Don’t take Nvidia’s word for it — try Nvidia’s awesome new image comparison tool instead

Nvidia wants you to have it away that AMD's FidelityFX Super Resolution (FSR) is a plain pseudo, nothing like the Nvidia Low-pitched Encyclopaedism Super Sampling (DLSS) proficiency its GPUs can use to boost the framerate and/or image timber of your games. To prove IT, Nvidia is trying an interesting new tactic aside releasing a remarkably herculean image comparison tool called Nvidia ICAT, for free, so that anyone bum see for themselves.

ICAT, short for "Ikon Comparison And Analysis," is a very acicular program that lets you regar a push-down stack of gameplay screenshots and even tape-recorded videos side-by-side, automatically lining finished the frames, and allowing you to fluidly pan across for each one scene, zoom in and out, on whol of those images or videos at the same time. It's got a draw-and-drop interface that makes pixel peeping child's bid, and as known pixel peepers, you'd advisable believe The Threshold is exit to be using it for to a higher degree right gaming graphics!

Nvidia's new ICAT tool in action. There's as wel an epitome slider mode.

Screenshot by Sean Hollister / The Brink

But Nvidia's hope, once again, is to reveal how poor AMD's FSR is compared to DLSS — or even Nvidia's own, separate spatial upscaling technique.

Just in case you necessitate a brief primer: Spatial upscaling techniques, like AMD FSR, run each of your biz's TV frames one by one direct a fixed algorithm — and they don't necessitate a special GPU to run. Nvidia's DLSS, on the other turn over, is a temporal upscaling technique that compares multiple frames and takes score of how things are taking possession a video game tantrum, and processes all that exploitation a neural meshing that runs exclusively on the Tensor cores you can only find in an Nvidia RTX GPU.

Nvidia illustrates how temporal and spatial upscaling differ.

But before you say "that's great Nvidia, but I can't actually buy your new GPUs!" you should probably know that Nvidia has its personal spatial upscaler as well, dubbed Nvidia Image Scaling. It's apparently been buried in the Nvidia Moderate Panel for some meter now, and now Nvidia is open-sourcing it on GitHub with its own SDK and support for every brand of GPU. Developers can natively integrate it into their games if they want, hardly like AMD's FSR.

And, the society's baking Nvidia Image Grading into its GeForce Feel for app so you can turn it on for whatsoever game, and adjust an in-game sharpness slider (so you can see the difference) exploitation Nvidia's sheathing.

Is it as good as Nvidia's DLSS? Non even incommunicative, and Nvidia's the offse to admit it. In a briefing with The Verge, product handler Joseph Henry Lin hammered us with lesson after example of how DLSS (especially the untried DLSS 2.3, which suffers from less ghosting around moving objects) not only beatniks the pants off of Nvidia's own spatial upscaler, but sometimes performs better than a pure image. Particularly when you're looking at thin objects with lots of edges that tend to glint:

The envisioned communications dish is one place DLSS has an reward.

Image: Nvidia

In person, I've sworn by native resolution all the way — simply Nvidia's ICAT made Pine Tree State queer sufficient to put under it to the test. So I fired up Deathloop and Back 4 Blood, two games that offer both AMD FSR and Nvidia DLSS, on a 4K screen at their highest quality modes.

In Back 4 Blood, I felt vindicated. All part of the image looked fuller, crisper, and high resolution at inborn 4K, up to and including the roof-mounted antenna:

You'll want to water faucet to zoom and/operating theatre download, this is a 4K image.

Screenshot by Sean Hollister / The Brink

There were evening parts of the scene where AMD FSR clearly beat Nvidia DLSS, like all of these wooden textures:

Maybe Nvidia's neural net needs more experience with woodwind instrument.

Screenshot away Sean Hollister / The Verge

But when I tried out Deathloop, it was the opposite story: almost everywhere I looked, DLSS was resolution and conserving inside information I couldn't even see at native resolution, and with far, far inferior glint from loosely knit aliased objects.

In Deathloop, the VW busbar in the background looks decidedly less cartoony with DLSS, and you can see outlying many point in distant textures.

Screenshot by Sean Hollister / The Verge

Native 4K, blown up to show detail.

GIF by Sean Hollister / The Verge

DLSS Quality, at the corresponding size. Less flitter, more detail.

GIF by Sean Hollister / The Wand

That said, I still byword moments in Deathloop where DLSS does weird things, too. Whorl up to the top of this post for one example: you can understand how DLSS totally metamorphic the sparkly finish of the Strelak Verso pistols and muddies the texture of the dirt ground in the mathematical same scene as my other screenshots and GIFs.

I'm not a believer in DLSS quite yet, but I'm intrigued. Nvidia's call is that, with enough learning, games sack actually wait better and run faster than native with DLSS turned on. For now, it's something developers have to enable on a per-game foundation, and not all games consume the same version to offer, merely that could change down the road. Nvidia's Lin tells me that Cyberpunk 2077 already features over-the-air DLSS updates — when you open the spirited, it can check with the Nvidia driver to see if there's a unweathered version and hot switch it at launch.

"Wherever you see DLSS not correspond native, that's what we'll be working happening," says Lin.

In the meanwhile, here's a unweathered video recording from Nvidia showing what DLSS can look like at its incomparable.

Update, 12:34PM ET: Added Nvidia's video.

Don't take Nvidia's word for it — try Nvidia's awesome new image comparison tool instead

Source: https://www.theverge.com/2021/11/16/22784404/nvidia-icat-dlss-spatial-temporal-upscaling-graphics

Berbagi :