YouTuber SirSethery recently posted a video demonstrating the differences between playing GBA games on all official Nintendo handheld consoles. As with all analysis (including my own!), I’d like to see these tests repeated by other members of the community before we call this “fact”…but Seth found some very interesting performance differences between the handhelds! Definitely check out the video for more info, but there’s a few things I’d like to discuss here.
The video does a great job showing the differences in screens and I’d completely forgotten about the voltage potentiometer trick on the SP’s. While both of those have been covered before, it made sense showing them in the context of this video…but what really interested me was the performance testing. First, it seems both the DS and 3DS run GBA games 0.12% slower than original GBA’s. While this can be considered “zero difference” in most scenarios, it’s absolutely something speedrunners need to know!
Seth also dug into overall lag testing and found some interesting data. Now, I need to start by saying these aren’t the most accurate lag test measurements, but they’re very close. I spoke to Seth and he’ll be upgrading his test rig to an LED-based version, however there’s one important thing to note: The current findings are consistent with each console measurement and while they might be a few milliseconds off, my gut feeling is the differences between the consoles are accurate. As always, I hope my fellow nerds jump in and confirm the results…but let’s take a look at what’s there now:
Seth’s tests show that the DS and 3DS consoles have just about a frame more latency than original GBA consoles. This is interesting, as I always assumed 3DS consoles ran GBA games in emulation, however it looks like both the DS and 3DS accomplish backward-compatibility through a hybrid approach. If it weren’t for the 3DS comparison, I’d say maybe the DS screens are just laggier, but that doesn’t appear to be the case.
One thing to note, is Seth found the GBA IPS “v2” kit to have more than twice the latency of the original screen. While I’m sure this is true for the panel he tested, there are so many of these on the market and each one could vary. I think it would be safe to assume most of these IPS kits need to buffer at least a frame to scale the image, however this looks like at least two full frames of buffer. I’d love to see more testing done on each kit to see the difference. Anyway…
Another interesting find is that original GB and GBC games seems to have a frame more latency on the GBA SP and yet another frame on top of that via the 2DS XL’s emulation. While the emulation lag makes sense, I’m really curious what’s causing it on the GBA – Is it the screen itself? Some sort of framebuffer to make it compatible with the GBA’s screen?
As I said at the beginning, we need people to verify Seth’s findings with the most accurate measurements possible. I can only go as far as using an LED and 1000fps camera, but I’m wondering what other methods exist that could be even more accurate? I’d also like to see how the GBAz and other video-out kits compare on low-latency displays! I guess this is just another sign that even all these years later, we’re still learning new things about classic consoles and need to keep testing!
