Previously, Jensen Huang zealously protected the technology from critics, trying to convince them they were wrong.
Blogger Daniel Owen was able to ask Nvidia evangelist Jacob Freeman a few questions about DLSS 5 and find out how the technology really works.
Although Jensen Huang previously stated that DLSS 5 does not perform post-processing of frames, but analyzes and "understands" the composition of the scene, after which it generates at the geometry level, the words of the head of Nvidia turned out to be far from the truth.
As it turned out, DLSS 5 works exclusively with the finished 2D frame and motion vectors for subsequent generation and stability. The technology does not have access to 3D geometry or scene depth and receives visual data only from the final image that the game engine outputs.
