Nvidia has demonstrated a new feature known as Deep Learning Dynamic Super Resolution (DLDSR) for its RTX series graphics cards. DLDSR is an improved version of the brand's upscaling 'DSR' feature which can downscale and upscale images for increased visual clarity at the cost of processing power.
What DLDSR does differently from DSR is that it will not consume a substantial amount of processing power. Instead, DLDSR works off of Nvidia Tensor cores, which are in the company's RTX graphics cards, combined with an advanced AI algorithm to upscale a game or video's image without the massive loss of frames-per-second (FPS) caused by DSR. In the demo shown by Nvidia the game Pray ran at 145 fps at native 1080p, at 108 fps when upscaled to 4k using DSR, and at 143 fps when upscaled to 1620p using DLDSR. This means that DLDSR can more than double the pixel count on screen with a negligible performance loss.
AI-Powered Video Upscaling
Nvidia Has Announced 'DLDSR' Technology for Its Graphics Processors
Trend Themes
1. AI-driven Upscaling - AI-driven upscaling technology is set to revolutionize image and video quality by reducing processing power without losing frame-rates.
2. Advanced Gaming Technology - The development of DLDSR showcases the advances being made in gaming technology, through the use of AI and improved algorithms.
3. Efficiency Enhancement - The introduction of DLDSR showcases how AI can be used to improve efficiency and performance, without compromising on quality.
Industry Implications
1. Gaming - DLDSR showcases the future of advanced gaming technology and opportunities for innovation within the industry through the application of AI algorithms.
2. Video Streaming - Video streaming platforms can leverage AI-driven upscaling technologies to offer higher quality video streaming experiences to their users at a reduced cost.
3. Graphic Design - AI-driven upscaling technology presents disruptive innovation opportunities for the graphic design industry, by improving the quality of image processing and editing workflows.