Nvidia just dropped some news about the latest version of Deep Learning Super Sampling (DLSS), and it looks like we’re moving well beyond image reconstruction and frame generation.
While the older versions were all about tidying up edges and squeezing out extra performance, this latest iteration, DLSS 5, is taking a massive leap into AI-driven territory. It might even be a leap too far for some players who prefer a more natural look.
The big headline here is what Nvidia calls a real-time neural rendering model. Essentially, it is designed to inject pixels with photorealistic lighting and materials on the fly. On paper, it sounds like a dream for anyone wanting their games to look more lifelike, but the early results are already stirring up a bit of a debate in the community.
When you actually see it in motion, the AI-generated lighting can certainly make things look more realistic, but it comes with a bit of a catch. Some of the footage suggests that the AI is being a bit too heavy-handed, changing the appearance of games so much that characters become barely recognisable. It starts to look less like a performance boost and more like a real-time generative AI filter that completely overrides the original artistic intent of the developers.
Despite the potential for things to look a bit artificial, plenty of big industry names are already on board. Major games developers and publishers like Bethesda, Capcom, Hotta Studio, NetEase, NCSOFT, S-GAME, Tencent, Ubisoft and Warner Bros. Games are set to use this tech in their games.
We won’t have to wait too long to see if it lives up to the hype or just looks a bit weird, as DLSS 5 is slated for release this autumn, somewhere between September and November.