DLSS or FSR support?

Get an AMD GPU and use upscaling on driver level. Together with AFMF, it's as smooth as butter. I have no fps drops, no stutters. Running RX6800 on 5700x3d 21:9 UW 34" monitor, all high or ultra graphical settings.
I've been using AFMF in LMU for a few weeks, never thought of adding Super Resolution into the mix. I'll give it a try, thanks!
 
The likelyHOOD of adding it to an existing engine is indeed very low.
I disagree. With drive towards a console version, that opens a lot of doors that would have otherwise remained closed. Vulkan/DX12, along with upscaling (FSR4?) in some form, seem much more plausible after whatever adaptations they're making for console (think PS5 + PSSR).
 
I disagree. With drive towards a console version, that opens a lot of doors that would have otherwise remained closed. Vulkan/DX12, along with upscaling (FSR4?) in some form, seem much more plausible after whatever adaptations they're making for console (think PS5 + PSSR).
Maybe or maybe not, PS technology is proprietary, which does not mean there is a direct translation to Vulkan or DX12. Or if there is I am not aware of it. Anyway, I would expect for the next game to see some good graphics/performance improvements, for LMU just small stuff shades of it here and there.

PS: I hope am wrong on it.
 
🤯
What you mean is that you hope that they rewrite the graphics engine for console. So you do agree.

Anyway, it would be nice if there were hints as to what they are actually doing.
If LMU wasn't headed for console, I'd agree with you. But with S397 openly working towards porting LMU to console, they're heavily incentivized to optimize this fork of the RF2 engine so that they don't need to degrade the experience (remember how ACC was on console). You can see they're already doing this, by adding occlusion culling in the upcoming March 31st patch (v1.3), and adding texture streaming in v1.2.

These are already major engine modernizations, which I applaud. I don't see them slowing down. DX12/Vulkan opens up proper multi-threaded rendering and async compute, both of which would be valuable for current consoles. Especially with IR similarly modernizing to DX12 soon, I am anticipating that LMU will see that upgrade, and with it, hopefully an option to trade MSAA for an ML-based upscaler. ML-based upscalers simply replace the AA pass. LMU uses MSAA, which has its positive tradeoffs but is very expensive in frame time, and generally loses in terms of temporal stability. It's why you rarely see it anymore (in newer titles).
 
adding "occlusion culling" after it has been standard for 50 years makes you optimistic and even applaud - I disagree! :ROFLMAO:

get me out of here!
 
Back
Top