- Nvidia has released its Maxine developer platform to improve video calls.
- The platform enables face alignment and eye correction for more natural conversations.
- Other supported features include dramatic bandwidth savings and video upscaling.
The COVID-19 pandemic means that video-calling has become more important than ever, as companies hold virtual meetings, schools hold online classes, and families connect over the internet in lieu of physical visits.
This widespread shift to video calling has laid bare a few major challenges for the tech, and Nvidia has now announced a solution in the form of its Maxine developer platform (h/t: The Verge).
According to the graphics colossus, Maxine is a suite of video conferencing software powered by Nvidia GPUs in the cloud. And the list of enhancements is pretty intriguing for the most part.
A more natural video-chat experience
The first two major features made possible by Nvidia’s Maxine are gaze correction and face alignment. We’ve previously seen the former on the likes of iOS, but the two features essentially simulate users looking at their camera in order to enable a more natural conversation. This theoretically means a participant doesn’t need to keep looking at the camera and can fully focus on looking at the other participants.
The platform is also able to slash bandwidth requirements by analyzing key facial points of video call participants. From here, the software then “intelligently re-animates the face in the video on the other side” using said facial points as a reference rather than streaming the entire screen over the internet. Nvidia says this technique can reduce bandwidth consumption “down to one-tenth of the requirements of the H.264 streaming video compression standard.”
What else does Nvidia Maxine offer?
Super-resolution is another area of expertise for the brand, as its Shield Android TV gadgets use AI upscaling to boost video quality from standard definition to 4K. A similar principle is in effect for Maxine, as Nvidia demonstrated the ability to boost a 360p video call to 720p (see the video at the top of the page).
The company also touted auto frame capabilities, ostensibly cropping into the camera in order to let it “follow” a video call participant as they move around. Other features enabled by Maxine include those ever-popular virtual backgrounds, the ability to filter out background noise, language translation, and virtual avatars.
Nvidia and video collaboration app Avaya have announced a partnership to use the technology, but we hope to see Maxine supported by other major players in some capacity. We’ve also recently seen the company release an RTX Voice app to filter out background noise, so here’s to some of these other features being made available to end-users in a similar way.
