Frame generation is established tech in games at this point. We already knew that Sony was looking at the tecnology for the PlayStation 6, but the company has filed a new patent that gives a better idea of how it'll work.

The patent, titled "Generation Super Sampling," (US-20260320740-A1), describes a mostly standard form of frame generation, where the model takes information from a traditionally rendered frame and then uses that information to generate an interpolated frame. The major difference is in how the model will be predicting how the frames will change.

Whereas DLSS frame generation works by taking rendered frames and then pulling motion vector information from the game engine, the suggested PlayStation model will also take input data from your controller to more accurately predict how you're going to be moving from frame to frame. The hope, according to the patent notes, is to limit the amount of latency or 'hitching' from frame to frame.

The patent also leaves the door open for multi-frame generation, which was introduced on PC with the RTX 5090 back in 2025. According to the patent notes, "In some embodiments, the predictions are not limited to an immediate next image in the image stream. The decoder system may be trained to predict any number of images before or after the input image."

It's hard to imagine why a console that'd presumably be attached to a TV needs to generate multiple frames, though. With PC gaming, multi-frame generation is only really useful when you're playing on a super-high refresh rate gaming monitor at 144Hz or higher.

Source link