I really thought Nvidia's DLSS 5 was going to be smarter than this
Nvidia's DLSS 5 technology appears to rely on a static 2D image and motion vectors, rather than a deeper understanding of 3D geometry or lighting. While impressive in its real-time application and consistency, its enhancements are based on inference from a single frame, leading to concerns about artistic control and potential homogenization of character appearances. Developers have limited options to tune the effect, primarily through sliders and masking, raising questions about the fusion of structured game data and generative AI.
- Entities
- Person: Jacob Freeman
- Game: Assassin's Creed Shadows
- Topic: Graphics Technology
- Person: Jen-Hsun Huang
- Topic: AI Upscaling
- Game: Starfield
- Company: NVIDIA
This article was reported and published by pcgamer.com. feed.gg links to it as part of a story cluster — full text, images and rights remain with the publisher.