Nvidia Announces DLSS 5, and it adds... An AI slop filter over your game
3mon 3d ago by piefed.ca/u/popcar2 in games from www.digitalfoundry.net
I'm completely speechless. This looks so terrible I thought it was a joke, but apparently Nvidia released these demos to impress people. DLSS 5 runs the entire game through an AI filter, making every character look like it's running through an ultra realistic beauty filter.
The photo above is used as the promo image for the official blog post by the way. It completely ignores artistic intent and makes Grace's face look "sexier" because apparently that's what realism looks like now.
I wouldn't be so baffled if this was some experimental setting they were testing, but they're advertising this as the next gen DLSS. As in, this is their image of what the future of gaming should be. A massive F U to every artist in the industry. Well done, Nvidia.
Check out this fun little nugget from further down in the article:
Nvidia actually used two RTX 5090s for its demos: one plays the game, the other exclusively runs the DLSS 5 technology.
An entire second GPU just to run it.
I think you have a typo there: An entire second GPU just to ruin it.
Can hardly afford or find one, they expect us to buy two? Fuck them
They have no intention to sell them to us. They'll maybe let us play it like this thorugh GeForce Now, or any other streaming service.
They don't want regular folk to buy PCs anymore.
So, yes. Fuck them indeed.
No. They want regular folk to buy PCs. They just have no idea what regular folk can afford. How much could a banana cost, 10 dollars?
Yeah, their sales team is incredibly fucked over from their only significant revenue being various AI farms.
Is there scarcity somewhere? Priced out completely, but they're available.
That's actually good news in my eyes, we definitely won't see this hit the consumer market for years.
Context is important. The following sentence is:
The use of two GPUs is required right now as DLSS 5 still has a long way to go in terms of optimisation - both in terms of performance and its VRAM footprint. However, DLSS 5 is designed for use on a single GPU and that's how it will ship later this year. Quite how scalable it is also remains to be seen, but in common with other DLSS technologies, Nvidia tells us that the computational cost scales with resolution.
Sure just double the vram and your AI can run, unfortunately you can't afford that vram because billionaires are running AI.
It's always funny to me when Nvidia releases new 8GB cards and new VRAM-heavy features as a reason to buy them.
8gig isn't even enough for 4k right now. There's a guy that took his 3070 up to 16gig and it really shows the cards are ram limited. And also it's a serious pita to change vram without melting the card.
But still the idea of DLSS was to claw back performance lost to raytracing, right? This is the exact opposite of that, it costs performance to sloppify the game. I just pray it's gonna be an optional feature in games and I can still use DLSS 4 instead.
2017: Buy 2x 1080 Ti for 1500 bucks, your build is GOATed, have fun spending half your time tinkering with your overclocks and fishing for the perfect SLI compatibility bits in Inspector
2026: This. It's a shame
SLI is back!
That was my first thought too lol
Ha ha thats great. Reminds me of PhysX back in the day where initially it needed a dedicated card.
Some of us are old enough to remember when your games would sound different depending on which sound card you had installed.
Ahh, I remember the first time I heard the intro music to Star Control through my Sound Blaster instead of through my motherboard's piezo speaker. Like the audio version of The Wizard Of Oz switching to colour.
Sound Blaaster!!
IRQ 3 DMA 8 bit (or what was that now again)
Ima bus out ma voodoo2 soon
This is shifting away from performance for dlss.
Features that need 2x the GPUs makes the stock go up 2x
Back to the ol pysx GPU days. Except physx made the game cooler.
Great the new SLAI.
Good, so my hardware won't support it. XD
as much as you and I both hate it and as shitty as it looks right now, i imagine that some sort of cloud hosted ai technology is the future of gaming
It sounds like indie games that will run on a potato and not require internet access will be the future of gaming. I’m completely done with aaa corposlop.
sounds like a good idea. it sounds like that's what I like doing and that's what I like playing and it sounds like it's good but I don't know man. you tell me that sounds good. I'm glad you predict the future
i hope that the universe will happen exactly as in this website In accordance with our upvotes
They tried this shit with Stadia. What would be the difference now that we also have AI? There is no reason to stream a shitty AAA game from a shitty company when you can buy and play indie hits on your own hardware.
I know the capitalists really want game streaming to be the future but every gamer with some common sense will reject that idea.
Makes perfect sense in context.
Nvidia is no longer a gaming company. They do not care about gamers. They are actively DRIVING AWAY gamers as fast as they can, because AI chips are more lucrative than dirty smelly tiny consumer products.
This is meant to show off real-time AI generation performance, which is massively marketable to AI companies. Not consumers. Nothing is about consumers.
oh i’m sure the "hire fans lol" gamer chuds will be over the moon about this
This looks so bad, like uncanny 3d porn bad.
Yup. You will soon pay premium for checks notes actual footage of people enjoying fucking?!
New Worms game looking hype though
Can't believe DF is so positive about this, just looks horrible. I've actually been quite positive about technologies like DLSS and FSR but this... no thanks.
Crazy to me to see people just now waking up to just how underinformed Digital Foundry really is to technical details, and how much they sold out.
Well, helps that you don't need to know how to turn a computer on to see that they're glazing some complete bullshit, just eyes.
They absolutely got a huge nvidia check for this one
The amount of glazing they do, holy fuck. They had me too for a minute before I started wondering if this is just an april fools joke
First time you’re witnessing nvidia coverage?
DF always glazes the industry. UE5, DLSS...
Eh, DLSS maybe, but they've been quite critical of UE5, especially its stutter issues. On the contrary, they're always happy when something isn't UE5.
My issue with UE5 is that it runs like shit, which DF tends to agree with, but it also looks like trash. Temporal effect are all over the place and make the image look like a muddy, ghosting mess. It only ever looks good on static screenshots
Idk why, but this reminded me of the South Park episode where everyone is using photoshop to show off their girlfriends.
Had to make the meme
The demos run on the back of two 5090 GPUs... At what point do we even acknowledge I shouldnt need to be using nearly 2kW of power on a gaming PC on a dedicated circuit in my home? Simply diminishing returns at a certain point and nvidia is bringing us way past that point into very unreasonable territory.
