I'm not so sure, this sounds more like removing the calculations not required as shadows and or rays don't have fluid movement in the first place.
I'm with derek on this. Shadows should be as fluid as the player. However, given I hardly have time to notice these shadows, I'm not surprised halving it or limiting it to 30fps was considered an option.I'm not so sure, this sounds more like removing the calculations not required as shadows and or rays don't have fluid movement in the first place.
If an object is moving so fast that it visibly outpaces its shadow in 1/30 of a second, it's moving too fast for you to see clearly . We need high FPS in games because our perspective of the game world can change rapidly, not because the game world itself changes rapidly.
Yes, in games with unlocked frame rates this could happen, example: shadows and reflections at 30fps on a game that runs at +120fps could be noticeable, but, on games with locked frame rate could be a great addition if its a option that the user can activate, more modest hardware could gain performance, specially first generation RTX Cards, example: Fighting games or any other game at 60fps, a 30fps RTXGI could help a lot on modest Graphic cards like the RTX2060. 1 of every 2 frames shouldn't be that noticeable.
Would look better on Turing and Ampere because those architectures have the RTX Cores which made them process RayTracing faster without having to use the base cores. Is a dedicated part of the hardware to do those calculations. Which until now not even one AMD card have. RayTracing is possible on current RTXless hardware, but the performance hit is too high to be viable.Even if the game only ran at 60fps a 30 fps update rate on shadows would be highly noticeable and it would only get worse the higher the frame rate gets. Hell it wouldn't even look right in a 30fps game.
Also since this is software why would it look better on turing or ampere than anything else? Surely big Navi would look better than turing and close to ampere.