Sony PlayStation's lead architect confirmed the use of Machine Learning to enhance frame smoothness, aiming to make gaming more enjoyable, with the earliest chance to see it on the PS5 Pro.
Mark Cerny, the lead architect behind the development of PlayStation 5 and PlayStation 5 Pro, told Digital Foundry that Sony is developing frame generation technology using Machine Learning for future PlayStation platforms. This technology uses AI to create new frames inserted between the actual rendered frames to make visuals smoother. This builds upon the current PSSR (PlayStation Spectral Super Resolution) technology that focuses on upscaling images on the PS5 Pro. The approach resembles PC technologies that enhance frame continuity, though technically it often comes with a slight increase in latency.
An interesting point is that it remains unclear whether the Machine Learning-based frame enhancement will launch for the PS5 Pro or be reserved as a highlight feature for PlayStation 6.
The launch of the PS6 is likely to be delayed due to ongoing RAM supply chain issues, making its release improbable before 2027. Therefore, the most probable scenario is that the AI frame generation technology will debut alongside new hardware in the future.
Notably, if Sony can integrate AI in this way similar to Nvidia's DLSS, it would produce images that are more accurate and visually smoother. Currently, some PS5 games use AMD's FSR3 technology to generate frames, but the key difference is that FSR3 does not use Machine Learning for processing; it relies on traditional frame interpolation methods.
However, industry experts believe that AI frame generation is not a solution for games with very low frame rates, because the system requires a stable base frame rate to provide sufficient data for the AI to generate complete and undistorted new frames.
/source:The Verge