Wednesday, February 01, 2012

eyeIO: New compression technology company signs up Netflix as its first customer

It's not unusual for developers to claim that they've improved the efficiency of video compression algorithms, but they usually result in one of two outcomes:
  1. The changes result in a new compression scheme that's not accepted as a standard, or
  2. The changes don't result in the savings claimed by the developers.
A Palo Alto-based startup is claiming that its new compression algorithms result in bandwidth savings of 20 to 50 percent with better quality, and that the output is 100% compatible with H.264, meaning that it can be supported without changes by tens of millions of existing devices. The correct response to such an announcement would usually be "I'll believe it when I see it", but the company, eyeIO, has signed up Netflix as its first customer.

FierceOnlineVideo reports that Rodolfo Vargas, Microsoft's former Senior Program Manager for Video, CTO of three startups and the former co-chair of Video Streaming and Internet Interactivity at the DVD Forum, approached Netflix with a rough version of the algorithms in September 2010. Netflix tested the prototype with a variety of content, and suggested that Vargas start a company to develop the technology. EyeIO started working with Netflix formally last June, but the companies' partnership was only announced today.

Vargas brought in Charles Steinberg, who's well-known in broadcasting electronics circles from his time as CEO of Ampex and President of Sony's Business and Professional Product division, and Robert Hagerty, the former Chairman and CEO of Polycom, to partner with him. EyeIO's market targets are fairly obvious from the backgrounds of the founders: PC and mobile video, broadcasting and videoconferencing. In addition, there's likely to be strong interest from cable and IPTV operators; eyeIO claims that a single 1TB 7200rpm hard disk can serve more than 400 simultaneous 1080p streams, which would have a big impact on VOD systems.

Netflix won't disclose how much content it has compressed using eyeIO; Vargas will only say that it's "a humongous amount". For its part, eyeIO didn't announce any products or services today, so it's not clear how the company plans to distribute its technology. Will it license its algorithms to hardware and software video compressor vendors, or will it sell its own hardware and software? Will it license its technology to cloud compression service providers? All of that remains to be seen.

I'm very curious to see how well eyeIO's technology actually works in third-party testing, which may come in a few months. For now, all we have is the fact that Netflix is using it--but given that company's bandwidth and storage demands, Netflix's endorsement carries a lot of weight.

Enhanced by Zemanta

No comments: