Computerphile explains Nvidia DLSS and upscaling via neural networks

One of the big draws for the Geforce RTX series of graphics cards is the DLSS technology – Deep Learning Super-Sampling. In short, it is about upscaling with the help of machine learning, which at best can give increased performance to the same or better image quality compared to playing at high resolution.

Exactly how DLSS works is not entirely easy to understand. SweClocker’s Jonas Thörnqvist described the technology in connection with the launch of the Nvidia Geforce RTX series, and since then the editors have looked more closely at the concept in both Metro Exodus and Battlefield V.

Now the nerd channel Computerphile is trying another review. A Youtube clip explains the idea behind DLSS, and what benefits Nvidia hopes to achieve. Presenter Mike Pound also highlights several challenges with the technology, such as the fact that Nvidia needs to be able to train the neural networks properly – for each game title.

So far, support for the Nvidia DLSS is available in two major game titles, Metro Exodus and Battlefield V. The reception has been lukewarm to say the least, with most being anything but impressive with the image quality of the technology. However, Nvidia promises continuous improvements, including an update to Metro Exodus promises a really refined end result.

Have you tested Nvidia DLSS in Metro Exodus or Battlefield V? Tell us about your experiences in the comments!

More about Nvidia DLSS