Gaming graphics have been fundamentally transforming over the past several years thanks to better hardware, modern and more efficient APIs, better engines, and the huge strides made in resolution upscaling. Nvidia was first to the party when it introduced its AI-based DLSS back in 2018. However, AMD has brought competition with its FidelityFX Super Resolution as well. Now, it’s Intel’s turn to show off what it can do with XeSS.
Team Blue first teased XeSS last summer with some impressive statements and a demo to back it up. Things went quite for a while though, perhaps in part due to the ongoing delays related to the Arc Alchemist GPUs. At any rate, we have some fresh info about how Intel’s upscaling technology works, compatibility, and when we’ll see it in some games so that we can try it out ourselves.
Intel’s Tom Peterson led the discussion on XeSS, which some may remember for his DLSS presentations. It’s with this in mind that it’s no surprise to hear about the many similarities between the competing technologies. XeSS uses AI neural networks to train algorithms for resolution uspcaling that can be used in gaming, just like DLSS. In very simple terms, the trained network is then able to take a lower resolution image and fill in the missing details for a final high quality image.
You should really watch the presentation, as Peterson does a great job of explaining it in greater detail in a way that wider audiences will still be able to follow. Spoiler alert, it’s actually magic.
The benefits of Intel XeSS are obvious. It allows a GPU to render frames faster at lower resolutions that involve fewer pixels. Gamers can use it to improve their overall framerates, or increase settings such as effects, shadows, lighting, and
Read more on pcinvasion.com