Is it cool they could do it at all?... I GUESS?! But they've definitely lost the plot here.
I just watched the Digital Foundry video on this and the way they fuckin glaze over this tech just disgusted me enough to unsubscribe.
Can't say I ever really trusted any Digital Foundry takes. They all seem to naval glaze over shiny things.
DLSS 6: You don’t even need to design your game. AI will generate it in real time.🤡
Was able to brace myself enough to skim through the video. Anyone that watches those clips and thinks that looks better has no place working on anything related to visual tech. Not only is the interpolation obvious and distracting, every scene, no matter the time of day, location, or anything else, has the exact same shitty lighting that washes out every shadow and color painstakingly added to the scene by actual professionals. Pro-tip: A white-balanced image is the starting point, not the end-goal. Making every game look exactly the same is fucking terrible, you hacks.
I can only hope this shit only finds its way into the AAAA dross that's not worth playing already.
Hmmm I also just thought of how this is it looking the "best" for the demo and that most of the time it will just be the makeup gun set to "whore"
Isn't DLSS by defintion always an AI slop filter?
Not really, DLSS mostly just reduces the resolution of a game and then upscales it back up. It does a pretty good job of making the game still look (almost) exactly the same. This, however, completely changes what you're looking at.
DLSS is short for Deep Learning Super Sampling, it does the upscaling using deep learning, what people also call AI. The upscaler has to be trained on images. Depending on how you train it you either get something that looks almost exactly the same as the game at a higher resolution or you get AI slop.
I'm aware of how it works, but the results aren't bad. Worst case scenario is you get some ghosting with DLSS, but it's far from what I'd call AI slop.
But it literally follows the same process. Why is one slop, but not the other? You’re being hypocrital.
One is upscaling the image while preserving it as much as possible, the other is applying a filter to try and "enhance" it by drastically changing the image and ignoring artist's intent. What's hard to get?
This isn't applying a filter, it's applying running the image through a transformer network trained on advanced lighting methods like subsurface scattering to make materials more lifelike. It seems to change artistic intent quite a lot on these existing games, but frankly I'm excited to see what creators do with a game designed from the ground up to utilize AI-enhanced lighting. The DF video also states that this is an early preview (hence the dual 5090s) that is expected to change over time.
it’s applying advanced lighting methods like subsurface scattering to make materials more lifelike.
It is not. It is approximating the results of training data consisting of output images that have been rendered with subsurface scattering. It isn't actually running the subsurface scattering algorithm.
Well that just sounds like subsurface scattering with extra steps!
.... this is AI we're talkin about, literally everything is trained. I thought that would be assumed, sorry for not being clear enough.
It's not. it's a completely different set of steps (at least at runtime). The Venn diagram circles don't touch.
For the joke to land it has to make sense in context. Otherwise it's just random references.
I'm well aware of the meme. You used it inappropriately.
If it was made for that the slopifier would be able to identify the light sources. Before that it is art and environment destroying irrelevant bullshit. From all the slop examples, the best Nvidia can deliver, it is shown that they ignore the lighting of the scene.
How is "upscaling while preserving it" not the exact same philosophy as "enhance by applying a filter?"
You just don't like the specific filter, it's very literally the same process.
Because a pixelated circle being upscaled is a circle, but a pixelated circle being turned into a high definition pie is no longer a circle, and that's especially problematic if the circle was just a cross hair or some other random circle like thing the AI thought was meant to be a pie.
Yes, both things are the same, but that's like saying you had a tiny spider in your house and you were okay because it killed mosquitoes in your house, so you should be okay with having a colony of bats since they are also animals and eat mosquitoes. Yes, both are the same, but the scales and the amount of intrusion are completely different.
If your training data has a pixelated circle as an input and a circle as output, your neural network will "upscale" your pixelated circle to a circle. If your training data has a pixelated circle as input and a high definition pie as output, your neural network will "upscale" your pixelated circle to a high definition pie. Even if it's the same algorithm in both cases.
Yes, that's precisely my point. The difference is in what the algorithm is trying to do, traditional DLSS uses the image rendered in resolution X as output and scaled down to X/2 as input (for example), so it's trained to upscale images, whereas this new thing uses who knows what as either, and clearly outputs something that is not an upscaled version of the frame.
Current DLSS intent: We can only render this at like 720p with enough frames, so let's do that and use AI anti-aliasing tricks so that when we present it at 4k, none of the jaggies are visible on-screen like they would be with raw 720p upscaling.
DLSS5 intent: Using our pile of stolen artwork neural net that we can now render at 60fps+ let's "reimagine" the entire look of the game as we present it on-screen, even if it was already running at 4k just fine.
TLDR; How big the neuralnet is and what your train it for matters.
Ideally you'd have a DLSS-like system trained specifically trained for only one game instead of a general system. Then you can train it on 4k with highest settings and you should get something that doesn't mess with the style of the game.
You're describing what DLSS 1.0 was I believe
Yeah, but they did that for like two games.
Yep. Maybe it could actually be "modules" that the individual devs submit with their game, essentially.
... How if flying a spaceship different from driving a car? They're both controlled applications of kinetic energy to move people or objects.
At the end of the day, it's all a pile of transistors and the only thing that is of import is the intent behind usage.
In one case it's saying you can use a neural net to take something rendered at resolution A/4 and make it visually indistinguishable from the same render at resolution A.
The other is rendering something and radically changing the artistic or visual style.
Upsampling can be replicated within some margin by lowering framerate and letting the GPU work longer on each frame. It strives to restore detail left out from working quicker by guessing.
You cannot turn this feature off and get similar results by lowering the frame rate. It aims to add detail that was never present by guessing.
Upsampling methods have been produced that don't use neural networks. The differences in behavior are in the realm of efficiency, and in many cases you would be hard pressed to tell which is which. The neural network is an implementation detail.
In the other case, the changes are more broad than can be captured by non AI techniques easily. The generative capabilities are central to the feature.
Process matters, but zooming out too far makes everything identical, and the intent matters too. "I want to see your art better" as opposed to "I want to make your art better".
