• Hey Guest. Check out your NeoGAF Wrapped 2025 results here!

Nvidia at Live GTC : DLSS 5

This is the kind of stupid shit people dumb enough will actually believe is true tho

They think oh it will make women look better and then suddenly champion DLSS 5.0

Unironically people are so dumb
Well, that's one facet of this drama.

Its not about making women look better, but make them not look ugly.

In other words: Negating the woke effect
 
HDj24LrXEAAQcRP


ITS THE SAME FACE GUYS HURRDURR
DLSS5 adds lip fillers

Mean Girls Halloween GIF
 
No wonder tons of wokesters are afraid, they should be :D

workaholics-i-like-that.gif
Glow up gen 😁 leftist and purple group will hate this.

If you really think the tech works like this u are retarded m8

This isn't made to 'beautify' female models you absolute goons

Better question is why are you? You have been on the attack all thread to multiple individuals. Its clearly the same model. If you aren't able to spot the differences and why they are there then I don't know what to tell you. The model is unaltered. As I posted to you from before this is directly from NVIDIA.

DLSS 5 takes a game's color and motion vectors for each frame as input, and uses an AI model to infuse the scene with photoreal lighting and materials that are anchored to source 3D content and consistent from frame to frame. DLSS 5 runs in real time at up to 4K resolution for smooth, interactive gameplay

"DLSS 5 is the GPT moment for graphics — blending hand-crafted rendering with generative AI"

From their own blog

You understand what generative AI is right? It's making shit up.
 
Last edited:
HDj6TzZaoAADlzr


Holy FUCKKKKK

Yeah, keep crying about Ai Slop.

Ai can give me that AI SLOPPY TOPPY

we have achieved fucking concept art quality visiuals in real time.

fuck you

the art looks the exact same, just fucking better. fuck off.
 
Last edited:
Is this going to be able to "remember" the last AI reconstruction? Or are characters going to look different every time I turn the camera towards them?
 
0BRu5pSjMHkTY1br.jpg


Doubt anyone will read that, but looks like Nvidia are just showing you how things could look. Devs have complete control.

That isn't "complete control". Not even remotely. The artist needs to define what they want visually. With actual concept art, raw game assets and their version of lights, cameras (like what lens to use, what type of camera, exposure etc) and post processing. Not by messing with masks and intensity sliders. You will never get tailor made/signature looks with that. You think any of the top studios would touch this with a ten foot pole and build their game from the ground up relying on it whenever the hardware capacity is there? Even indie devs that have a real art department would throw it out if they are presented with sliders and dropdowns to "control". They'll let Nvidia make it available as a feature if there is some partnership with them, but not get fundamental adoption like other aspects of DLSS does.

It's still a good start to start exploring this approach though. And the fact that they are able to pull this off on existing hardware is promising. But this isn't the game changer people are hailing it to be. It's just the start of another experimental journey in rendering. It will eventually become the norm once there is actual complete control and lighting consistency that is locked in throughout the entire game. But to do that, every dev needs direct access to massive amounts of compute, so that they can cook their own model using Nvidia's underlying tech, just like how LLMs are being used and built today in the private sector, outside of public facing chatbots. Doesn't seem like we are there yet.
 
Last edited:
HDj24LrXEAAQcRP


ITS THE SAME FACE GUYS HURRDURR
Its ok if it isnt, dlss5 one looks like real life 9/10 hot chick(spoilers, they use make up, and quite a lot) while left one looks like gamey character/doll.
Ask 10k men, almost every1(95%+) will easily will chose dlss5 model/face, hell i think even total chads like young brad pit/leo or middleaged clooney wouldnt say no to dlss5 grace, thats how crazy goodlooking she is in that picture, assuming only jesus is perfect 10/10 she is easly 9/10 here :)
 
Yeah that's about the use i expected for AI in terms of image quality. Don't have anything against it but not sure if i'd use it, i hope the option for the "classic" DLSS to remain in future games.
 
This is absolutely amazing technology. People who think this clashes with the artistic intent: they're not changing any assets here you know! It LOOKS like that because the lighting improvement is SO dramatic!

It's basically how those games would look if you applied CG level path tracing to them. You're talking a couple of generations of RT improvement in a single step. It's a gigantic step forward for rendering.
True, it just looks like A.I. deep fake slop in the demos shown, except in Assassins Creed screenshot.
This look so bad to me, "DLSS on" picture looks like someone put fake face on top of the screenshot in photoshop.
Exactly.
 
It's very impressive and something that will be bananas crazy if they get it to work flawlessly in realtime. But I fear that Nintendo games will be cursed if they ever use it. Especially Mario games. Hopefully there'll be a slider so that you can choose how much you want it to change the game.
 
Before 2019: Ok team, we have to optimize the game or it will run like shit!
Between 2019 and 2025: Ok team, AI will make the game runs good, I don't care about performance, but we still need to make good textures, mesh and 3D models.
Starting from 2026: Ok team, fuck optimization, fuck textures, fuck 3D model, just develop a cheap basic character model, something like a mannequin, AI will make it run and looks good.


Gaming is doomed
 
Last edited:
So many games have interesting visuals and style because they were created with the limitations of the system in mind.

This is just a sad end to video games for me. Pure generic slop.
This, but also even without talking about tech limitations, I just really prefer looking at things created by someone, with an art direction and intent in mind, not something generated artificially....
That's also why, without talking about AI like this, I also never enjoyed procedural content, randomly generated maps etc etc...

I just want hand crafted things, with a real intent and artistic thought behind it. It seems like most people don't care about this though sadly.
 
Last edited:
This is like the Marathon players coming into the CCU thread saying they are really enjoying themselves and we should shut up about the CCU
  • Some are being extreme contrarians
  • Some are actually liking what they see here (your marathon analogy)
  • Some are only basing their opinions on whether or not another group likes it or hates it (the zero-stancers)
 
Is this going to be able to "remember" the last AI reconstruction? Or are characters going to look different every time I turn the camera towards them?
Yes. The AI reconstruction uses the underlying information to inform its interpretation. As long as the underlying information for a given object doesn't change, the reconstruction will be identical every time.
 
Is this going to be able to "remember" the last AI reconstruction? Or are characters going to look different every time I turn the camera towards them?

nope it won't.

it already uses a shit load of memory, if you wanted to make it consistent you'd literally have to keep previous generation in memory and even save it into storage.

this will literally look different every time you start the game, or just open the menu and close it again.
 
Capcom's subpar early PS4 era character modeling is in dire need of an upgrade anyways. They can feed in Grace's real life model as better reference next time.
So be it, but that still doesn't mean I want Nvidia robots doing it. Are people really going to run this filter for their first Elden Ring 2 playthrough? What would it do to the very intentional faces of a Kojima game? I think it's cool as a novelty. I'd like it for replays of old games, but it's not a GPU seller for me. I just want what the developer modeled. But I downgraded my monitor resolution to avoid using current slop scalers. Seems like this stuff is still a ways out before it reaches a mid-range secondhand card that I can buy in a Del Taco parking lot, but I see more value in it than I do the currents slop scalers. Just for fun replays of old stuff. I'd never play a new release for the first time with color grading and facial features that the artists didn't intend to be there.

To me it's a cool novelty like when you unlock filters after beating a game. I'm not anything close to an AI hater. I want AI used to ease the busy work of game creation. I want indie devs to have their manpower multiplied by it. I've said a bunch of times the only next gen from here is photorealistic AI games, and I stand by this. This is just a weird middle ground.
 
Last edited:
Top Bottom