Because AMD has been milking their fanboys with software heuristics. Thats all Ive seen. Inferior hardware for 2 generations. First no RT hardware then no ML upscaling hardware. No DLDSR (could not live without). Go ahead. Hate on me like I have no point. Good luck on your next hardware adventure.
Going to be cool to see how this is used if it turns out to be true. It doesn't change the fact that your still wrong that correcting the visual errors in FSR 2 requires hardware. As those errors are already correctable in regular TAA without specialized hardware. https://aws.amazon.com/de/blogs/gametech/anti-ghosting-with-temporal-anti-aliasing/ https://www.elopezr.com/temporal-aa-and-the-quest-for-the-holy-trail/ *edit* I think that's all we got stuck on the last time we went back and forth on this. It's not that hardware is bad or that AI is bad(it's not). Just that neither is needed to correct how FSR 2 looks. The proof of that is in versions of TAA that have managed to correct those issues seen in FSR 2.0.
You would just code a fallback to something like FSR2 for hardware without the functionality That's assuming it's used like we are expecting.
If AMD is smart, they just build it into FSR3 and not require the developer to do anything. It just automaticly switches based on hardware detected.
Fair point lol XeSS is doing something similar. If you have the hardware it will use it, if you don't it uses another method to accelerate it. So it might not be FSR 2.0. Was just an example.
Well, AMD gotta create some incentive for people to buy RDNA3 and that is FSR3.0 FSR2.0 is just a freebie that AMD just toss out like FSR1.0, no further update, it's up to modders what they do with FSR2.0 LOL