What…? It’s more like chemical vs nuclear rocket. You’re not even comparing the same thing while these are both the same things with different views. You don’t like this one, so suddenly it doesn’t met your arbitrary conditions to be acceptable, so now you’re coming up with incorrect analogies to try and make a point. Great job!
And you didn't even read past the first sentence I see.
Saying they're the same because they both use a neural network is roughly equivalent to saying things are they same because they're both manipulating kinetic energy.
Not all answers are easy. This new dlss looks like it was trained on stolen work. Old dlss had a neutral network that was tuned before the plagiarism machine became popular.
Piracy is not stealing.
It is when it's used by corporations for profit, IMO. Not for individual private enjoyment.
Oh yeah? Well vegatables are both in pig troughs and on dinner plates. Why's one slop and not the other? They were grown with the same process!
Because one is shitty and the other isn't.
If the vegetables are the same, they aren’t slop. Pigs aren’t fed vegetables, they use rotten vegetables. Your analogy doesn’t work, if you actually comprehend the basics of it….
If the vegetables weren’t rotten, yeah most people would eat the “slop” since it’s just vegetables, you would let good food go to waste just because the “name” you’re arbitrarily and incorrectly using for all pig feed?
Are you really asking why compressing and uncompressing art made by a human being is different from slop produced by the slop machine?
One exists to reconstruct an image as closely to the original as possible while saving space, the other is meant to insert arbitrary changes to the initial image and produce something else.
I don't like AI but christ Lemmy is getting annoying lately with kneejerk "slop" claims for anything with the letters AI in it. A lot of this stuff has been used for ages and yeah, they're leaning into the current hype but the over reaction is just ridiculous (see: the "open slop" list of open source projects that includes those that have the audacity to allow developers the ability to use AI line completion)
It genuinely diminishes actual concerns with AI tech when people are losing it over things that have existed long before the current bubble but just have AI™️ on the package now
deep learning isn't really the same thing as a large language model. People call LLMs AI.
Llms aren’t the only type of ai….
AI isn't real, I'm just saying what people call AI is pretty much LLMs or anything that does NLP. No one looks at DLSS and says 'thats ai'
"Science fiction AI" isn't real. AI is most definitely a thing. From the Oxford dictionary:
artificial intelligence = the study and development of computer systems that can copy intelligent human behaviour
By definition, a chess program is AI.
pretty irrelevant to my point, honestly
AI isn’t real
It is real. Better like this?
Your point being you’re making up your own definition of ai…?
you're out of touch with your neighbors
LLMs were (and are) marketed as "AI".
I agree. DLSS isn't, though. It's not AI though. Deep learning is like a close cousin
DLSS actually uses Machine Learning models to do the upscaling, so in fact there is no AI Slop here.
Not really,
Nvidia just calls everything DLSS...
Like, it's basically an anthology label at this point. If they think it's a good idea, they call.it DLSS #
For example DLSS 4 was frame generation, nothing to do with super sampling.
You could call it temporal super sampling.
It does a pretty good job of making the game still look (almost) exactly the same
Isn't that just displaying the image with extra steps? Why is my PC using all this extra processing power in order to make it look (almost) exactly the same?
I think that's accurate. It's making something out of nothing, which will certainly be graphics but not necessarily exactly what the game is supposed to look like.
No, and if that's your opinion you don't know what DLSS is
While it may have used machine learning, it was definitely not in the 'slop' category. I generally think of slop as things which try to imitate some kind of creative or human element (like the enhancements from DLSS 5), but FSR and earlier DLSS used machine learning to replace anti-aliasing like MSAA, etc., through super-sampling and temporal technologies (frame gen kinda sucked though). So, to answer your hopefully literal question, DLSS has, in the past, not been a AI slop filter.
Yes, jensen huang recently tried to defend it.
Is the DLSS AI slop going to lead to homogenization of game aesthetics? That's really fucking lame.
Incel Shading
it's incredible that this can be done in real time on a consumer GPU.
this is also stupid and completely useless
Won't it be kinda strange for gamers who buy a game and find that none of the characters look like the ones on the Steam storefront? 👀 Many of these characters are legendary and have been in long-running serieseses, imagine booting up the latest Uncharted game only to find that Burt Reynolds is the main character now, all because you turned on DLSS.
Downright disrespectful of the artistic toil that goes into perfecting characters, lighting, colour grading, etc. Like maybe there was an artistic choice made to not have the character look like an AI ad for an AI-generated OnlyFans model?
It used to be a tool to gain FPS while retaining graphical fidelity. Now it's aiming to be a full-on infidel; replacing art to the extent that it's barely recognisable. What will DLSS 8 be like?
‘Artists have the final say’
They already had the final say.
New Mario game looks 🔥🔥fr fr
While everyone is raging on about how a woman looks in video games, I find it even crazier that it changes the fantasy feeling of Oblivion with a hyperrealistic look. Fuck art directions, every game will look the same now with DLSS 5.
the inability to read the room, lol. and I thought I was autistic
Fuck nvidia
It feels like homogenization of art. The filter pays no attention to the original art style and just conforms it to nividias Ai style.
We can't play games unless we play exactly as the developers intended (constant bug fixes, online requirement, anticheat etc. for non competitive games), but Nvidia can ensure games are portrayed differently, entirely shoving a middle finger in artists' faces? And you raise a valid point --
AI is trained on what, exactly, because AI filters always enhance facial definition, add makeup like shadow, blush and lipstick, and full body upscaling adds flat stomachs with abs, perfectly symmetrical D cup breasts, in this case blonde highlights are apparently more appealing than the original.. It all stinks of the generic 2010's Western white heterosexual man's idea of the perfect woman's body. It's like every beauty filter is made by Jack Black's character from Shallow Hal.
It also seems to change the lighting on faces as if they are lit by a studio ring light.
Also in their #1 demo image it changes the girls hair so it’s brown hair bleached leaving a highlight pattern instead of 100% white.
Yeah, there's a bias in these image generator AIs, because model photographs are:
- easy to get in bulk and in relatively high resolution etc.,
- relatively uniform (even lighting, facial symmetry etc.), meaning they reinforce existing patterns in the training data, and
- relatively easy to generate good-looking results with them, for obvious reasons, so more likely to be optimized towards by anyone involved in the training.
