Both are increasing resolution, but the difference between internal resolution and display (buffer) resolution is whats at stake here, and the difference is huge.
Remember, when Sony and MS were all like "we have ourselves 4K resolution consoles this generation", and then maybe 2 games ran at those resolutions, but - everything did _output_ 4k resolution?
Thats what we are looking at here as well. We want our games to be calculated at a higher internal resolution, which is where all the good stuff is at - we dont care so much about "display (buffer) output" resolution, because thats just the 2D output images (frames) uprezed ("scaled"). Thats the same every TV does - internally at no cost. Its easy.
Getting a game to run at a higher render resolution is much more costly performance wise. (And needs HW acceleration (dynarec), on the Switch, and pretty much everywhere else.)
What does higher internal resolution do - for starters, it creates nicer edge definition on polygon models, so every round surface looks better. But the current emulation cores dont stop there, they upsample textures, apply anisotropic filtering, and when this is then filtered on the 2D output image level, it looks much better, because the initial result a binaural filter (f.e.) gets applied to has much more detail and resolution to begin with.
Also higher internal resolution is the best form of anti aliasing available, also the most costly (performance wise
) - which is why DSR (nvidia "brand" term for upresing older PC games using their drivers, really is much more than just hype.
Look at this Video of Tomb Raider Legend (old game
) run at 4K resolution (with FXAA and 16x anisotropic filtering), and compare it to any normal 1080p output video of the game on youtube. Everything looks nicer.
For PSX the difference is huge as well, because those 3D models had such a low poly count on the original hardware - that even double resolution (720p roughly) makes a huge difference.
See:
Both output at the same 720p resolution, one just scales the display output frames, the other one the internal resolution.
Now display (buffer) output resolution we dont care so much about, thats just scaling the 2D frames that the GPU outputs in the end.