Perhaps it’s a coincidence, but it seems somewhat suspicious to us that Nvidia decided to announce its new DLSS 5 upscaler on the day Sony revealed plans to roll out its new PSSR 2 upscaler.
The highly anticipated upgrade for PS5 Pro, a collaboration between PlayStation and AMD as part of its Project Amethyst alliance, has generally been attracting rave reviews.
Tech experts Digital Foundry already waxed lyrical about the implementation in Resident Evil Requiem, while it’s now heralded the improvements in Silent Hill F as “transformational”.
Watch on YouTubeSubscribe to Push Square on YouTube168k
But AMD has been playing catch-up to Nvidia for quite some time, with its proprietary FSR solution only recently gaining ground on DLSS.
Today’s updates to PSSR don’t necessarily move the technology forward, but instead clean up many of the issues exhibited by the original implementation. In almost all of the examples shown so far, we’re seeing a cleaner, more stable upscale.
DLSS 5, which you’ve no doubt seen by now, attempts to take that technology to the next level.
We’ll defer to Digital Foundry’s excellent expert analysis for the nitty-gritty, but Nvidia claims it’s retaining original assets and implementing photo-realistic lighting to achieve truly next-gen visuals.
There’s no doubt that this is a seismic change for video game presentation; you won’t need to zoom in to see the difference. But many are already beginning to question the impact on artistic integrity.
A comparison of Grace from Resident Evil Requiem is perhaps the most egregious. While it apparently is the same model under different lighting conditions, her eyes appear bigger and lips fuller on first glance.
It does, at least to us, look like the character has been run through some kind of Instagram beauty filter.
And the response has not been particularly kind on social media.
Gamers Nexus, a well-known tech YouTuber, seemed to sum up the sentiment best:
“DLSS 5 makes this look like an AI generated dating profile picture used to scam an old person in another country. Just looks like every other AI generated image of a ‘person’. No character or soul to it. Art loses what makes it impressive when it all looks like generated slop.”
We’re hesitant to go as hard because, effectively, what Nvidia’s demonstrating here is still a work in progress for a product not due to ship until the end of the year.
But in response to the initially negative reaction, it has already started to do some damage control.
A pinned comment added to its YouTube trailer roughly an hour after it was uploaded attempts to quell concerns:
“Important to note with this technology advance – game developers have full, detailed artistic control over DLSS 5's effects to ensure they maintain their game's unique aesthetic. The SDK includes things like intensity, color grading, and masking off places where the effect shouldn't be applied. It's not a filter – DLSS 5 inputs the game’s color and motion vectors for each frame into the model, anchoring the output in the source 3D content.”
So effectively it’s saying that game developers will still retain control over how DLSS 5 is leveraged, and the outcome may not be as egregious as what we’re seeing in these initial side-by-sides.
It should be added that this technology, as it’s been announced today, is reserved for Nvidia’s 50-series line of graphics cards – state-of-the-art products that cost considerably more than a PS5 Pro on their own.
Therefore what we’re seeing here is a taste of the future – or, at least, Nvidia’s vision of it.
Ultimately, AI-based upscalers are here to say. Existing versions of DLSS, FSR, and Sony’s newly upgraded PSSR are posting remarkable results, producing clean, artistically-intended images generated from less computationally intensive sources.
DLSS 5, however, at least on the evidence shown so far, may be taking things too far.
It seems that Nvidia, in its attempt to one-up AMD today, may not have received the reaction it entirely anticipated.