Nvidia promises more accurate video with its AI upgrade

Raw performance is of course key when it comes to buying a decent graphics card, but in recent years Nvidia and AMD have shifted their focus on associated software features in and out of games. From Nvidia, for example, there is the Broadcast package, which can reduce microphone noise in real time and replace the background in a video stream from a webcam.

► Quick Quiz: RTX Voice meets Kenneth Drums

During the ongoing CES, the Green Team is lifting the veil of a novelty that falls under the broad RTX umbrella. The name is Video Super Resolution, which is, in short, an AI-based upscaling feature for streaming video. The upcoming feature is very similar to the upgrade which is part of the DLSS package, in which artificial intelligence is used to create a higher resolution image based on a lower original resolution.

With DLSS, the upscaling is something that allows for a higher frame rate by essentially running the game at a lower resolution, but in the case of RTX Video Super Resolution it’s just about improving image quality. Nvidia believes the technology partially handles the compression elements, and partially provides significantly more detail in the image compared to a classically upscaled image.

A shorter video shows the difference between upscaling 1080p to 4K UHD and when doing the same upscaling with RTX Video Super Resolution. Judging by the images, the lift could be clearly noticeable, something that all owners of Geforce RTX 3000 and RTX 4000 graphics cards could consider in the near future. The functionality will be released sometime in February and will be available for free, but initially only works with the aforementioned graphics card and Chrome and Edge browsers.

Read more about Nvidia at CES: