New PS5 patent could be a GAME-CHANGER for console rendering technology
PS5 games may look outstanding if the technology from this patent ever comes to fruition…
A new PlayStation 5 patent filed in 2019, first made public on 23 July 2020, hints that the console may utilise powerful image reconstruction technology to improve performance when it launches later this year.
The patent which is available from here, has yet to be fully translated into English, but details from its abstract give some indication into the potential this technology has.
This is the latest in a long line of Sony patents released in the last few months. While some, such as this interesting robot patent, are not likely to come to fruition, we’ve got hope for this one.
What is the Image Reconstruction patent? What does it mean?
As we mentioned above, the full patent is in Japanese, and so much of what we can go off is the abstract, which states:
An information processing device for acquiring a plurality of reference images obtained by imaging an object that is to be reproduced, acquiring a plurality of converted images obtained by enlarging or shrinking each of the plurality of reference images, executing machine learning using a plurality of images to be learned, as teaching data, that include the plurality of converted images, and generating pre-learned data that is used for generating a reproduction image that represents the appearance of the object.
This confusing block of text hints that Sony could utilise a type of AI rendering technology to boost the image rendering quality and resolution while maintaining a solid frame rate.
NVIDIA uses a similar type of AI rendering with its Deep Learning Super Sampling (DLSS) technology to create a similar image overall.
If this technology is present in the PS5, it would lead to a significant graphical upgrade on PS4 games. The PS4 Pro and Xbox One X use checkerboarding to increase the input resolution. This rendering technique relies on nearby pixels to ‘fill in the blanks’ and upscale images, whereas DLSS relies on a neural network and machine learning instead.
NVIDIA explain their “supercomputer trains the DLSS model to recognise aliased inputs and generate high quality anti-aliased images that match the ‘perfect frame’ as closely as possible.”
To put it simply, if the PlayStation 5 utilises this technology, it will benefit the visual output of the next-gen console without relying on 4K or ultra-HD resolutions.
Will the PS5 actually use this patent?
Sony has a habit of filing patents that seemingly never lead to anything tangible, meaning there is a huge possibility that this technology will not be used in the PS5.
As the patent was filed in January 2019, however, the timing of its release does not rule out its inclusion. We’ll be keeping a lookout at the next few PS5 events to see if there’s any reference to this technology.
If you want to see this technology in action, we’d recommend checking out this Digital Foundry video comparing DLSS to Checkerboarding in Death Stranding.