Overview
- GeForce evangelist Jacob Freeman told PC gaming YouTuber Daniel Owens that DLSS 5 takes a 2D frame plus motion vectors as input and is trained end to end to interpret scene semantics from a single frame.
- The remark conflicts with Jensen Huang’s assertion that the tech is not post-processing at the frame level and instead provides generative control at the geometry level under developer direction.
- Nvidia’s description says the feature ingests each frame’s color and motion vectors to add photoreal lighting and materials anchored to source 3D content.
- Scrutiny grew after an official screenshot of a Starfield character exhibited a grotesque artifact dubbed a “giga-nostril,” with observers also noting apparent eye-color inconsistency.
- Eternal Darkness creator Denis Dyack called the reveal a mistake, warning the system could override a game’s art direction and weaken incentives for big-budget production unless Nvidia rethinks its approach.