NVIDIA is the AI company now. There's no turning back.
Finally, Gamers™ are getting a filter for all the female protagonists that aren't hot enough. Gaming is healing!
Slopvidia
The absolute tone deafness of the AI industry man. Everything about this is terrible. From destroying artistic intent, over making everything look slop to being insanely resource hungry in this economy. There is everything to hate about this. The house of cards can‘t collapse soon enough.
And this is why indie games will probably have a second golden age.
So it uses up way more hardware and power whilst not improving the part of the game were the fun is: gameplay.
What's next NVidia, an AI driver that plays the game for you?!
I get why some people would like this, but Nvidia has lost touch with their gaming market.
AI users are their market now, and this is what they want. This is an advertisment to lazy game publishers/developers like Bethesda.
Yeah no at this point nuke the whole fucking company.
Someone reported your comment.
that's glorification of violence and then there's nuclear options....
You don't have a pocket nuke, do you?
Last I checked I'm not the president lmao
"Make it shitty"
It doesn't look like the same character anymore. The right side just looks like something AI hallucinated that is vaguely similar. If that were a real person you wouldn't be able to recognize them anymore.
Also, what the fuck do you call that? "Standard Human Face" ?
AI has never fixed shit ever. It always just makes shit up. I wouldn't even use it to remove the Starbucks cup from Game of Thrones, cause I don't know what the fuck it will put there instead. Whatever it "thinks" is supposed to be there, I guess.
They’re being very careful about cutting almost immediately after the DLSS is turned on. It all looks like printed media, not actual video. There’s something very off about the way the details just sort of track along the normals of the surface. Maybe something I’d enable as a novelty in a few rare cases, but certainly not something I’d use to replace any existing solutions.
Edit: There may come a moment in time, like it did for me with RTX (specifically Control) where a game finds a way to work it into the very idea of the look of the game and manages to make DLSS 5 look jaw-droppingly gorgeous. But that will be up to artists, not the geeks at NVIDIA trying to shortcut the artistic process. I can see it being a tool for some rare games, and those are the few I’m actually really looking forward to.
You can see more of it in action in the Digital Foundry video.
The stills are dogshit enough but this actually looks so much worse in motion lmao
Yeah, that Resident Evil scene looks like a mock-up. The Oblivion remake and Skyfield look like it's just the lighting and specular.
Don't call it DLSS!
Call it AISSS (AI Slop Supersampling)!
AiSS: Arbitrarily Interfering Super Slop
ASS
Aaaaaaaaasssssss!
What they were thinking?
ASS SLOP?
Maaaan.... I really wish they'd fuck off with this shit, but it's only going to get worse because this is how they're going to push cloud gaming given it takes 2 5090's to achieve this, and dummies will help them because she's hotter now.
A long way down this road there are a bunch of morons voluntarily plugging themselves into the matrix, dick first.
oof the dislike ratio on this video is brutal. 23k likes but 49k dislikes, at least according to PiePipe.
as it should be. DF has instantly lost all the credibility they worked so hard to earn for so many years.
Meanwhile my windows PC restarts if I launch a game with kernel level anti-cheat, after updating my GPU driver a few months ago. Just tried the new driver, instant reboot.
MS and Nvidia are fucking entirely made of vibe-coding morons, and apparently testing is no longer needed.
DLSS 5: For when 85% fake frames and 75% fake pixels just isn't enough!
To be pedantic: multi-framegen can do at most 75% fake frames. You can, however, get more than 99% fake pixels on your real frames with game overrides.
How much does this increase my carbon footprint?
Am I able to compete with the likes of Taylor Swift or Elon Musk if I use this?
What happens when you apply this to a game in a cartoonish style? Is it going to look like those unsettling renders of the Simpsons characters with life like human skin?
It's probably gonna turn all characters into AI Anime Waifus.
Like all generative AI, it probably has 5 modes: deepfakes, uncanny valley, disneyghibli slop, shiny hentai, and schizocore
Give me the schizocore please
Thanks! I hate it!
AI filters on the end result that basically kills the soul of the art direction. I can see a game parody that capitalizes on this.
Nvidia actually used two RTX 5090s for its demos: one plays the game, the other exclusively runs the DLSS 5 technology. The use of two GPUs is required right now as DLSS 5 still has a long way to go in terms of optimisation - both in terms of performance and its VRAM footprint. However, DLSS 5 is designed for use on a single GPU and that's how it will ship later this year.
So much easier when you can just throw 2x the hardware at things. Let's wait and see how it works on a single card.
Also, guessing it will be set up so developers can turn it on and off as needed in their game. If they make it a user-facing global setting, it will quickly turn into a meme generator.
Edit: Option B it is: https://www.gamingonlinux.com/2026/03/nvidia-dlss-5-has-become-the-source-of-many-memes-as-the-backlash-continues/
Leaving aside the question of artistic intent, does this not lead to every player having a different experience? Or maybe even every session? DLSS 5 seems to fill in a looot of details and even changing things. How are they making sure it's the same for every player?
This is not even remotely close to being a released product so far as I can tell. It's running on two RTX 5090s in series.
Came here to post this, I'm glad it's blowing up. What new monstrosities they unleash upon us... I don't know how to even list all the things I have problem with with this.
Oh god no, I hate ai slop
Important details from a post-demo writeup:
During the demo, the DLSS research talked through the level of granularity available. Developers don't just get an on/off switch. They get intensity controls that can be dialed anywhere, not just full strength. They get spatial masking, so they can set the water enhancement to 100%, wood to 30%, characters to 120%, all independently within the same scene. They get color grading controls for blending, contrast, saturation, and gamma. All of this runs through the existing SDK, which means studios already using DLSS and Reflex have a familiar pipeline to work with.
The demo showing the tech running at 100% is not going to look the same as full games built with it over the next year before release.
Another thing to keep in mind is that the only thing it's changing is the lighting effects. The models aren't changing at all (even when this looks hard to believe).
Yes, at full strength the effect at times looks pretty bad (anyone remember when devs could suddenly use bloom effects and entire games looked like Vaseline was smeared across the screen?). But it's not going to be flipped on at 100% across the board for most games.
My guess looking at the demos so far is that a lot of material lighting like stone, metal, etc will have it at higher strengths and characters, particularly faces/skin, will have it considerably lower (the key place where it's especially uncanny valley).
Aside from all the other issues surrounding it, every demo I've seen so far is creepy uncanny valley stuff that I found unsettling and wouldn't want to be playing.
They have completely lost their minds
Leon looks like that Russian guy with the massive chin.
RTX on
Reading through the comments, this is definitely an unpopular opinion, but while I don't love it, it does look more "realistic" to me, even though it definitely needs to look more like the source material.
The original image is bad because the game is rendered at a low resolution and upscaled likely using DLSS 4.5, I think this is intentionally deceptive. We know they have to be using DLSS 4.5 due to path tracing being enabled and running in real time which is extremely GPU intensive even for the 5090. That smoothes away every texture. Nvidia is trying to tie the gaming industry into being "AI default" instead of using normal rendering techniques, which, if done correctly look much better than this.
Yeah, probably the main reason it's getting the little bit of praise that it does is that they're showing it off on games with fairly flat-looking skin shaders. Unfortunately a problem with this sort of thing is that getting that "2023" image is the result of giving a whole team a huge amount of time to model one man's face. If you're Bethesda and you just want to get NPCs into Starfield, it would be a similar amount of work. A bit less, since the first people already gave a talk on it, but still much more work then just getting a diffuse BRDF with some subsurface scattering and calling it good. But you also need a process that can be applied to every single NPC...
And looking at Striking Distance Studios, the company where that 2023 image is from:
In February 2025, it was reported that most of the studio's developers had been laid off.
Yeah, I think it's safe to say that the work those people put in will never be directly reused.
Another reason the DLSS version looks a bit more realistic there is because of the specular highlights on the eyes, for example. They probably aren't reflecting anything real, or else they would be there in the original. But the AI knows that specular highlights add realism and are plausible in this scene, so it puts them there. That's something that an artist could do if given a specific shot and camera angle, but in the general case they can't really do that without causing problems.
There are specular highlights on the eyes in the original, they are just much more subdued.
Oh, yeah, I didn't see those. I think my point still stands though, really those specular highlights shouldn't be that bright, but the AI can figure out that it's plausible for them to be brighter and that it would fit the target style better.
That's a really good point. I'm not even sure what the game is, I had assumed it was something old based on the before image, which would be a sensical use case for this imo.
It's the new Resident Evil 9 (Resident Evil Requiem)
Nah, total recall, the triple boobs scene, to be ignored cus it's cool sci fi
It looks "realistic" to you because your brain has already been fried by slop content. This is a feature for people with zero sense or taste. So I am sure some people will absolutely eat it up.
You make a good point in the second half of your comment, but I'm not sure why you had to call me an idiot first. I don't watch short form content where AI is apparently rampant, and I haven't ever intentionally seeked it out elsewhere. As far as people being fried by AI goes, I've gotta be near the lower end of the spectrum.
My take isn't even gung-ho on the feature, and I only commented on a single screenshot.
Not even true. People would think it looks more realistic 10 years ago too.
Exactly. This is one of the few good uses for AI. It looks less like a beauty filter and more like a realism one. If it can make the rest of the game look as real, I'm all for this. I just hope it doesn't require a GPU upgrade, because I was lucky enough to score a 4090 at launch for MSRP, and I plan on keeping it for at least another 6-8 years.
Given AI's training data biases it's not true reflection of reality but more of streotype projection. I don't think its effect on PoC character, unconventional demihuman types, or any "semi-real" artstyle will accurately enhance artists' original intent.
Brother, get your eyeballs checked. Nothing about that looks photoreal.
When did I say it looked photo real? All I said was that it looks like a realism filter
I'm not sure I'd go as far as saying it's a good use of AI. Also, if you'd read the article (or even just the comments) you'd know that this required 2 5090s. Your 4090 is not going to cut it.
I guess I will just keep playing factorio. That's a game which doesn't involve any automatic generation involving intelligence ... Wait... Fuck!
/s
To be clear I don't think factorio has any Gen Ai. It does have a lot of automation, a lot of generation and requires a lot of intelligence. I thought it was just a clever bit of word play.
FACTORY MUST GROW!!
Yes on factorio, even though I am so bad
If you think you are bad, you are not. There is no wrong way to play and you may want to have different goals. Do you want fastest time? Smallest area? Most powerful against hoards of bugs? Conquer all planets? Science per minute? Self imposed limitations?
All are valid and all could be fun.
The unoptimized demonstration requires two 5090s, yes. Final version might be less demanding. Let me believe.
Brother leave the basement every now and then
Close one, I almost just downloaded a mod instead of burning money.
Wow, all of those I unreal engine "next Gen" Mario and Zelda videos were right.
Digital Foundry somehow still praises it 🤦🏻♀️
slop
it just seems to be fucking the lighting up while also slopifying beauty stereotypes onto women making them look homogenized and not like the artists want (not even just adding detail, it changed hair from blond to bleached brown and changed face bone shape)
who actually even wants this, like, actually wants to use it??
As we already know, this is the south park timeline, so DLSS 6 (or a later one like 7 or 8) will most likely have undressing functionality.
AMD should thank Nvidia for giving them customers (i've never actually used, and never will use, an Nvidia GPU, i have intel integrated gpu + amd gpu rn (but this pc is pretty old), previously i used a Mac. hopefully ill be able to get a new thing with a good AMD GPU)
Grown people expend energy on this?
This looks like legitimate ass. RAM prices are through the roof in part because of this shit - we've regressed heavily.
I don't want to hate on it, but at the same time why is everything AI just hyper-realistic uncanny valley perfectionism BS?
I know people harp on AI for stealing art, but honestly whose art are they stealing? Because its really bad. If we're going to live in a dystopia can they at least steal art of people who are good at it.
"This does not just apply to high end graphics. We're already seeing that a little bit with Starfield, which is a more primitive title" lmaoooo
Bethesda should be offended that their flagship product that runs on their archaic engine was just roasted like this
Todd is too busy swimming in his vault Scrooge McDuck style.
