It makes a certain kind of sense. The marketing for these products has always revolved around denigrating software emulators as total, inaccurate trash, which is completely unfair for a lot of reasons (byuu used to have a great blog post about it but I think it is gone). But of course, you've gotta convince people they "need" to spend hundreds of dollars on emulation when it's available for free otherwise. Turns out they took the message to heart.
The Pocket is a unique product here. Analogue spent effort characterizing and reproducing the color and LCD response profile matrices of the target systems. Having 10x integer scaling allows for very detailed sub-pixel simulation. It's 2022 and we're getting sub-pixels in our sub-pixels.
Big whoop, right? My 500 W PC with a 40" OLED panel can do that with ease. Well this can do it in a handheld with 7+ hours of battery.
I watched the DF video. I don't watch LTT because, while the topics are often interesting and the resources and effort is usually there, the technical chops are not. Linus never developed a culture of putting heads down and making sure it's right with a thorough technical understanding. There are too many options for me to waste time on creators not creating well enough.
The LTT video[1] is also from their ShortCircuit channel, their unboxing/ first-impressions channel and should not be confused with their review content.
I personally watched the 54 minute in-depth overview from My Life in Gaming[2], and it was clear that LTT/ShortCircuit overlooked or were incorrect in some parts of their initial impressions.
EDIT: Specially in regards to the display, they spent no time at all to experiment with the 3 built-in GBA display modes; if they used the modes that replicated the original GBA or GBA SP (AGS-101) screens, they would've had a more color-equivalent appearance. MLiG in their video spent 10+ minutes on display modes alone.
> EDIT: Specially in regards to the display, they spent no time at all to experiment with the 3 built-in GBA display modes; if they used the modes that replicated the original GBA or GBA SP (AGS-101) screens, they would've had a more color-equivalent appearance.
Apologies, missed that for sure. There could still be a multitude of other factors at play (backlight brightness, saturation setting, sharpness setting, etc.)
> But the broader point that the channel is specifically unboxing/first-impressions and not a review is definitely accurate.
It's a long path to go down. Retro gaming is a boutique cottage industry. It's not at all like audio because it isn't snake oil. People perceive differences in CRTs, sound chips, analog video and audio drivers, etc. Does any of it matter? Does anything matter?
I like to tinker and I'm a display junkie. Playing retro games revitalizes me by bringing my headspace back to when I was young. Little me loves these toys. It would be trivial for me to fall down a cynical nihilistic cliff any moment of any day. If a utilitarian wants an explanation, they can have that one.
The blind tests aren't even the end-all here. The science of human perception limits and characterizing the error are the two fronts to look at. We've long passed recreating what a human ear can hear. We have targets for frequency range, dynamic range, and group delay. Hit those targets and you've won. There is more on the art side of things to advance, namely in simulating directionality, but that isn't what the Tweeter salesman is usually selling you.
The imperfections of the audio chips in old consoles are easily audible to humans. If the goal is accurate recreation then assuming the audio chips and driving circuitry is perfect is insufficient.
Video recreation is still far below what the Human Visual System can detect. We are not masters of the universe yet.
I remember that post; Byuu conceded that FPGAs will always have an edge latency-wise, but was really annoyed that people acted like FPGA magically made things "more accurate" than something like bsnes, saying FPGAs can still be subject to the same bugs or mistakes as emulators, and are harder to develop. I got the impression that Byuu did not like the Analogue people very much.
Hardware emulation is also just as impenetrable as the original systems themselves. Software emulation serves as documentation for the hardware and is therefore much better at preserving that knowledge.
Sticking another emulator under a decompiler is probably at least a couple orders of magnitude easier than looking at an FPGA bitstream. Particularly for the kinds of people who hack on emulators.
With all the stuff like runahead and frame delay showing up in Retroarch I'm not convinced that the latency thing is even true except if we're confining ourselves to low-powered devices.
Yep, actually in my only interaction with Byuu (RIP) they were explaining emulator runahead to me, and they seemed really excited about it, saying that it had the potential to have even lower latency than official hardware.
That said, I will concede that there is definitely an audience for low-power devices obviously, and maybe FPGAs will fill that niche.
runahead is brilliant but it relies on save states being trivially fast to create, which isn't broadly true especially as you get to newer & newer consoles. And strictly speaking needs to be adjusted on a per-game basis as it really depends on the game's internal latency in responding to button presses.
But that also doesn't necessarily offset the latency of getting inputs into the emulator in the first place or in getting video frames from the emulator to the display.
FPGA doesn't really solve your other two issues, nor is it practical for anything post-PlayStation/Saturn/N64, so those objections are not very strong. Past that point the games are designed more with a certain amount of lag in mind anyways.
Are you arguing from a theoretical or practical standpoint? In practical cases the FPGA products do solve my first two issues, as the video output path of the emulation goes directly to the display output in the FPGA. Meanwhile in all software-based emulation products, the video output path goes through a normal Linux graphics stack, which tends to be crap (and not getting better with the silly X11 vs. Wayland fight that keeps Linux's graphics stuck in a bad spot).
That's not an inherent limitation of software emulation, no, but for all practical purposes it is as nobody is doing specialized software emulation for a specific set of hardware to bypass the normal OS/graphics stack. It's instead "throw linux on it, and fullscreen an off the shelf emulator"
> FPGA doesn't really solve your other two issues, nor is it practical for anything post-PlayStation/Saturn/N64
I agree with most of your point here, bus is that particular one true? Doesn't the Gamecube (for example) use a PowerPC chip? I would be very surprised if isn't an effort to make FPGA implementations of the PowerPC.
I realize that there's a lot more to making a clone system than "just recreating the CPU", but why would making an FPGA for something after the N64 be impractical?
As I understand it, the hardware they’re using isn’t up to the task and it would be prohibitively expensive. The complexity is also greater but perhaps that isn’t insurmountable.
I don't know a lot about FPGAs, but I guess I just assumed that they're like basically everything else in tech where they get cheaper and better as time goes on. If that's correct then I don't see why we couldn't have a GameCube or Xbox clone system eventually.
I think this might be more important as time goes on as well. Getting cycle-accurate emulation of anything more complex than the Dreamcast I think will become pretty prohibitive to do in-software. I suspect something like CEN64 might end up being the last cycle-accurate emulator out there (though I would absolutely love to be wrong on that). The advantages of FPGAs would make themselves substantially more apparent at that point than the Gameboy or NES.
As someone who has purchased the some analogue hardware, I will say that a reason I bought the mega sg and super nt is less about pixel-perfect accuracy, and more about buying an appliance.
As a software engineer, I spend all day dealing with bugs. I have a good gaming PC and a PS5, and spend much more time on my PS5 because it always just works. No tweaking settings, no crashes, no compatibility issues, no incompatibility with my controller.
When I get home from a hard day's work, I want to play something on an appliance. Analogue devices are appliances, and emulators (and even the Mister) are not.
I can understand this mindset perfectly, but it bothers me when it's mixed in with confused stuff about how playing on a cycle-accurate emulator is somehow a perversion of the original experience.
It's also worth noting that while people tend to simplify the discussion down to the US, the legal status of software emulation is... a lot more ropey in some non-US territories.
In the UK for example, there is no legal ability to format shift for end users at all, so ultimately you can't really do legal software emulation, at all.
(Do not take this as legal advice that hardware emulation is permitted. It's complicated.)
