There has been some controversy surrounding NVIDIA’s DLSS 5 showcase, with a number of gamers expressing their disapproval of generative AI. However, Jensen disagrees with these critics and asserts that they are ‘completely mistaken’.
NVIDIA’s CEO Asserts DLSS 5 Empowers Developers to Achieve Realistic Results Based on Input Data
The unexpected unveiling of DLSS 5 at this year’s GTC event caught many by surprise, given its focus on enterprise. By leveraging neural rendering, DLSS 5 enhances visual quality by processing it through an ‘AI layer,’ aiming to create ultra-realistic environments. While some gamers were impressed by the visual upgrades brought by DLSS 5, a significant portion of the gaming community opposed NVIDIA’s efforts, labeling the new upscaling technology as ‘AI slop’. Jensen, however, has a response to these critics.
First and foremost, they are completely mistaken. This is because, as I have clearly explained, DLSS 5 combines the controllability of geometry, textures, and all aspects of the game with generative AI. It is not merely post-processing or frame-level post-processing; it is generative control at the geometry level.
– CEO Jensen Huang of NVIDIA
Jensen’s response directly addresses the perception that DLSS 5 is just a visual quality filter, arguing instead that the technology is seamlessly integrated into the 3D engine’s pipeline, enabling the creation of what he terms “generative control” data. Furthermore, DLSS 5 ensures that the AI layer involved does not engage in ‘guesswork’ when enhancing visuals; it processes structured data such as the 3D skeleton (geometry), character movement (motion vectors), and scene depth to generate textures that depict realistic scenarios, rather than the ‘AI slop’ narrative circulating.
Another key point emphasized by Jensen in responding to criticism of DLSS 5 is that developers have the freedom to decide how willing they are to incorporate the upscaling technology into their games, hence why he refers to it as “content-control generative AI”. Jensen has previously likened DLSS 5 to the ‘ChatGPT moment’ of upscaling, where developers provide structured data to generate a realistic output.
DLSS 5 represents a significant advancement in rendering, although the primary source of criticism lies in the noticeable changes to textures once AI is introduced. It is incumbent upon developers to ensure that the original artistic vision remains intact with DLSS 5, achieved by feeding in structured content that aligns with this vision.
News Source: Tom’s Hardware
Follow Wccftech on Google to get more of our news coverage in your feeds.