That's great to hear! That's actually my primary issue with the PT CP2077 path - while it improves visuals a bit personally it doesn't do it enough to compensate for the large amount of noise it adds. If this takes care of that better I'm happy to hear it. I'm definitely interested in seeing Alan Wake 2 as well as I'm a big fan of Control.I saw it live Just Yesterday where I could Test it on and off in Alan Wake, CP2077 and Portal RTX vs. NRD and other denoisers - DLSS 3.5 really wrecks NRD and any other denoising I have ever seen. Really impressive!
Beyond massively boosting inner surface quality and detail and light propagation from undersampling, it also cut down in Image/lighting Retention and ghosting.
The concept certainly makes sense and I certainly believe it's a place that ML can do a good job. That said, it's hard to imagine this isn't a major chunk of dependencies and inputs/outputs and this time right smack in the *middle* of a frame. Not to re-open this conversation but... this really cannot be the long term solution here. We need the hardware exposed properly so that games and renderers can implement similar passes. I realize NVIDIA would love to just replace game renderers with their own custom one (and with all of these things it is getting close), but this is not a healthy long term direction for the industry, nor gamers.
Looking forward to that!We should have impressions in latest df direct
Last edited: