With DLSS 5, NVIDIA presents a new generation of its AI-based graphics technology. Instead of just improving performance, the technology will actively contribute to image quality in future. A neural rendering model generates lighting, materials and details in real time and adapts them to the game world. The aim: photorealistic graphics, which until now have been more familiar from Hollywood productions. For game developers and gamers, this could mean a decisive step towards cinematic real-time graphics.
- AI-based neural rendering creates more realistic lighting and materials
- Real-time processing up to 4K resolution for smooth gameplay
- Analysis of scene elements such as skin, fabrics or hair for more precise rendering
- Support from major publishers and integration into numerous upcoming games
NVIDIA DLSS 5: AI changes the way game images are created
For years, NVIDIA has been developing technologies to bridge the gap between real-time graphics in games and elaborately rendered movie sequences. While visual effects in Hollywood often take minutes or even hours per frame, games usually only have around 16 milliseconds per frame. Pure computing power is therefore not enough to generate photorealistic images in real time.
This is where DLSS 5 comes in. The new version combines classic rendering techniques with generative AI. Instead of having to fully calculate every pixel, a neural model analyzes the scene and adds lighting, material properties and visual details. This is based on color and motion data from the game engine, among other things.
The AI has been trained to recognize complex scene elements and display them realistically. These include skin with subtle light scattering, fabrics with fine glossy reflections and hair with precise light-material interactions. At the same time, the result remains deterministic – a decisive factor for games in which image content must be exactly reproducible.
An important difference to generative video AI: DLSS 5 does not generate its results offline, but directly during gameplay. This allows scenes to be dynamically adjusted without losing control over style or art direction. Developers are also given tools to control intensity, color corrections or masking in a targeted manner.
Support for large studios and first games with DLSS 5
According to NVIDIA, the technology will be supported by numerous major studios. These include Bethesda, CAPCOM, Ubisoft, Tencent and Warner Bros. Games. The first announced titles with DLSS 5 support already show the direction in which graphics quality could develop.
Games that are set to benefit from the new technology in the future include
- Assassin’s Creed Shadows
- Starfield
- Hogwarts Legacy
- NARAKA: BLADEPOINT
- Phantom Blade Zero
Of particular interest is the ability to depict complex lighting situations more realistically. Scenes with backlighting, diffuse daylight or dynamic shadows could look much more natural than before.
Integration remains comparatively straightforward for developers. DLSS 5 continues to use the familiar Streamline framework, which is already used for previous DLSS versions and NVIDIA Reflex. This allows studios to integrate the technology into existing projects without having to rebuild their entire rendering pipeline.
You are currently viewing a placeholder content from YouTube. To access the actual content, click the button below. Please note that doing so will share data with third-party providers.
DLSS 5 also marks a strategic evolution of the DLSS series. While previous versions were primarily used to increase performance through upscaling or frame generation, the focus is now more on image quality.
Conclusion
DLSS 5 could represent an important step towards photorealistic real-time graphics. The combination of classic rendering and generative AI creates a new approach that should improve both performance and visual quality. Supported by major game developers and several announced titles, the technology is expected to spread quickly. According to NVIDIA, the launch of DLSS 5 is planned for the fall, probably initially on current GeForce RTX graphics cards.
