Apple Acquires Startup That Uses AI To Compress Videos

Image credit: Lauren Desicca/Correspondent/Getty Images

Apple quietly acquired Mountain View-based startup WaveOne, which was developing AI algorithms for video compression.

Apple did not acknowledge the sale when asked for comment. However, his website at WaveOne shut down around January, and several former employees, including one of WaveOne’s co-founders, are now working in various machine learning groups at Apple.

WaveOne’s former head of sales and business development, Bob Stankosh, announced the sale in a LinkedIn post published a month ago.

“After almost two years at WaveOne, we finalized the sale of the company to Apple last week,” Stankosh wrote. “We started our journey with WaveOne when we realized that machine learning and deep learning video technology could change the world. We have added our own technology to his portfolio.”

WaveOne was founded in 2016 by Lubomir Bourdev and Oren Rippel. The two set out to embrace the decades-old paradigm of video codecs and leverage AI. Prior to joining the venture, Bourdev was a founding member of his AI research division at Meta, where he and his Rippel worked on Meta’s computer responsible for his moderation, visual search, and feed ranking of content on Facebook. I worked on the vision team.

When it comes to standard algorithms for video compression and decompression, compression happens on the content provider side (like YouTube servers) and decompression is handled by the end user’s machine. While this is an effective approach, new codecs require new hardware specifically built to speed up compression or decompression, slowing the propagation of improvements.

Image credit: WaveOne

WaveOne’s main innovation was a ‘content-aware’ video compression and decompression algorithm that could run on AI accelerators built into many mobile phones and an increasing number of PCs. By leveraging AI-powered scene and object detection, the startup’s technology is essentially able to “understand” a video frame, e.g. You can prioritize faces at the expense.

WaveOne also claimed that its video compression technology is robust against sudden interruptions in connectivity. This means you can make a “best guess” based on the available bits, so your video won’t freeze if your bandwidth is suddenly capped. Details are simply not displayed during the period.

WaveOne claimed that their hardware-agnostic approach could cut video file sizes by up to half, resulting in better effects in more complex scenes.

Apparently, investors have realized the potential. Ahead of the Apple acquisition, WaveOne raised his $9 million from backers including Khosla Ventures, Vela Partners, Incubate Fund, Omega Venture Partners and Blue Ivy.

So what does Apple want from AI-powered video codecs? The obvious answer is more efficient streaming. Even small improvements in video compression can save bandwidth costs or allow services like Apple TV+ to deliver higher resolutions and frame rates, depending on the type of content being streamed.

YouTube already does this. Last year, Alphabet’s DeepMind adapted a machine learning algorithm originally developed for playing board games to the problem of YouTube video compression, reducing the amount of data video-sharing services needed to stream to users by 4%. reduced.

Perhaps soon we will see similar innovations from the Apple-owned WaveOne team.