NVIDIA announced a new technology that uses artificial intelligence to improve the quality of low-resolution videos for better display on high-definition screens.
The technology, called RTX Video Super Resolution (VSR), is available in NVIDIA graphics cards through an update to GeForce software version 541.18 released Tuesday.
According to the company, 90% of the videos viewed online are 1080p or lower. When these clips are displayed on a 4K screen, the browser will enlarge the content, which will cause frame errors, which is also a problem that VSR technology needs to address.
This feature is currently only supported by the latest versions of Chrome and Edge, which are based on the open source Chromium browser. Nvidia notes that it has not tested the feature on other Chromium-based browsers.
The company says the feature is available for its Series 30 and 40 graphics cards and will later work to bring the technology to Series 20.
NVIDIA VSR technology supports four levels to choose from, but higher levels may drain your computer's battery and negatively affect the performance of other applications. Therefore, NVIDIA recommends that users choose the layer that best suits their hardware.
Similar to other AI-based techniques for improving video quality, VSR uses deep learning techniques to generate high-resolution images that are close to their original resolution thanks to a neural network trained on a large number of images with different levels of compression that can be restored.
The neural network used in VSR improves accuracy and reduces errors when scaling videos. These defects include fuzzy edges, high-frequency detail, and print aliasing.
Another feature reported to be introduced by Nvidia to improve image quality is DLSS. VSR differs from this in that it is specifically designed to improve video quality.