Apple Acquires WaveOne Startup Working on AI-Powered Video Compression - MacRumorsOpen MenuShow RoundupsShow Forums menuVisit ForumsOpen Sidebar
Skip to Content

Apple Acquires WaveOne Startup Working on AI-Powered Video Compression

by

Apple recently acquired WaveOne, a California-based startup that was developing AI algorithms for video compression, according to TechCrunch.

WaveOne Apple
Apple did not confirm the acquisition, but former WaveOne executive Bob Stankosh said a "sale of the company to Apple" was finalized earlier this year. In addition, the report notes that WaveOne's website was taken offline around January and that several of the company's employees now work at Apple on various machine learning teams.

WaveOne was working on "content-aware" video compression and decompression algorithms to reduce the size of video files, according to the report:

Leveraging AI-powered scene and object detection, the startup's technology could essentially "understand" a video frame — allowing it to, for example, prioritize faces at the expense of other elements within a scene to save bandwidth.

It's unclear how much Apple paid for WaveOne or when the startup's technologies might be integrated into Apple's software platforms.

Tag: TechCrunch

Top Rated Comments

markfc Avatar
markfc
39 months ago
Middle out compression here we come! 😂
Score: 12 Votes (Like | Disagree)
Vol Braakzakje Avatar
Vol Braakzakje
39 months ago
They should invest in AI to improve Siri instead of more compression...
Score: 8 Votes (Like | Disagree)
K
kenny911
39 months ago
Someone finally acquired Pied Piper
Score: 8 Votes (Like | Disagree)
bankshot Avatar
bankshot
39 months ago
If this is prioritizing faces over other scene elements, it sounds like the primary application will be for FaceTime over very low bandwidth connections. Doesn't sound like anything that would be used for general compression of video files or for streaming like TV+.
Score: 5 Votes (Like | Disagree)
Adoniram Avatar
Adoniram
39 months ago
Interesting… AI-based upscaling programs have come a long way (see: Topaz VEAI) and I use them with varied success.

If you could use a known compression algorithm, and train an upscaling (or decompression) algorithm based SOLELY on that known compression, I could see truly remarkable results. Platforms like YouTube would really benefit, especially if the decompression algorithm became a standardized part of various OSes or apps.
Score: 4 Votes (Like | Disagree)
W
WestonHarvey1
39 months ago

Middle out compression here we come! 😂
I hope it has a high D2F ratio
Score: 4 Votes (Like | Disagree)
Read All Comments