNES AYE]

But that could be done with a fantasy console that exists purely in FPGA.

But then it's really not the SNES imo, at least not for me personally.

Now, I don't personally care about some technological debate of it using a new chip being real SNES or not, as that's a debate for people coming from a totally different place to me, but I have this thing where playing on a PC or a clone console or a smartphone or even a Switch, etc, simply isn't the same thing as playing an actual SNES, which is the particular childhood feeling I want to experience again. When it comes to SNES, for me at least, only playing on the real hardware or the SNES Mini capture that feeling in modern times. And it's that feeling of playing SNES that I'm always looking for personally.

Maybe it all comes down the purity and simplicity of it all, just sitting in front of the living room TV, plugging in a game cartridge or at least starting a game with none of the convolution and fuss of modern times, picking up that authentic SNES controller, and I'm off, transported to my childhood again.

So, to go back to the original question once more, any cartridge that plugs into a SNES and plays a SNES game absolutely captures that SNES feeling for me, regardless of what's inside the cartridge really. Although, not if the end game is truly so far beyond anything I could imagine actually running on a SNES that it in and of itself it no longer looks, sounds, or feels like playing a SNES game at that point again. So there is some limit in my mind as to how far a game can go in terms of graphics and such before it stops even being a SNES game. Playing Crysis on SNES that is Crysis in basically every way other than output colours and resolution just isn't even remotely SNES for me. Playing some hypothetical version of SNES Star Fox using that guy's ray tracing engine somehow would be acceptable to me for some reason though. It's a bizarre one for sure. But I have my own internal feeling on what seems right there basically. Doom Definitive Edition, yes. Xeno Crisis, yes. A hypothetical re-release of Star Fox on a physical SNES cartridge running on that guy's engine, crazily yes. Playing something like Call of Duty Black Ops 6 or Fortnite or Tears of the Kingdom just output via the SNES on some old TV or whatever, nope, not even remotely, not even if it's doing so via a cartridge plugged directly into a real SNES console.

I'm just sharing my own thinking in on things there. Each to their own though. And my personal view is that everyone is and should be free to enjoy playing SNES or experiencing something related to SNES however they see fit.

[GValiente]

I've been looking at that Doom SNES (Super FX 3 ver.) thread, and though that was a remaster of a historical port that originally used the real Super FX 2 of its time, it still made me wonder about what you people think about aftermarket SNES cartridge expansion hardware in general.

For me, if a game uses a modern coprocessor, it's no longer retro, and it interests me as much as the games for modern platforms (not much).

For me it's not about the technical merit of avoiding the use of coprocessors, I wish more homebrew used the SA-1 for example.

[tepples]

To avoid the display compositing and audio mixing lag associated with Microsoft Windows and Android

But that could be done with a fantasy console that exists purely in FPGA.

I'd be interested to see how many people own a Super NES vs. how many own a MiSTer.

[93143]

Playing Crysis on SNES that is Crysis in basically every way other than output colours and resolution just isn't even remotely SNES for me. Playing some hypothetical version of SNES Star Fox using that guy's ray tracing engine somehow would be acceptable to me for some reason though.

I suspect our views are more similar than one might have thought, even though you're coming from a player's perspective rather than a programmer's.

The Super RT would have been ridiculously expensive in 1995, and would probably have required its own power supply, but it was probably possible. It's an edge case for me - I probably wouldn't use it myself, but I can respect it and it would be cool to see somebody make a game with it.

The MSU1 is fun as an option, or for projects that demonstrate that storage capacity was the only obstacle to the SNES doing a particular thing (Super Road Blaster is a good example). But it is a little implausible. We went over this in another thread, and concluded that while the MSU1 data capability could be approximated with a buffered CD-ROM reader, the best way to actually meet the spec in-period would be a tape drive. Which is hilariously impractical, but the MSU1 spec doesn't actually specify a maximum seek time, so...

Then again, in a different thread, I argued that the MSU1 could technically be considered a "legacy" add-on, even though it was designed well outside the commercial era. I've tended to consider storage memory a less important restriction when compared to computing power and even RAM, perhaps because it's so easy to design a bankswitching chip to blow past the limits; indeed the NES is famous for most of its games using such chips. This makes cartridge cost the only real restriction, making it a soft limit rather than a hard one, that moved quite a bit even during the commercial lifetime of the SNES. Technically you could have done a zero-latency MSU1 quite easily in the '90s; it would just have been impractical to use the entire 4 GB address space...

I wish more homebrew used the SA-1 for example.

I wish more commercial games had actually taken advantage of it instead of just using it for copy protection or to save optimization time or whatever was going on there. If I ever get to it (sadly, it's probably safe to say I won't), I have an F-Zero game planned that's a perfect fit for the Super Accelerator.

I'd be interested to see how many people own a Super NES vs. how many own a MiSTer.

Real SNES here. The CPU started acting weird on my old one, so I bought another one. It's been very helpful for some of the hardware-pushing tricks I've been up to; you can't do definitive tests on an FPGA clone.

[KungFuFurby]

Real SNES, version 2-1-3. Yes, it's even hooked up to a CRT TV, but that's mostly because of the surviving video output cable, though one of the speakers on the TV in question failed.

My word on the modern cartridge debate? To each their own: it is clear that they continue to give the hardware its own life, even if they go modern. I know that, at least on the graphical side, there are bound to be some bottlenecks here and there. Plus, from a nostalgia perspective, there is such a thing as uncanny valley, and not everyone is going all-in stylistically in that direction. That's fine by me, since people are still tinkering with the hardware well after the original commercial era.

Interestingly, when it comes to hardware, to me it narrows down how you'd be able to also let it run on the emulation side... that is, the many users running on the PC, simply because it isn't implemented.

The MSU-1 sounds like the ultimate in data storage in audio. The data storage part has undefined transfer rates if I recall correctly, and the audio is just one stereo file at a time at a fixed sample rate. That's perfectly fine for what the chip set out to do. I don't consider myself to be the target, mostly because part of the reason why my nostalgia landed on the SNES in the first place is because I found later games to be too expensive filesize-wise, especially when instead of a physical disc or cartridge you have to store it on the hard drive and consume memory that way... and in my case specifically, I find myself to be a cheapskate in the memory department, so I find just taking a few megabytes for a SNES game to be a good deal. So to me, I see it both ways: from a player's perspective, well, it's either physical or digital, and digital costs memory. From a hardware perspective... well, to each their own.

[Oziphantom]

Coming from the Commodore scene gives me a different view I guess. In C64 land we are very staunchly "STOCK or GTFO" as the whole point of doing something on a C64 is too see if it can be done, to push the machine beyond what it should be. And if you just bolt on a new fancy chip then what is the point? If you wanted an Amiga go make something for the Amiga

Or "I wanted an Amiga but almost no one has an Amiga in my country." It's like devs wishing more people in Latin-alphabet countries owned a PC Engine or TurboGrafx-16, and resorting to either making "wonder-mappers" for NES, such as Rainbow and MXM, or severely underusing the SNES.

This seems like an odd juxtaposition, you are inspired by a machine and want to make something for it, but then stopped because not enough people you know have it. In a world of emulation that problem is gone, but if you make something interesting on the platform then you could introduce more people to that platform in said language and boost its popularity.

Putting a couple of Ghz+ ARM cores into a snes cart to run the game, why? what is the point, you haven't made a SNES game you have made a game that uses a SNES as a graphics card. Great, why did you choose a SNES in this case?

To avoid the display compositing and audio mixing lag associated with Microsoft Windows and Android. Or to avoid the lack of physical buttons on mobile.
[/quote]
If the SNES is not powerful enough/to too hard to develop for, the far more popular PS2 and Wii say high. People happily take 2D Arcade games on those platforms, so while you might need to up the graphics colour depth, it just removes the hassle of making graphics as you are now 256 colours or 15bit colour, and the CPUs are powerful enough to handle anything you would want "a beyond normal lazy coded snes product" to do.

[Oziphantom]

But I am open to "what ifs", what if they put this concept that was in another thing from the era into a cart, what would that have done, how would it have evolved a SNES game. But I can't actually think of anything from the era that the SNES is missing.

The SNES-CD... which is why Near developed MSU1.

I don't really see the SNES-CD as moving the needle much, it gives you a tonne of storage, but what do you do with it? Even on the PS1 we were making 6-8MB games and then filling the disc with FMV and RedBook Audio to "fill it". Granted it makes it more economically viable for large RPGs or Visual Novel type games, it lowers the cost of manufacturing allowing smaller print runs and hence more experimental games, so it has a big impact on the economics of video game development.

[Oziphantom]

To avoid the display compositing and audio mixing lag associated with Microsoft Windows and Android

But that could be done with a fantasy console that exists purely in FPGA.

I'd be interested to see how many people own a Super NES vs. how many own a MiSTer.

I would think that a actual SNES still out numbers MiSTer, SNES is a very common machine, very common with speed runners where they need hardware. While MiSTer is very expensive. However I expect the new $100 clone MiSTers might change the split and bring the MiSTer closer to winning, especially as it gets very expensive to get a good image on a composite signal as tvs have dropped analouge inputs for some time now.

