[This article was published on 08 month 0 by the author of the little black box @青, and may not be reproduced without permission!] 】

Gamers are looking for better graphics and higher frame rates, but the two are always a difficult contradiction to reconcile. In order to solve this problem, NVIDIA, AMD, and Intel have launched their own "black technology", that is, frame generation technology.

DLSS: NVIDIA's Technology Pioneer

DLSS stands for DeepLearning Super Sampling, which means Deep Learning Super Sampling, which is an image rendering technology created by NVIDIA for the RTX graphics card series. It uses deep learning algorithms to reduce the burden on the graphics card by reducing the rendering resolution of the game, and then uses deep neural networks (DNNs) to intelligently upscale low-resolution images to high resolutions, thereby enhancing the clarity and detail of the images. At the heart of DLSS is its pre-trained deep learning models that are able to identify key elements in an image, such as edges, textures, and patterns, and improve them as they zoom in.

The strength of DLSS lies in its efficient performance boost and image quality optimization. With DLSS enabled, frame rates can be significantly increased while maintaining or even surpassing the quality of native high-resolution rendering. For example, in the action-adventure game Shadow of the Tomb Raider, DLSS is turned on, allowing players to enjoy a smoother gameplay experience in complex environments. In addition, DLSS also supports a variety of mode selections, such as "Quality", "Balanced", "Performance", and "Super Performance", which players can adjust according to their hardware configuration and needs for the graphics.

FSR: AMD's open-source solution

FSR stands for FidelityFXSuper Resolution, which stands for Super Resolution Technology, and is an open-source super-resolution algorithm implemented by AMD based on the FidelityFX Image Quality Toolkit. Unlike DLSS, FSR does not rely on AI neural network models, but instead uses ingenious math and algorithms to achieve super-resolution of images. FSR enables superior visual quality at a low system performance overhead, allowing games to make more efficient use of graphics card computing resources.

The strength of FSR lies in its open-source nature and wide compatibility. As an open-source technology, FSR can be adopted by more game developers, providing players with a richer choice of games. At the same time, FSR supports a variety of resolutions and image quality modes, allowing gamers to flexibly adjust according to their hardware configuration and needs.

XeSS: Intel's Laggard

The full name of XeSS is XeSuper Sampling, which means Intel Super Sampling Technology, which is a machine learning-based image upscaling and enhancement technology launched by Intel. It is mainly used in the gaming field, and aims to optimize game performance through AI algorithms and hardware acceleration to improve picture quality or game frame rate. XeSS allows games to render at lower resolutions and then leverage AI technology to upscale the image to a higher resolution, significantly reducing the computational burden on the graphics card.

Similar to DLSS and FSR, XeSS also supports a choice of image quality and performance modes, allowing players to adjust to their needs. As Intel deepens its partnerships with more game developers and hardware vendors, the XeSS technology ecosystem will be even better.