Nvidia leans on DLSS improvements to make up for a lack of GPUs at CES

In a surprising turn of events, Nvidia opted not to unveil new GeForce graphics card models during this year’s CES, a first in several years. Instead, CEO Jensen Huang delivered a lengthy keynote that predominantly spotlighted the company’s robust AI operations, with gaming announcements relegated to a separate video released later in the evening. The highlight of Nvidia's presentation was the introduction of DLSS 4.5, which enhances the existing upscaling and frame generation capabilities of its hardware. This update features a new second-generation transformer model, designed to improve pixel generation predictions by training on an expanded dataset. Bryan Catanzaro from Nvidia emphasized that this advancement is particularly beneficial for image quality in Performance and Ultra Performance modes, where the upscaler often relies on lower-resolution source images. In addition to DLSS 4.5, Nvidia also announced improvements to its Multi-Frame Generation technology. This upgrade boosts the number of AI-generated frames per rendered frame from three to an impressive five. The new 6x mode, combined with Dynamic Multi-Frame Generation, allows for a flexible number of AI-generated frames. This means more frames can be produced during graphically intensive scenes, while simpler scenes will see a reduction in generated frames, optimizing performance. However, it’s important to note that the enhancements come with certain limitations. The upgraded Multi-Frame Generation function requires an RTX 50-series GPU, as the previous 40-series models can only generate one additional frame per rendered frame, while older models lack this capability altogether. Furthermore, a reasonably high base frame rate is still necessary to avoid lag and rendering issues. Despite these constraints, the technology remains a valuable asset for enhancing the performance of fast-paced games, though it may not be a solution for games suffering from unplayable frame rates.