Nvidia is launching a new feature for all RTX GPU owners today: RTX Video HDR. Much like how Nvidia’s RTX Video Super Resolution can upscale old, blurry web videos, RTX Video HDR uses AI to convert SDR color space video to HDR. You just need an HDR10-compatible monitor with HDR enabled in Windows.

RTX Video HDR is part of today’s 551.23 Game Ready driver release for the new RTX 4070 Ti Super launch. The Video HDR feature works in both Microsoft Edge and Google Chrome, and during my testing on an HDR600 monitor, I noticed it added subtle color detail to 4K YouTube videos that weren’t uploaded in HDR. The effect might be more obvious on a better HDR monitor, but it was difficult to notice unless I toggled it on and off multiple times. I found similar results when I tested RTX Video Super Resolution previously to upscale old YouTube videos and sharpen them up.

If you have an HDR monitor and an RTX card, you can download the latest drivers from Nvidia and head into the Nvidia Control Panel to enable RTX Video HDR. It’s available under the Video > Adjust video image settings in the Nvidia Control Panel, in the same location where you can enable RTX Video Super Resolution.

Like many of Nvidia’s AI-powered features, RTX Video HDR requires an RTX card since it leverages the tensor cores that Nvidia first started shipping in its RTX 20-series GPUs. Nvidia also uses AI techniques for its Deep Learning Super Sampling (DLSS) system to improve frame rates and image quality in games. Nvidia also launched its Deep Learning Dynamic Super Resolution (DLDSR) system in 2022 to improve image quality by rendering games at higher resolution. Nvidia Broadcast also includes an Eye Contact feature that uses AI to make it look like you’re making eye contact during video calls.

Nvidia’s latest 551.23 driver release also adds support for the Ultra Low Latency Mode with DirectX 12 games. It’s a mode that’s worth enabling in games without Nvidia Reflex support, as it reduces the render queue directly at the driver level instead of the game managing the queue.