I can see this being jarring as current AI implementation of this kinda thing in videos leads to inconsistent facial features which can have an uncanny valley effect.
techno-fascists are incapable of making decent art, hence why they love this shit. They don't care about Grace as a character or what her design can tell us about her, she needs to be "upgraded" to look like the most generic model ever because they don't care to actually understand anything about the characters or story or fucking anything at all
Thank you for finally providing the context for this image. I've seen it all over the place but I don't game on PC so I wasn't going to take the time to look it up.
Be a lot more popular applied to digital porn
Changed the bone structure of the face and the clavicle.
If claims about video game characters setting standards for beauty, hold any substance, then the same logic should be calling this racist propaganda.
How about Half Life Cinematic mod, but for every game?
These images and press release is not for gamers, but for shareholders to “feel” like their investments are going to pay off.
Omg someone tell that young person her lips are looking thin!
LOL The right image looks so haggard, hints at the potential training on NSFW material those certain features being changed on Grace's model, particularly the emphasis on her lips, hair, and gaunter features. I actually prefer the left, as it seems more in the confines of the game's original portrayal of Grace; the game's background looks better in the left image vs the right image which seems cartoonish in nature. Yikes, NVIDIA is making a big argument for me to fully convert to AMD (which has better open drivers). Realism is fine and all, if that was the original intent of the game's designers.
Looking at the first comparison I thought it was pretty good, but weird that they changed her hair though.
Then I got to all the comparison shots in the NVIDIA blog post and noticed it basically just brightens dark scenes, relights characters and eats background detail, this looks very half baked to me.
Lol it gave her low income girl eyes and dyed roots
It gave her Mar-a-Lago face
The starfield demo looks better with dlss5 on, but that says more about starfield than it doesn about dlss5.
This was definitely something I predicted like a year or so ago. Richard Burns Rally will continue to be the default rally sim, but they will never bother to update the 2004 game engine. Rather at some point, AI will literally start MITM capturing all the 2004 graphical output and converting it to photorealistic modern graphics that are output to your screen instead.
EDIT: But yeah, this is such a "fuck your game, fuck your artistic intention, eat our AI slop and like it!"
Ew. It looks like ai porn. I feel so bad for the teenage boys who are jerking off to that shit not realizing they're going to royally screw up their attraction to real women.
Peak grift. Hopefully it's all downhill from here.
Instagram filters for everyone!
Can I have this for teams? I work with so many ugly people that I need to paint them over in real time 😁
I don't care if 2 50x somethings need to run on 100 percent half the day.
Can you turn it off?
Holy shit that looks bad. I was about to say come on guys its not that bad and then I watched the video. There must be no gamers at nvidia because I cant imagine anyone wanting to add 70ms of latency to make their game look like this.
I've seen prerendered indie games do this deliberately. I do NOT understand it, at all.
Its one of those things that I really hope they make optional, a fun thing to turn on and play with but not always over every game ugh.
I honestly want to See ts applied on terraria
It’s bad now, it’s gonna get worse soon… I do feel that modern video games have plateaued in recent years, with the addition of this I do believe that we are staring down the barrel of a Grand Canyon sized crash.
I would looooove to play some AI-transformed Balenciaga games lol
Of course the whales are going to buy those cards.
@popcar2
It's so bad. Artstyle ruined AND is running off of a SECOND GPU. How did Nvidia okay this...?
Also sad to see Digital Foundry approve this. I thought they were better than that.
Just to be clear, this is what you have a problem with: https://streamable.com/j0ryqe
They should really check out that one app that changes whole faces. It's pretty amazing. /s
They no longer care about making the best product to get the consumer to buy because they know if they make hot girls you can control they’ll sell millions
Why make art when live like king from sex machine
I don't know the games or the characters but I think they both look as good.
It makes them look like two different people, but they both still look good.
the memes are strong with DLSS-5. 'for the low low price of..' $1000.00
This is how they made the Demon's Souls remake
damn look at all the free marketing nvidia is getting with the memes. Idk when people will learn that if you aren't into something it's best to ignore it and not generate buzz about it.
The only good application of DLSS5 I could've think of is Euro Truck Simulator 2, hyperrealism would pair well with the game.
What did you think dlss was before..?
Advantage: Better graphics
Disadvantage: Needs two 5090 graphics cards.
Apart from that, I'd like to see a Minecraft version of that 😃 just for fun.
Hey at least it still runs on your machine.
DLSS 6 will run exclusively on the cloud because RTX 6 series won't ship with any RAM or compute nodes since they all went into enterprise cards lol
It does not generate a new image, thus generates no AI slop in my book. It is literally an image filter with AI, which is what DLSS already was. This circlejerk slop has torpedoed any discussion about the actual flaws - the uncanny valley effect during animation, the ridiculous amount of GPU power this technology seems to need, how absurd it is in the actual economy - and has substituted it for its circlejerk hallucination that this is just them sneaking image generation into DLSS.
-
It takes color and motion vectors and passes it as input through AI - same as previous versions of AI.
-
It generates no new shapes or geometry.
-
It just applies changes to lighting and materials, basically creating a mask over the original.
If you call this "image generation", you might as well call any filter that. This is not the generative AI the circlejerk slop is hallucinating. It's an evolution of what DLSS was able to achieve before.
I've actually just been corrected, this was referred to employ some form of generative AI by Jensen. It's also significantly different enough to what I generally thought of as AI slop and my issues with it that it could also be said that I am a supporter of generative AI now.
The second one didn't look too different from the original, but the first and third look so far from the source material, it's jarring.
Looks fine to me
actually helps out in the sports games. ea has the same lightning like they had in 2006, so this takes it up to speed to 2020. however real games gonna look arse. alao 2 x 5090 cards required, donno who the fuck is the demographic. also thanks for killing dual gpu support so you could bring a shit one back.
Insanely impressive that it’s able to do that in real time.
My phone can apply facial filters for years in realtime.
Your camera also runs at 4k 60 fps without a GPU. You're already photorealistic. Are you really stupid or just pretending?
