In response to a question from tech YouTuber Daniel Owen, Nvidia’s “Head of GeForce Ecosystem” Jacob Freeman revealed some concerning details about the DLSS technology. Freeman explained that DLSS 5 does not understand the nature of models, textures, or light sources in games. Instead, it only works with a 2D image sent to the player, making alterations without context.
For example, in Resident Evil Requiem, if Grace Ashcroft’s face appears to change, it’s because DLSS 5 assumes that’s how it should look, without reference to the 3D model. This raises questions about control over DLSS 5’s output.
Freeman indicated that developers can adjust the intensity and color tones of DLSS 5 but provided little information beyond that. This lack of clarity about how DLSS 5 functions and interprets 2D images poses concerns about maintaining artistic direction in games.
Freeman highlighted the challenge of ensuring DLSS 5 accurately interprets materials like metals or woods based solely on 2D visuals, without additional context from developers. The potential misinterpretation of textures could lead to visual discrepancies in games.
Despite claims from Nvidia CEO Jensen Huang that DLSS 5 can control geometry, textures, and other elements using AI, Freeman’s responses suggest otherwise. The debate over DLSS 5’s role as a filter or an advanced tool remains unresolved, raising doubts about Nvidia’s communication about the technology.
Whether DLSS 5 is deemed a filter or a more complex tool, the current rendering results are underwhelming. It remains to be seen if Nvidia will provide clearer explanations and if developers will receive adequate tools to manage DLSS 5’s impact on their games.
