Since we didn’t evolve with eyes on the backs or the sides of our heads, it’s no surprise that Bill Banta and his team had trouble anticipating all of the problems they would run into with their 360-degree video camera.
The former Apple procurement manager—who was in charge of more than $500 million in annual spending on parts for iPhone cameras—is the CEO of Centr Camera, a San Francisco startup with a broad vision. Literally. The company is developing a handheld, torus-shaped device with four built-in video cameras that capture the surroundings in four directions at once; software inside the camera stitches the pictures together into a single, very wide video stream (at the camera’s highest resolution, it’s 1,080 pixels high and 6,900 pixels wide).
There was just one little problem with the prototype version of the device. Unless users mounted it on a tripod or a helmet, they had to hold the camera by the side—and that meant there was always a hand blocking part of the picture.
“No one wants to see your hand,” Banta says. The problem became acutely noticeable after the company, previously known as Stealth HD, distributed 100 copies of its prototype to beta testers. Somewhere in every beautiful wraparound view, there would always be a big ugly hand, like the finger in the corner of a clumsy snapshot.
“We realized we had to come up with a shape that allowed you to hold the camera without disrupting the view,” Banta says. The result is the new Centr camera, which looks like a black plastic donut and fits perfectly over your thumb. The company pledges to deliver the earliest “developer” units to backers of its new Kickstarter campaign by November, with production units to follow next February.
If the selfie was the big personal photography fad of 2013-2014, walking around with a toroidal video camera on your thumb could be the big thing of 2015. “We designed it with the idea of a ‘GoPro Plus,’” Banta says. “We said, ‘How do we create a device that people can use to take awesome imagery, like they do with a GoPro, and then allow them to do more, to take these panoramic videos?’”
The goal of Centr’s Kickstarter campaign, launched one week ago today, is to raise $900,000—which is on the high end for a crowdfunding campaign, but not an outrageous amount for a major hardware project. As of this morning, the company had already collected $349,000 in pledges, taking it more than a third of the way to its goal, with 24 days remaining in the campaign.
Most backers are pledging $299 to reserve their own Centr camera, which means the company needs about 3,000 interested customers to meet the goal. Banta hopes a publicity blitz will keep the pledges flowing in. Tomorrow’s stop: the Hardware Alley section of TechCrunch Disrupt in New York.
I first met Banta in 2012 at Stanford, where he was finishing his MBA and Stealth HD had just won $20,000 in a student business-plan competition. The company’s idea at the time was simply to build a portable camera with enough processing power to merge four high-definition video streams in real time, in contrast to older 360-degree-view technologies that stitch recorded images together in post-processing. (The computationally difficult part is making the four images line up at the seams and achieving a consistent exposure and color balance across the four cameras.)
Stealth HD sent early versions of the technology to customers like sports broadcasters, who used it on soccer fields to capture more of the action, and the U.S. Army, which tested a helmet-mounted, night-vision version of the camera for reconnaissance. But even as the startup raised seed funding, won government contracts, and grew to more than a dozen employees and contractors, its focus shifted to the consumer market.
The phenomenal success of GoPro—the San Mateo, CA-based startup that has captured about half of the U.S. camcorder market and recently announced plans to go public—had demonstrated that people love to make videos of their outdoor adventures, from surfing to snowboarding to stunt piloting. Banta’s team reasoned that even in these civilian situations, four cameras are better than one. By recording a 360-degree view of the action, their device would free users from the need to keep the camera pointed at something interesting. They could find the best sections later, using standard video editing software like Final Cut Pro, Adobe Premiere, or iMovie.
That was the theory, anyway. In January, the company assembled 100 prototypes and gave them out to testers. “They all broke immediately,” Banta says. “They froze up, or were recording corrupted video, or the cameras would get out of sync or adjust independently for exposure. There were some late nights getting that resolved.”
After some updates to the software inside the cameras, “the content started to look good, and we were able to learn a little more about best practices,” Banta says. It turned out that holding the camera level to the horizon is fairly important; otherwise, strange distortions in perspective result. Testers told the company the prototype was way too heavy, and that it needed raised buttons that they could feel even if they were wearing gloves. They also wanted to be able to change out the battery pack in the field, rather than having to plug in the device to recharge it.
Then there was the hand-blockage problem. Quite a few testers wanted to use the prototype to make selfies, but if they held the device at arms’ length, their own hand would always be in the way.
In short, the prototypes turned out to be far from production-ready. But building 100 of them was worth the effort, Banta says. “If we had just built five, we wouldn’t have run into these issues,” he says. “The feedback was hugely valuable.”
Based on that feedback, the company decided to scrap almost everything—including its old name—and came up with a new, smaller design built around next-generation mobile image processing chips from Movidius. The San Mateo company’s “vision processing units” or VPUs are also inside Google’s Project Tango phone, an experimental Android device designed to make it easy to model 3D spaces.
The camera Centr plans to manufacture using funds from the Kickstarter campaign will have four 5-megapixel apertures, a thumb-size hole in the middle, a thumb-activated record button on the top rim, a removable battery back, and a digital “bubble level.” (It’s a cool-looking ring of LED lights that turn green when the device is level, and red when it’s not; the effect is reminiscent of the power coil implanted in Tony Stark’s chest in the Iron Man movies.) The device will connect via Bluetooth to a smartphone app, where users will have the ability to change camera settings and preview what the sensors are seeing.
The thumb-friendly center hole can also accommodate an expanding gasket that attaches to a tripod or a GoPro mount. That means the device can be bolted to, say, a bicycle or a skydiving helmet.
Banta says Centr picked the $900,000 number carefully when it was preparing for its Kickstarter campaign. The idea was to stay below the million-dollar level, which Banta calls a “psychological barrier,” while at the same time asking for enough money to get the device ready for mass production—a classic stumbling block for hardware projects on Kickstarter. For Centr, the to-do list will include tasks like integrating the Movidius chips into the camera’s motherboard and getting all of the company’s existing image-processing algorithms running on them.
“We wanted to not set a goal so low that we wouldn’t be able to deliver,” Banta says. “We also needed to know that there was enough support from a market standpoint. There are always 500 people in the world who will buy an awesome gadget. But if we can get to 3,000 backers, we will feel better about the market as a whole.”
The company picked a new moniker to go along with its Kickstarter effort because “we wanted a name that reflected what the camera could do,” Banta says. “We’re able to create this 360-degree video that you experience like you are in the center. So Centr makes sense as a name and a brand.”
But unwrap the wraparound video in an editing program, and you get a panorama—and Banta says one of the most unexpected surprises from the beta-testing period was the emergence of what he calls “the panoramic selfie.” “People really like to pull pano-selfies out of these video,” he says. “Those are images you can’t get with your GoPro or your iPhone, and would be very difficult with any other type of camera.”
Confirming, perhaps, that today’s generation of amateur digital photographers and videographers still see themselves at the center of the world.