NVIDIA's AI Upscaling: A Game-Changer for Content Quality

NVIDIA's AI upscaling technologies can help free us from low-quality content

We're all constantly chasing better image quality. Cameras aim to provide higher megapixel counts. TVs have leapt from HD to Full HD and to 4K, 8K, and beyond. We've gone from black and white to color and now are seeing HDR expand even further on the color experience. Regrettably, in a lot of areas, the quality of the content simply isn't keeping up with the displays' capabilities. But just as technology can serve as a tool to create better visuals at the beginning of the creation pipeline, recent advances in AI technology are making it possible to create better visuals at the end of the pipeline, just before the final product is displayed. Take streaming video for example: Absolute mountains of content stream at low resolution because they were never recorded at a higher resolution. Many streaming services also don't provide 4K HDR video to PCs, and others still limit the quality delivered based on the subscription tier you're on. So even if you're using a dazzling display with support for 4K HDR visuals, you may end up stuck with a lackluster experience. NVIDIA's AI upscaling technology stepped in several years back with DLSS to to turn lower-quality game visuals into higher-quality ones. It has also worked to help creators develop high-quality content by enhancing their workflows, speeding up previews of realistically ray-traced content to speed up the creation of new content. Now similar technology is in stock for video content. With NVIDIA RTX Video Super Resolution and RTX Video HDR, you can spruce up all sorts of content. Whether you're watching an old 480p video or a 1080p one, NVIDIA RTX Video Super Resolution can analyze the video and scale it up to 4K. The technology leverages recent RTX GPUs' Tensor cores to intelligently expand the image, filling in pixels between the existing ones with a mind for what's displayed in each frame. This avoids the unsightly artifacts common of bicubic upscaling, which more or less just stretches each pixel to fill the right amount of space on the output display. In the same vein, NVIDIA RTX Video HDR can take content with a standard dynamic range, analyze it, and turn it into an HDR video signal for a high-end display to make use of. This can see content that would have looked drab on a modern display get a huge boost in vibrance and show sizzling highlights. These technologies can come in clutch for PC users, as they're available in a lot of the places where you're likely to encounter lower-quality content. Local video playback in VLC can be enhanced by these technologies, and videos streaming in Chrome, Edge, and Firefox can also have both RTX Video Super Resolution and RTX Video HDR applied to them. Of course, not everyone has their PC hooked up to their TVs, which may well be the display most in need of high-quality video sources. So NVIDIA has similar video upscaling technology available on its NVIDIA SHIELD TV and SHIELD TV Pro, which are both seeing discounts for Amazon Prime Day from July 16-17. From the creation of video and game content to the final output, NVIDIA's AI technologies are providing tools to enhance the experience. Creators can work faster, and viewers can upgrade their experience. This all comes from advanced computational capabilities baked right into the heart of NVIDIA's RTX hardware and the many powerful tools NVIDIA has created to leverage that hardware. And DLSS, RTX Video Super Resolution, RTX Video HDR, and NVIDIA SHIELD TV's AI upscaler are just a small set of the AI tools NVIDIA is building. To stay up to speed on all the latest developments and get digestible explanations of new tools and features, tune into NVIDIA's AI Decoded blog here.

Decoding NVIDIA RTX AI-Accelerated Upscaling - ExBulletin

Editor's note: This post is part of our "AI Decoded" series, which demystifies AI by making the technology more accessible and introducing new hardware, software, tools, and acceleration for RTX PC and workstation users. Video is everywhere. Today, approximately 80% of Internet bandwidth is used for streaming video from content providers and social networks. Although screens are getting bigger and supporting higher resolutions, almost all video quality is 1080p or below. The upscaler helps to sharpen streamed video, leveraging the AI of the NVIDIA RTX platform to significantly improve image quality and detail. What is an upscaler? The large file size of video makes it harder to compress and transmit than images or text. Platforms like Netflix, Vimeo, and YouTube get around this limitation by encoding the video, a process that compresses the raw source of the video into a smaller container format. The encoder first analyzes the video and determines what information it can remove to meet the target resolution and frame rate. If the target bitrate is insufficient, the video quality will be degraded, losing detail and clarity and introducing encoding artifacts. Smaller files are easier to share on the Internet, but they will look worse. Typically, software on the viewer's device will upscale the video file to match the native resolution of the display. However, these upscalers are very simple and just multiply pixels to meet the required resolution. While they can help sharpen the edges of objects and scenes, the final video usually contains encoding artifacts and can look unnatural and oversharpened. AI knows a better way The NVIDIA RTX platform uses AI to easily remove video artifacts and upscale. With RTX, you can easily remove video artifacts and upscale. The process of AI upscaling analyzes images and motion vectors to generate new details not present in the original video. Rather than simply adding more pixels, it recognizes and enhances patterns in the image to provide more detail and video quality. Before processing begins, we first need to remove image artifacts. Artifacts, or unwanted distortions and anomalies that appear in video or image files, are caused by over-compression and data loss during transmission or storage. The NVIDIA AI network helps remove image artifacts and eliminate blocky areas sometimes seen in streaming video. Without this first step, the AI upscaler may end up enhancing the artifacted image itself instead of the desired content. Extra Large Video Just as putting on prescription glasses instantly makes the world appear clearer, RTX Video Super Resolution, one of NVIDIA's latest innovations in AI-enhanced video technology, helps users see the world in streaming videos more clearly. Click on the images to see the difference between Bicubic Upscaling (left) and RTX Video Super Resolution (right). Available on GeForce RTX 40 and 30 Series GPUs and RTX Professional GPUs, it uses AI running on dedicated Tensor Cores to remove block compression artifacts and upscale lower-resolution content up to 4K to match the user's native display resolution. RTX Video Super Resolution enhances every video you watch in your browser. Artifact removal combined with AI upscaling technology makes even low-bitrate Twitch streams look amazingly clear. RTX Video Super Resolution is also supported in popular video apps like VLC, allowing users to apply the same upscaling process to offline videos. Creators will soon be able to use RTX Video Super Resolution in editing apps like Black Magics Davinci Resolve, making it easier than ever to upscale lower-quality video files to 4K resolution or convert standard dynamic range source files to high dynamic range (HDR). Welcome to High Dynamic Range RTX Video now also supports AI HDR. HDR video supports a wider range of colors and adds greater detail, especially in the dark and light areas of an image. The problem is, there isn't a lot of HDR content online yet. Simply turn on the feature and enter RTX Video HDR, where our AI network will convert standard or low dynamic range content into HDR and perform correct tone mapping so images look natural and preserve original colors. Introducing AI to every field RTX Video is the latest implementation of AI upscaling, powered by NVIDIA RTX. Members of the GeForce NOW cloud streaming service can play their favorite PC games on nearly any device: GeForce RTX servers around the world first render and encode game video content, then stream it to players' local devices, just like streaming video from other content providers. Members with older NVIDIA GPU-powered devices can also use AI-enhanced upscaling to improve the quality of their gameplay. This means you can enjoy the benefits of both server-rendered gameplay with RTX 4080 class GPUs in the cloud and AI-enhanced streaming quality. Learn more about how to enable AI-enhanced upscaling on GeForce NOW. NVIDIA SHIELD TV takes this one step further, using AI neural networks processing directly on the NVIDIA Tegra system-on-chip to upscale 1080p quality or lower content from nearly all streaming platforms to the display's native resolution. This means that with the push of a button on your remote, users can enjoy improved video quality for content streamed from Netflix, Prime Video, Max, Disney+ and more. SHIELD TV is available now for discounts of up to $30 in North America and $30 or $35 in Europe as part of Amazon's Prime Day event running July 16-17. European Prime members with an eligible SHIELD TV purchase will receive a free one-month GeForce NOW Ultimate membership, allowing you to stream GeForce RTX 4080 class PC gameplay directly to your living room. AI has improved video quality like never before, setting a new standard for the streaming experience. Generative AI is transforming games, video conferencing, and interactive experiences of all kinds. Stay up to date on the latest and next trends by subscribing to the AI Decoded newsletter. What Are The Main Benefits Of Comparing Car Insurance Quotes Online