Lighting in video games encompasses a lot of things these days. I'm not a graphic designer but, to my untrained eye, it looks like this tech is pushing hard on things like subsurface scattering to make materials appear more realistic. It doesn't appear to be changing the actual polygons. This picture further down seems like a more realistic example of what it would end up looking like when the technology is complete. I think they picked that particular first image as an example BECAUSE it's such a stark change.
Ok but the included picture added eyeliner and a red lip stain.
It Hollywood glossified it.
Zoom in on the before; the textures are fairly low res so it's hard to see at a distance but she clearly has eyeliner along her top eyelid. I can't really tell if the intent was for the dark undereyes to be eyeshadow or bags from being exhausted. But you can tell the AI is pulling from source coloration, not just adding things willy nilly like a beauty filter. The lip coloration, idk, that could be an artifact of the early technology or a hallucination. But to me it appears that she had some sort of lipstick on because it doesnt extend all the way to the corners of her mouth. I don't think contouring is the right word but maybe someone here who wears makeup could add some detail?
edit actually looking even more closely, i can't tell if that's eyeliner or just her eyelashes being dense. Either way you can see what the model is trying to replicate. I don't think it looks good but it isnt doing a "beauty filter". If the model is adding eyeliner, it appears to be confused based on the thick black line in the source image. Again, these are really low textures compared to what it's trying to output so hopefully this improves over time like every other DLSS tech has.
It's not contouring, it's adding muscle lines, that I interpret as pure scaling up. But yeah as someone who wears makeup, I stand by my position especially on the lip color. It just reminds me far too much of the incel redesign of Aloy from Horizon Zero Dawn. It also plumps the lips and gives them more variance from center to edge.
And I want to be clear here, I don't think they did this on purpose. I think that the training data is more likely to use more conventionally attractive women's faces including ones with makeup and so it accidentally does some light beautification while in there.
Yeah, I dont know what the word is. My point is mostly that you can clearly see what it's trying to recreate from a low-quality source. It seems to be keeping a very close adherance to the original looks? If you lay the images on top of eachother you can see the same creases around her eyes, the same nose shape, the same jawline, the same skin-colored makeup at the corner of her mouth to make her lips appear more defined. I think you're spot on about the "training on conventionally attractive models" which is the part that I'm hoping improves with time. I don't think applying this to existing works is appropriate at all but apparently it will use nvidia streamline so the mods will come whether we want them to or not sadly.
I am so not sure about your points, if they are all the same, and AI just tries to add more details, then should AI disregard the source material and make a more plain interpretation to appease certain viewpoints?
DLSS of course does use machine learning (transformers or CNNs) which you can call AI if you want, but I think there may be a bit of a misunderstanding about what this new version is actually doing. If you watch the video or read the breakdown they note that DLSS 5 is NOT replacing or altering the original graphical assets in the game with generated ones. There’s no new textures or 3d models being changed here. It only alters the lighting of the game to add more realistic shadows and reflections etc. It’s just that it does that so extensively that it gives the impression of being an almost entirely different image. Kinda nuts that lighting can do that but that is apparently what is happening here.
That said, arguments about artistic intent are still very relevant here since it’s a very different look to how the original devs may have intended.
So, two things,
which you can call AI if you want
This is from their official blog post, emphasis mine:
DLSS 5 is the GPT moment for graphics — blending hand-crafted rendering with generative AI to deliver a dramatic leap in visual realism while preserving the control artists need for creative expression
So, Nvidia themselves are the ones calling it AI
Second thing,
DLSS 5 is NOT replacing or altering the original graphical assets in the game with generated ones
I think they low key shot themselves in the foot with this particular comparison. The AI screenshot is actually from a frame (or however many frames) later in the scene. If you look closely you can see her mouth has actually opened slightly, and she appears to be in the process of rotating her head.
The end result is her lips are unquestionably larger in the AI image. Like if you bring the comparison slider directly to the middle, the bottom of her lip is lower while the top is higher. It has a pretty profound effect on the overall image.
Combine that with the hair highlights and makeup being so much more pronounced and the whole comparison is waaaay too similar to the “anti-feminism” filter thing that went viral in recent memory
I bet you're all being used to manipulate stock prices.
"AI Slop"
I'm curious. How do people think DLSS actually works? Do y'all think the Nvidia fairy comes and sprinkles magic dust on your GPU?
I think the one on the right looks better.
Please don't burn me as a heretic.
Looks cool to me, photorealism is coming. Not really a gamer, btw
Yea, though I walk through the valley of the shadow of death, I will fear no evil…I think the right picture looks better.
That's the valley you walk trough blind.
I’ve since seen videos and a screen shot of Arthur Morgan. This is terrible.
I’m an atheist now.
All the rage focuses on faces, and yes, some look rough. But nobody is talking about the environments and textures. I think they look incredible. Old games may have been a bad example for them to give. Games built with this in mind from the start? I'll hold judgement until I see a game built for this from scratch, not a demo of a 15-year-old one.
I dislike AI just as much as the next person, but I don't get all the hate for this image. It looks fine.
Do you guys turn your graphics settings to lowest to make sure not to interfere with "artistic vision"? Or do you just not understand the technology at all?
The dev can literally control the vectors for DLSS 5 and test it for themselves until it looks how they like. If you don't like modifying frames, you should disable all post-processing shaders and most modern anti-aliasing too.
I'll likely never own a 5090 at all. Definitely not two of them.
But while a lot of you are hating how that before and after looks, the after image if just straight up nicer looking and more realistic.
It shouldn't look so realistic? Man; I'm in my 40's. Game devs have been trying to make people in games more realistic looking my entire life. Her before image wasn't less realistic due to artistic choice. It was because it couldn't be done easily or run on a system.
Lara Crofts tits didn't start as triangles for artistic reasons.
An ocean of hate in this thread... I'm excited for this tech, and hope there's a way to inject it into classics that haven't aged so well because they aimed for realism at a time when it was unachievable. Now we can actually see how those games might have looked had they been fully successful.
