Stitch by Stitch
To be sure, Facebook isn’t ready to stream live spherical video from the 1.6 billion people who use its social network. Funneling all that video across the Internet and stitching it together takes time, and the tech needed to shrink this time is not yet in place.
In the beginning, cameras like the Surround 360 will mostly produce 2-D monoscopic video that you can view on a laptop, desktop, or phone. Facebook’s social network and Google’s YouTube can already serve up this sort of video. But Cabral’s camera designs can also generate 3-D stereoscopic images that you can view through virtual reality headsets such as the Samsung Gear and the Oculus Rift. Though these headsets are hardly ubiquitous, they too are here today. The next logical step, as Cabral says, is livestreaming.
Others are pushing in the same direction. Google and Nokia have built professional-grade cameras similar to the Surround 360, which Facebook says you can build with about $30,000 in off-the-shelf parts. Companies like Ricoh offer far less expensive devices for consumers, while at least one company, Orah, already offers a camera meant to do live spherical video (though you should expect less than perfect quality). The difference is that Facebook is sharing its designs.
It’s a strategy the company has pursued before. Facebook shares the blueprints for the computer servers and new-age networking devices that drive its various Internet services on the back-end, and it open sources much of the software that underpins its online empire. In the long run, that helps improve these technologies and drive down the cost of the hardware—all which feeds the progress of the Internet as a whole and Facebook in particular. An open source camera can work in much the same way. But this is new territory. Cabral admits that the company doesn’t quite know what effects the project will have.
Mix and Match
The camera’s design is a little different from others on the market. In addition to the 14 lenses arranged in a ring around the device, it includes fish-eye lenses on the top and bottom, allowing the camera to capture a scene from any angle. And Cabral says that, more so than other 360-degree device on the market, it’s built so that anyone can quickly and easily and reliably capture video out in the field. In other words, he and his team built it not to break. But potentially most useful to the market as a whole is the software that processes the video from all those lenses, stitching together the final image.
In this market, says Shahar Bin-Nun, CEO of HumanEyes, an Israeli company that will soon release a 360-degree, 3-D camera of its own, “the software is more complex than the hardware. Not all hardware manufacturers can build the right software.” His company has no plans to use Facebook’s designs or software, but he believes the stitching software could be useful if it can process video from cameras beyond the Surround 360. Cabral says it can, at least in theory. If it doesn’t work for one developer, they’re free to modify it.
The whole point is that anyone can use Facebook’s open source designs—or just the ideas behind them—to build something new. The possibilities are myriad. Though the stitching software is designed to upload videos directly to Facebook, you could modify it to plug into other services, like YouTube. The software can run on a single (if rather beefy) Mac desktop. But you could also run it across a vast farm of computer servers running the open source Linux operating system, processing videos in bulk. It doesn’t run on Microsoft Windows machines, but according to Cabral, porting the code to Microsoft’s OS is “not a big task.”
Facebook has even shared the specs for the “ruggedized” machine it uses to remotely the drive the camera out in the field, via a fiber optic cable. You could place a camera inside, say, a zoo cage, and then run a cable to this machine on the outside.
What’s not feasible is building a system that could livestream high-fidelity 360-degree video. Even if you had enough computing power to put all those images together quickly enough, you’d still have to move the data across the Internet. For now, the Internet isn’t quite up to that. But Facebook, like many others, is working on that problem. The know-how is there. The market is there. In the end, live 360-degree video is just a matter of more, faster machines and bandwidth. Which is another way of saying it’s inevitable.