By Evan Zimmer
Last month, Nvidia revealed its latest version of DLSS, which has drawn negative criticism from gamers and developers alike. The criticism stems from the tool appearing to alter the art direction and style of the games Nvidia used to show off the tech, including Resident Evil Requiem and Starfield.
Since the DLSS 5 announcement, Nvidia CEO Jensen Huang has responded to public outcry a handful of times, saying gamers are entirely wrong about their stance on the tool. In his most recent comments, Huang, unsurprisingly, once again defended DLSS 5, saying it’s just another way for developers to improve visual fidelity. He took it a step further by saying he views all AI-generated content as “beautiful.”
However, developers like Quin Henshaw, a Unity development engine expert and Unity Instructor at Vancouver Film School, view the latest iteration of DLSS as a step in the wrong direction. From his standpoint, DLSS 5 will only have a negative impact on game development and art direction as a whole.
What is DLSS?
Nvidia’s deep learning super sampling (DLSS) is an AI-powered graphics technology that is designed to both increase visual fidelity and improve performance. It uses AI to render games at a lower resolution and then upscale the image to make it look better. It’s similar to Sony’s PSSR, which also leverages AI to go pixel-by-pixel to upscale resolution.
DLSS uses a type of AI called a deep neural network (hence the “deep learning” part of its name) to enhance images. DLSS was trained on supercomputers to recognize high-quality images and learn how to reconstruct what lower-quality images were missing, therefore upscaling the image.
It is not something developers use to build games, but it’s something that runs with the game in order to improve its image and performance. Developers are able to tune it to various degrees, but DLSS has to be implemented and enabled by developers in order to work.
DLSS improves performance by reducing the amount of work a computer’s GPU has to do. If a game is rendered at 4k, that means the GPU is processing millions of pixels. With DLSS, the game could be rendered at 1440p or even 1080p, and then employ AI to upscale it to 4k.
The prior versions of DLSS haven’t been problematic. In fact, image upscaling is becoming increasingly common, and is generally viewed by developers as a positive technological advancement in game fidelity. “All of that upscaling tech is pretty slick, and as much as [NVIDIA is] leading the forefront with it, pretty much every software developer has their own version of that now,” Henshaw said.
AMD, for example, offers FidelityFX Super Resolution (FSR) as an upscaling tool. However, it doesn’t require dedicated AI hardware like Tensor Cores, which is what DLSS needs to run.
“I've been pretty on board with a lot of the tech that Nvidia has created in the last few years,” Henshaw said. He's used prior versions of DLSS to create tech demos in the past. Generative AI was first introduced into DLSS in 2022 with DLSS 3. It marks the point where Nvidia went from AI-assisted rendering to AI-generated frames.
DLSS 3 used generative AI for frame generation. The GPU creates a frame to go between frames rendered by the game’s engine to improve image quality and performance. DLSS 3.5 took it a step further by allowing lighting generation via Ray Reconstruction. DLSS 4 improved on the tech with refinements to frame generation and greater integration with Ray Tracing technology.
However, DLSS 5 does more than simply upscale an image via generative AI to make a game run smoother or look better. It’s starting to more aggressively influence how the final image appears, potentially changing aspects of how a scene looks. That’s what developers and gamers are taking issue with.
“It's not something that I could see any serious development studio, either at the indie level or the triple A level, really get into, unless there's a big push from publishers to cut down on costs,” Henshaw said.
In a video by YouTuber Daniel Owens, Nvidia's Jacob Freeman provided a few more details about how DLSS 5 works. It isn’t really changing anything on the game engine level, but it is essentially taking a 2D image and running it through its generative AI technology to tweak the image. Artists can control things like color gradients and filter intensity, but they don't seem to have control over the end output. Freeman also stated that this is an early preview of the technology. DLSS 5 is expected to release sometime this Fall, so it stands to reason that the full extent of developer controls will be made clear around that time.
Read the full article on GameRant
This article originally appeared on GameRant and is republished here with permission.
