AMD announced FidelityFX Super Resolution 2.0 (FSR 2.0) in March of this year. Unlike NVIDIA’s Deep Learning Super Sampling (DLSS), FSR does not use AI to train models, but uses traditional image processing algorithms. The advantage is that it is easy to integrate into games and supports old graphics cards, even those of competitors. DLSS requires specialized hardware on the Nvidia graphics card to reconstruct the image, and it can even reconstruct a higher pixel count than the native image. FSR relies on and is subject to the basic pixels of the game, and it is difficult to really challenge DLSS. The source code for FSR 2.0 was published on GitHub last week under the MIT license. Last weekend mod developers released the FSR 2.0 version of Cyberpunk 2077, which can replace DLSS for users with older Nvidia graphics cards.
This article is reprinted from: https://www.solidot.org/story?sid=71957
This site is for inclusion only, and the copyright belongs to the original author.