fbpx

NVIDIA’s Deep Learning Super Sampling Is Getting Smarter

As a gamer who hasn’t cared for anti aliasing ever, supersampling was a nice improvement at higher resolutions. Jagged edges that might show up when AA is off, even at 1080p, 1200p, 1440p, or even 4K, would usually disappear with supersampling. That process was refined by NVIDIA back in 2018 as DLSS, deep learning supersampling.

It’s deep learning because NVIDIA’s method uses AI processing to improve the performance of supersampling on a per-game basis. The fine folks at WCCFTech took a look at how the technology has improved over time, and thanks to both ray tracing with RTX videocards and increased performance at higher resolutions, it’s harder than ever to see the difference of picture quality relative to performance.

What’s always been cool about supersampling is it provides 4K quality at lower resolutions. When lead editor Shawn Sanders told me to try it out two years ago, I didn’t think much of DLSS above standard supersampling. So he told me to turn down my resolution to 1080p, and while I could tell the difference between that lower-res, the performance boost was almost worthwhile.

If you’re squeezing every ounce of performance from your NVIDIA GPU, DLSS is a smart way to do it. And based on the tests run by WCCFTech, it’s definitely improving too. We’ll be testing it more with a new machine coming in soon, so stay tuned to learn more. And read the full article to learn more about how DLSS has improved over the past few years.

Looking to get in on the action? Asetek liquid cooling technology powers several NVIDIA 20-series RTX video cards from EVGA. Check here for details.

Share on facebook
Share on twitter

SHARE

Share on twitter
Share on facebook