[Hojo_Norem]

Putting a couple of Ghz+ ARM cores into a snes cart to run the game, why? what is the point, you haven't made a SNES game you have made a game that uses a SNES as a graphics card. Great, why did you choose a SNES in this case?

To avoid the display compositing and audio mixing lag associated with Microsoft Windows and Android.

I'm curious, realistically how much of that is a 'show stopping' problem, on the scale of "Not really a problem unless you are using the worst monitor out there" to "bare metal or go home"?

EDIT: In a strictly non-emulation scenario.

If it is a big problem then I don't notice it myself. I pull the trigger and pretty much instantly see my modshards turn the Thargoid in front of me into chunky, caustic salsa.

[Pokun]

To avoid the display compositing and audio mixing lag associated with Microsoft Windows and Android

But that could be done with a fantasy console that exists purely in FPGA.

I'd be interested to see how many people own a Super NES vs. how many own a MiSTer.

I'm very sure that has got to be a real SNES. Although it wasn't as popular as the NES, the SNES was still a highly successful system back in its day so there still got to be many people that still got one in their attic, whether or not they are still playing it anymore.

The MiSTer on the other hand is more of a niche retro toy than a mainstream video game console and I'd bet that people that own a MiSTer are likely to be enthusiasts that are also quite likely to own and play a real SNES. More casual players that just wants to play some SNES once in a while probably owns a SNES Mini or uses other emulators that are more user-friendly than the MiSTer is.



Regarding the real-deal-or-go-home-discussion, I don't think the whole point is to push the system to its limits is the main point for everyone, although it is definitely the main point for many of us. In my case I just have a passion for these systems, I'm not really interested in pushing any boundaries of what they can do, nor is that within my ability in the first place.
But authenticity is still as important to me though, so yeah I'm not very enthusiastic about a new aftermarket game that uses modern technology to push boundaries beyond what was possible back in the day. I mean I'm not against it, but it doesn't excite me either as it feels like you can just put anything in a cartridge.
Using modern technology for more practical reasons is not the same thing though and is therefore fine, like using programmable logic to replicate mapper chips that are no longer produced.

Makign a fantasy system in FPGA is also something I think is interesting. Say an improved version of the SNES would be fine, I wouldn't call it a SNES anymore though as it no longer fits the definition.

[93143]

To avoid the display compositing and audio mixing lag associated with Microsoft Windows and Android.

I'm curious, realistically how much of that is a 'show stopping' problem, on the scale of "Not really a problem unless you are using the worst monitor out there" to "bare metal or go home"?

I have a two-year-old Windows 11 laptop with high-end specs (RTX 3080 Mobile, OLED 4K @ 60 Hz with GSync). I typically use three frames of runahead in Mesen (with VSync off), and I think it feels about right. It does glitch, though, when the prediction it was using turns out to be wrong and it has to teleport some sprites and start (or stop) a sound effect partway in. This is okay for something like Super Mario Bros. 3 (which I've, uh, played quite a bit of lately), but a game like Gradius III suffers because of the player sprite's instantaneous acceleration, and Tetris is borderline unplayable at high levels. Using four frames of runahead also feels about right, but the glitching seems way worse for some reason. Audio lag is substantial; I have to set the latency to 60 ms to avoid crackling.

I also have access to an old Sony Bravia 40" LCD TV. Playing the N64 version of Mario Golf on it is pretty much impossible, as the lag is something like 1/6 of the swing meter. Thank goodness we kept our old Panasonic Tau...

Somebody did the math a while back and concluded that Mike Tyson in the original NES Punch-Out!! cannot normally be beaten by skill if played on a modern flatscreen TV, because his tell doesn't leave the average human brain enough time to react...

Using modern technology for more practical reasons is not the same thing though and is therefore fine, like using programmable logic to replicate mapper chips that are no longer produced.

Strongly agree. Function is what's important to me. As long as it might as well be the old technology from the perspective of both the programmer and the player (this includes not doing stupid things with 3.3V parts), the manufacturer can get there however he wants.

Makign a fantasy system in FPGA is also something I think is interesting.

Ultra 64 with 100 MHz CPU, 66 MHz RCP, fixed memory controller, clamped blender and no multitexture offset bug.

[tepples]

Coming from the Commodore scene gives me a different view I guess. In C64 land we are very staunchly "STOCK or GTFO" as the whole point of doing something on a C64 is too see if it can be done, to push the machine beyond what it should be. And if you just bolt on a new fancy chip then what is the point? If you wanted an Amiga go make something for the Amiga

Or "I wanted an Amiga but almost no one has an Amiga in my country." It's like devs wishing more people in Latin-alphabet countries owned a PC Engine or TurboGrafx-16, and resorting to either making "wonder-mappers" for NES, such as Rainbow and MXM, or severely underusing the SNES.

