Lecture 23: Virtual Reality (98)
Franklin-Zhang0

I think maybe there's another way to solve this problem, which is to first generate relatively low-quality frame and then use super resolution technology to reconstruct the high-quality frame. This technique has been widely applied in the gaming industry, e.g. the FSR of AMD and the DLSS of Nvidia. Therefore, maybe it can also be used in VR.

keeratsingh2002

@Franklin-Zhang0 I think that's a really interesting point! DLSS/ FSR'S frame generation is popping up in a lot more triple A games and it seemingly works really well. However, as mentioned in the slide and speaking from personal experience, the additional frames introduces quite a noticeable amount of overall system latency. I can't image how that would carry over to using a VR headset, as experiencing latency delay over a controller or a keyboard/mouse must feel a lot different from experiencing it while swiveling your head around in a VR simulation.

omijimo

in my experience, frame generation does add consistent input lag. since we want to minimize input lag for vr headsets, i feel like it would be a better tradeoff to have slightly lower image quality for better input latency

You must be enrolled in the course to comment