This “article” is pretty light on details. What are the actual technical difficulties being faced? Is it a lack of documentation for N64 architecture? A lack of thorough reverse-engineering of the original console? Wildly varying hardware between cartridges? Lack of interest?
It’s that N64 emulation was built around High Level Emulation focusing on plugins and “getting individual games working properly”.
The legacy of this being that there is little to no mature, accurate and performant low level N64 emulation -as is the case with other systems that we now have fast, accurate emulation of the base system.
A significant problem being the way individual N64 titles would often have their own microcode written that differed from what Nintendo provided developers as a base.
MVG is a big proponent of emulators. But even after all his investigating the current state of 64 emus, he comes to the conclusion that you’re still better off buying a used N64.
I’d love to make something highly portable that could run N64 games alongside other retro consoles, so just buying old hardware wouldn’t quite cut it.
One of MVG’s issues is that N64 emulation just doesn’t cut it on portables like Anbernic devices.
And also that cores like those used in Mister or Analogues FPGA don’t cut it - they can start commercial games but can’t do homebrew software (which does work on 64 hardware). So again, have been built to target individual games microcode and not emulate the N64.
He’s definitely lamenting that it’s unlikely we’re going to see improvement in this space due to how much this emulation relies on HLE, but if 64 emu devs can get properly into LLE like PS1 and PS2 emulation did, then it can start to get better.
Decompilation is a solution of course. Those decompiled games can even be run as ports on portable devices (using software like Portmaster), but decompilations are few and far between.
That’s a problem with the current state of emulators, yes, but what are the technical barriers to implementing proper low-level emulation, including handling microcode as mentioned earlier? Imagine that I’m asking this as an embedded systems engineer.
It’s probably worth looking at what Ares is doing with N64 emulation in this case.
Ares aims for accurate emulation (as an overall project goal) and I understand their N64 emu is pretty good. So if there is as lesson to be learned it may be there.
As a hobbyist who’s profession is well outside of tech, this is about as good as you can get from me !
Thanks for the tip, I’ll take a look.
I’ve recently just dipped my toes into emulation, and spent a month writing a decent NES emulator from scratch (we’ll see how well it performs on a ULP ARM core, but I’m optimistic), so I’m starting to think about how far I can push the envelope.
skimmed through the article apart from not playing on the original hardware hasn’t n64 emulator being around for awhile…even to merely to try a game out
There aren’t performant, accurate low level emulators. Everything relies in custom plugins that may be tuned for specific games. The PS Vita should have the horsepower for N64 emulation, but it doesn’t run well at all. Even FPGA cores can’t run home brew that works on original hardware.
There’s so much more to it than just running Mario Kart 64 and Goldeneye.