This seems like an odd juxtaposition, you are inspired by a machine and want to make something for it, but then stopped because not enough people you know have it. In a world of emulation that problem is gone

People not familiar with a platform's existing library lack an intuitive sense of the hardware's limits. I imagine that people who didn't grow up with PCE might complain that PCE games' visual complexity falls in an "uncanny valley" between NES and Super NES capability, or that it's "NES with cheating". In addition, in the specific case of the Amiga, two companies (Cloanto and Hyperion Entertainment) have been asserting copyright in the system software to keep the platform closed. As I understand it, any commercial release of an Amiga game on a modern platform needs a license from both. See this post by wyatt8740.

[Oziphantom]

Or "I wanted an Amiga but almost no one has an Amiga in my country." It's like devs wishing more people in Latin-alphabet countries owned a PC Engine or TurboGrafx-16, and resorting to either making "wonder-mappers" for NES, such as Rainbow and MXM, or severely underusing the SNES.

This seems like an odd juxtaposition, you are inspired by a machine and want to make something for it, but then stopped because not enough people you know have it. In a world of emulation that problem is gone

People not familiar with a platform's existing library lack an intuitive sense of the hardware's limits. I imagine that people who didn't grow up with PCE might complain that PCE games' visual complexity falls in an "uncanny valley" between NES and Super NES capability, or that it's "NES with cheating". In addition, in the specific case of the Amiga, two companies (Cloanto and Hyperion Entertainment) have been asserting copyright in the system software to keep the platform closed. As I understand it, any commercial release of an Amiga game on a modern platform needs a license from both. See this post by wyatt8740.

For most games, you would be able to just trap and do background emulation of any kickstart functions you would need. Unless you are making a "workbench" game, to which you would then need to ship workbench which would also be an issue.

[Hojo_Norem]

To avoid the display compositing and audio mixing lag associated with Microsoft Windows and Android.

I'm curious, realistically how much of that is a 'show stopping' problem, on the scale of "Not really a problem unless you are using the worst monitor out there" to "bare metal or go home"?

I have a two-year-old Windows 11 laptop with high-end specs (RTX 3080 Mobile, OLED 4K @ 60 Hz with GSync). I typically use three frames of runahead in Mesen (with VSync off), and I think it feels about right. It does glitch, though, when the prediction it was using turns out to be wrong and it has to teleport some sprites and start (or stop) a sound effect partway in. This is okay for something like Super Mario Bros. 3 (which I've, uh, played quite a bit of lately), but a game like Gradius III suffers because of the player sprite's instantaneous acceleration, and Tetris is borderline unplayable at high levels. Using four frames of runahead also feels about right, but the glitching seems way worse for some reason. Audio lag is substantial; I have to set the latency to 60 ms to avoid crackling.

I also have access to an old Sony Bravia 40" LCD TV. Playing the N64 version of Mario Golf on it is pretty much impossible, as the lag is something like 1/6 of the swing meter. Thank goodness we kept our old Panasonic Tau...

Somebody did the math a while back and concluded that Mike Tyson in the original NES Punch-Out!! cannot normally be beaten by skill if played on a modern flatscreen TV, because his tell doesn't leave the average human brain enough time to react...

I agree whole-heartedly that when emulating or using OHW incurs penalties unless you take steps. It's one of the reasons I snagged a OSSC Pro.

However, in the context of shoving multi GHz co-processors into a retro machine to heavily bypass that machine's limitations? You are no longer making a game for that system. In fact, the SuperFX only gets a passing grade because Nintendo released it. Oziphantom was asking "why?". If you need a multi GHz processor to run your game then you are either a terrible programmer, making a 'modern' game with retro aesthetics or want something to serve as anti-emulation DRM.

My question was referring to the second possibility. Computer operating system overhead has been a thing since operating systems went pre-emptive multitasking. Developers have had to account for that for decades and to me and I dare say many others "display compositing and audio mixing lag" hasn't been much of an issue.

I have adjusted my original question to better reflect my thoughts.

[Nikku4211]

It seems like everybody's talking about extreme high-end hardware meanwhile when I created the thread, I was thinking more along the lines of some sort of embedded processor that isn't anywhere near as powerful as whatever NVidia or AMD's pumping out or even Intel's stuff.

Like maybe an ARM chip (not the GBA's) or something, though of course even that would still be very powerful for hardware meant to interact with 90s hardware.