Published in Graphics

Nvidia’s DLSS could have a full neural rendering

by on26 September 2023

AI in the gaming business

A future version of  DLSS technology is likely to include full neural rendering, according to Bryan Catanzaro, a Nvidia VP of Applied Deep Learning Research.

According to Toms Hardware Catanzaro mentioned the plan at a round table discussion organised by Digital Foundry.

DLSS has appeared in several different products over the years since it first appeared in the RTX 20-series GPUs. Many have wondered if there was much point sticking the technology like the Tensor cores being included in gaming GPUs. Certainly the first-generation DLSS, and ray tracing for that matter, were pretty pointless.

DLSS 2.X improved the tech and made it more valuable, leading to it being more widely used -- and copied, first via FSR2 and later with XeSS.

DLSS 3 debuted with the RTX 40-series graphics cards, adding Frame Generation technology. With 4x upscaling and frame generation, neural rendering potentially allows a game to render 1/8 (12.5 per cent) of the pixels. Most recently, DLSS 3.5 offered improved denoising algorithms for ray tracing games with the introduction of Ray Reconstruction technology.

Catanzaro said that full neural rendering was a way forward for the tech. Nvidia has been working on this since at least 2018.  During a demo, the UE4 game engine provided data about what objects were in a scene, where they were, and so on, and the neural rendering provided all the on-screen graphics.

"DLSS 10 (in the far far future) is going to be a completely neural rendering system," Catanzaro added. The result will be "more immersive and more beautiful" games than most can imagine today.

Last modified on 26 September 2023
Rate this item
(1 Vote)