How AI could cut bandwidth in video conferences by 90%

It's no secret that global Internet traffic has dramatically increased thanks to a pandemic forcing everyone to work and learn from home.

It's difficult to find exact, comprehensive figures, but it's clear that as a result of COVID-19, Internet traffic is up generally by 30% to 50%.

And video conferencing is likely a major culprit. After all, Zoom's revenues jumped by an unbelievable yearly 355% in the second quarter, to $663.5 million, after COVID-19 made the videoconferencing platform part of many people's daily routine. Zoom says it now has over 300 million daily users. Before the pandemic, that number was 10 million.

And that's just Zoom. According to data from Synergy Research Group for the second quarter of this year, spending on unified communications or collaboration tools increased by 7% from the second quarter of 2019, reaching over $12 billion.

That's why Nvidia's relaviately quiet announcement of Maxine is so noteworthy. According to the chipmaker, "Maxine is a fully accelerated platform SDK [software development kit] for developers of video conferencing services to build and deploy AI-powered features that use state-of-the-art models in their cloud."

And what exactly does that mean? "Video conferencing applications based on Maxine can reduce video bandwidth usage down to one-tenth of H.264 using AI video compression, dramatically reducing costs," the company said.

Nvidia explains that Maxine can basically take a snapshot of a user's face at the beginning of a stream and then use artificial intelligence (AI) to map that face across a series of data points. Then, as the conversation commences, those data points are sent across the network and reassembled at the other end into a face. The result is a significant reduction in the bandwidth necessary to transmit video conferencing streams because all that needs to be communicated are the data points rather than the millions of ever-shifting pixels that actually comprise the video stream.

"With AI-based video compression technology running on Nvidia GPUs [graphics processing units], developers can reduce bandwidth use down to one-tenth of the bandwidth needed for the H.264 video compression standard," the company said. "This cuts costs for providers and delivers a smoother video conferencing experience for end users, who can enjoy more AI-powered services while streaming less data on their computers, tablets, and phones."

According to Nvidia, the implications are significant: The company said that more than 30 million web meetings take place every day.

Interestingly, cutting down bandwidth is only one of the applications for Nvidia's Maxine. The company said the technology can also improve video conferencing by supporting instant language translations, raising video stream resolutions, and even using AI to shift users' faces so it looks like they're looking into the camera rather than at their screen.

Nvidia said developers and others can now apply for early access to its Maxine platform.

Related posts:

Mike Dano, Editorial Director, 5G & Mobile Strategies, Light Reading | @mikeddano

Be the first to post a comment regarding this story.
Sign In