Off the top of my head, I think this would be great in most Elder Scrolls games, most notably Daggerfall. Almost anything released for N64, with it's extremely flat textures, or PSX with it's heavily pixelated textures. Maybe GoldenEye you'll actually be able to make out other characters from further than 20' or Final Fantasy IX won't look like 3D blobs hovering over static, pre-rendered scenes.
I'm excited to replay so many games without feeling let down by my nostalgia filter when actually seeing them again!
Now let the waves of down-votes roll over me like I'm at the beach, for having the audacity of disagreeing with the irrational and incensed mob.
If this tech appeared four or five years ago, it would be all praise around here...
AI hate is becoming "hate for the hate" just because it's the dominant opinion.
"Artistic intent..." When did you guys care about artistic intent before? Did you ever use a mod to change visual aspects of a game?
The textures are flat on purpose so the AI filter looks "more detailed" in comparison.
They don't have to be like this, see this Callisto Protocol 2022 screenshot no weird AI filter needed.
The textures are flat on purpose so the AI filter looks “more detailed” in comparison.
Definitely. Boot up the game and look at the actual graphics. Their "OFF" pictures look terrible.
It looks like it applied some make-up and changed the hair colour.
bruh it looks like shit lmao
I think a bigger issue is the fact that they override the original artists' intentions and expressions. It's like if you painted a painting and AI is like, hey let me just touch up everything for you that I think you fucked up.
It's not the quality that is the point here, as much as that sucks too.
it doesn't look "more realistic", it completely changes the character's appearance: hair that looks dyed, redder & fuller lips, more prominent cheekbones, a more defined chin, and eyes that instead of kinda looking away avoidedly, now stare at your soul through the screen (although that one might be just the lighting, idk)
basically this except more subtle, but the idea is the same, make her look more attractive under the guise of realism. it's gross on its own, but the implications are even worse imo, just giving gooner chuds the tools to turn any game into the porn of their dreams, fuck art amirite? it's literally turning everything you see on the computer into a unique serving of slop, every copy of mario 64 is personalized except instead of interesting analog horror each player gets to discuss how big peach's tits were on their version
If I was an artist that worked on the game and had this shit forced on my product it would make me fucking furious!
How about the part where they had to use 2 RTX 5090's to do this? One for the game and one for DLSS 5?
Please try and justify that.
Easy, if it needs that much GPU and even one GPU is hard to find, then it could make someone rent their games from Nvidia to use their GPUs...
Oh. Did you mean justify for the user? Lol, no, only the interests of nvidia matter.
Supposedly this trash will run on a single GPU at launch, but this still feels like a subtle push toward shit like geforce and taking ownership away from the consumer
It honestly just looks like color, contrast, and sharpening changes... Basically what could be done with a reshader app.
I hate ai too but this seems like it's just blatant anti-ai outrage. In the end this works against the honest anti-ai movement.
Really just seems like "live" ai powered reshading...
I still don't understand all the extreme anti-ai cult in these types of contexts. Oh well.
Whatever ai did, the right image looks better than the left.
I honestly think the anti ai thing has a lot of crowd mentality aspects to it. There's reasons to dislike some of what big ai business is doing, more reasons to push for regulation. But this specific example? I got nothin...
People have repeatedly in this thread talked about how it also added details that were not hinted in the original, and in part it looks like adding makeup, which could totally undermine a character or setting if they are unlikely to care about our have time for makeup.
Characters that have barely survived in the wilderness for weeks somehow wearing lipstick and eyeliner and eyeshadow... That's the sort of thing that can happen with this approach.
If only Leonardo de Vinci had a camera, can you imagine how much better the Mona Lisa would have been? /s
Yea it looks way better on the right. lol. People just wanna be mad. Just assume they're bots and ignore the chirping
I too have a complete and total disregard for artstyles and developer-intended aesthetics
It's like when people look at a film deliberately animated with a particular framerate, see the version where someone had AI interpolate the frames to make it like 120 FPS, then say it looks objectively better than the original, because the only metric they can value a piece of media by at that point just seems to be "I need more frames"
Like In Into the Spider-Verse, when Miles is animated at 12 FPS, and the (more experienced) Peter Parker is animated at 24 FPS, but after Miles improves and gets a better hold of his skills, he gets animated at 24 FPS, too. The lower framerate subconsciously makes us interpret his movements as more choppy, inexperienced, and imprecise, on top of his existing animated movements, to even better sell the plot point of his inexperience.
Meanwhile, many people's TVs have motion smoothing, which entirely destroys this effect and makes the film fundamentally less communicative as a result, even if on the surface people just say "It's smoother so it's better."
Lmao. I definitely said that.
I too imagine people saying things I disagree with so I can stay upset and disagreeable
I mean that's exactly what you're saying if you genuinely thing that yassified filtered dogshit with atrocious lighting actually somehow looks better and retains the artists' intent
You don't think Grace would have dolled herself up to go investigate the burned out hotel where her mother was murdered as soon as she finished having a panic attack?
I don't recall even mentioning the "artist's intent" ya weirdo. lol
Such an odd way of communicating and interpreting others.
So why don't you explain what you meant because it sure sounds to the rest of us that you're basically "yah brah totally hotter this way YOLO hang loose wooo!" and completely missing the point on how destructive this is to the creative process and the art that goes into making a game.
This is Nvidia doing this, right? I don't see why they wouldn't try and improve the graphics in games when using their graphics cards. DLSS has been criticized as it's continued to improve. If people use it it's because they prefer the graphics with the minimal impact on FPS.
Again, I don't think I've commented on Ithe impact on artists. Since you insist, I'll comment on it. Y'all raise interesting points. I'm not sure what's in the best interest for super stressed, underpaid, and underappreciated game developers. If this leads to them selling more games because they can run on cheaper hardware then maybe it's great. If it completely distorts some face they worked really hard on then maybe that sucks.
That said, I'm just not gonna pretend the one on the right doesn't look more realistic and I wanted to disrupt the echoes
And as everyone knows, you MUST enjoy something EXACTLY as the creator intended. That's why when you go to restaurants you're not allowed to ask for substitutions. Your preferences don't matter in the slightest, because you risk offending the chef.
I'm going to assume they're confused about which is left and which is right.