I’ve seen a lot of people lately saying that upscaling (fsr, dlss, etc.) is a bad thing, including some calling it ‘fake frames’, which is probably due to them confusing it with frame generation.

What upscaling does is take an input (a frame rendered at 1080p, for example) and attempt to improve it by generating more information (bringing that 1080p frame to 1440p). this does make things a little fuzzy, but it also frees up resources to allow stuff like improved lighting to be rendered which makes games like cyberpunk able to be rendered at a decent framerate without a $5,000 gpu.

Frame generation is different. It takes an input as well (same 1080p frame, for example), but it doesn’t improve the frame. It makes a new one based on that frame, sometimes several. These actually are ‘fake frames’, and this is what the people who called upscaling fake frames were really talking about.

I won’t lie, upscaling is definitely a crutch and the goal should be to be able to render that cool stuff at native resolution. however, the tech that can render that stuff is too expensive to be worth buying unless you have money to throw away, which real people typically don’t. it’s up to you whether a little fuzziness in the graphics is worth it to you, but the fact is it’ll give you the leeway to choose between higher framerate and prettier lighting. without it most people are stuck just setting their graphics to ‘no’, because they can’t afford the kind of processing power making things look good at native resolution takes.

Part of why I am making this post is that I wanted to see what other people think of this take, and more importantly get feedback so I can improve the take later. I’m currently running a laptop with a 1650, and I’ve had it for years. I’m used to balancing frames and quality and making compromises, and upscaling tends to be one of them that’s worth making.