MiSTer FPGA Forum

Disclaimer: I'm not implying I expect all following ideas to happen nor in same cases that they even SHOULD. Just 'technology daydreaming'...

With the ao486 core just amazing me, I wondered what the Mister actually CAN'T do.

Not for the reasons we'd expect (PS4) from sheer power or complicated headroom, but from other oddities. If it's running a word processor now, a 1949 EDSAC, PDP1, where else can surprise cores one day be inspired from? The Sega Pico one is interesting how the other functionality will be incorporated into it in new ways. It's also odd loading virtual paper tapes in cores.

Limitations like arcade pinball are obvious- even if the electronics and scoreboards can be, it would be hard to play without the physical items' experience without it becoming virtual pinball. I imagine Electronic Battleship would need/prefer physical pieces in lesser regard, but a Speak N Spell or Simon electronic device in FPGA might appear some day where a graphic like the Altair 8800 is used. We have a Vectrex core despite the special screen, and Adventure Vision got made (Nt Mini only for now). Though we have Madrigal's hardware simulators, would screens like Coleco's 1982 mini Donkey Kong handheld venture more into simulation?

MUSIC DEVICES:
Is the Roland MT-32 in FPGA (or MUNT) the only music device you'd like to see? For musicians, would people want 80's keyboard synths, drum machines, or even mix and matching sound chips in other consoles (imagine Streets of Rage having Yamaha, FM or added dual SID chip output).
I have a pointless one I'd still love - a 1978 Mego 2-XL robot. It's no more than an 8-track player with light up led blinking eyes, but there is an online simulator with copies of the tapes to play the comedic quiz games: https://www.2xlbot.com Imagine that Twitch stream.


FMV?

If the DE10 Nano is great with video decoding, what are the technical limits (if any) where CD based FMV games (Dragon's Lair, Sega CD) are possible, but NOT for Laserdisc games (Pioneer Laseractive)? If the Genesis core goes beyond the original spec to include mouse support for lightgun games, could some newly invented "standardized ODE" type of FPGA core be baked into a core to make an otherwise unplayable console work (hopefully that makes sense)? For example, a 1987 Action Max would have the console electronics done in FPGA (two digit score and 'target hit' voice sample), a graphic of a red sensor in the lower right screen corner, and some ability to playback video files for the 5 games.

Would Laserdisc games even be possible assuming one converted analog to digital 'rips' of the games? Action Max is static playback but Laserdisc jumps around. I am excluding FMV games where you ONLY need a VCR, DVD or BluRay player and not lightguns and additional hardware. This site has many console's FMV games: https://fmvworld.com/console_vcr.html

What devices may not be possible due to scarcity, and or no schematics are known to exist?
I imagine the 1985 RDI Halcyon is about one step higher on developer's To Do list than a dishwasher's control panel, but maybe in the spirit of preserving the PDP1 it could be possible if schematics were available? The TWO games (NFL Football LA Raiders vs SD Chargers and Thayer's Quest) were released on CED format, so of course non-scratched up copies for dumped digital conversions would also be needed.

Sega's Time Traveler Hologram game obviously wouldn't be possible, though I'd love someone to make an IBM 5100 core just to make John Titor not need to time-travel back from the year 2034 to 1975.

seastalker wrote: ↑Mon Aug 17, 2020 4:39 pmSega's Time Traveler Hologram game obviously wouldn't be possible

As far as I know, the hologram effect is something like a large mirascope built into the cabinet, and has nothing to do with the board hardware or video signal (aside from any flipping required, and the illusion probably being broken if there's a background filling the frame). So while a MiSTer core obviously can't reproduce the effect, it's in the same sense that the Arkanoid core can't reproduce a spinner; it's simply outside the scope of what a core does.

It's less about what is technically possible, and more about finding a dev who is interested in working on these obscure systems

These systems are not always documented... and the interest in "simulators" can be limited. I believe somebody was looking into a Game & Watch core though (they've been emulated and documented with MAME at least), but no idea what status it is in now.

FPGA music devices exist already but those who made it are not big on open source (they sell it as a whole product).

I wouldn't be surprised if somebody does a VTech VSmile core one day.. it's a toy console with crappy games, but has interesting controllers.

What is and isn't possible generally boils down to a few things. I'm not the most qualified person to answer this, but I'll break it down as I know it:

1) Space. If the design is too big for the fpga, it's dead in the water. After about 75% full, the design becomes difficult or impossible to compile stably. This will include most of the 6th generation and higher systems other than GBA which is relatively simple for that generation.

2) Speed. The FPGA is only capable of having clocks up to a certain speed before it loses stability. Some systems require very high clocks to function, and sometimes because of the configuration of the memory, a higher clock than the original system needs to be used to create transparent memory access. If the speed requirements are too high, then the core can't be done (N64)

3) Memory. Both memory bandwidth (how much can be read at once) and memory speed (how long it takes to get a response from memory) play a role in what is possible. MiSTer has three types of memory: BRAM which is very fast, but there is only around 700kb of it, SDRAM which is either 32mb or 128mb or absent which is fast, but not as fast as bram, but very consistent, and ddram which is 512mb of for the fpga, and is very fast for burst reads but very inconsistent for single byte access read. Each type of ram has it's pros and cons and can be challenging, but this can also limit what cores are possible. This was the major challenge in the neogeo core for instance. If someone were to attempt a Saturn core, this would likely be the largest consideration.

4) Special hardware. If there is a special display, servo, or other bit of hardware that is impossible to emulate in any way, then this is an obvious limitation as well. Most things like vector displays, tapes, or punch cards can be emulated through various means though. Laserdiscs are digitally encoded composite which can be easily decoded on the fly (we did some of this when working on the tg16 palettes). VHS can be handled similarly if needed, should digital rips of these things be available. Things like a hardware MT-32 are possible, if someone spends the time to make it, but that would be a fairly huge task, and also very unlikely to fit into something like the 486 core which already uses almost all the usable space.

5) Out of Scope. That said there's also things that SHOULDN'T happen because they are out of scope. By this, I mean you wouldn't want your microwave making ice cubes, because that's not what a microwave is supposed to do and that's not what it should be designed with in mind. Adding an ice cube maker to a microwave would almost certain add a lot of cost , size, and complexity to a device that shouldn't have had it.

In the same token, MiSTer has a pretty clear role as a device that emulates old computers, consoles, and arcade games on an fpga. It doesn't have the hardware to be a good software emulator (no video acceleration and a processor about like a pi zero). It doesn't have the hardware to read carts (not enough GPIO, cores aren't written for it.) It's not meant to be a video player, as it can't read the old media and there's better more purpose made devices that can do this much better. It's not meant to be an upscaler like OSSC, etc, etc.

Hope that helps clarify what limits there are on the de10's hardware.