Today Facebook dropped a bomb on the VR Video world. Things are speeding up even faster.
360-degree Video is hot these days. The world of Cinema/film production, live broadcasting, corporate marketing, documentary filmmaking, and journalism, are all clamoring to figure out where this new medium of expression fits in to their existing ecosystem. Is VR Video a flop like 3D TV or is it a fantastic new way of capturing and viewing the world that will make our flat screens look antiquated in only a few years’ time? Will we wonder in a few years how we possibly lived without it? Are we now in another exciting time similar to the launch of the iPhone, where everyone who gets in early will be innovators and pioneers in a space that will be completely mainstream in a short time?
Only one year ago, there was basically just one 360 camera on the market: The Ricoh Theta. This little pocket-sized marvel of hardware and optical engineering really nailed one thing: It made (sort of low resolution but still good looking on phone) 360 photos. Then they updated it to make videos. It’s still very much the low-end 360 camera, but it has crossed the threshold of acceptable quality for enough use cases that it has become the best selling camera at B&H. Yes, this weird little camera has legs.
Now we have lots of 360 cameras out there. Tiny ones, huge ones. Audaciously expensive ones. Some are for sale, some are not. Some are real, some are vaporware. Some are shipping, some are nearly shipping. The world of 360 cameras is a bit like the automobile industry 100 years ago. All shapes and sizes, no standards, nobody really agrees on what is going to work best or what is going to stick. It is a fantastic world where even small startups are making a big splash and getting the world’s attention.
And today we welcome a new 800 pound gorilla to the scene. This 800 pound gorilla is blue, and it’s giving a thumbs up 🙂
The Facebook 360 Camera
17 cameras, 6K resolution, with 30 gigabits per second of raw capture. This thing is formidable, and as Facebook says, they want to jumpstart the VR video ecosystem by open sourcing the plans for this camera so that anyone can build it themselves. The stitching code will also be open sourced, and it is a doozy: using optical flow algorithms, the stereo (so called “3d”) image derived from these cameras will be first-rate.
What is the end game with this camera, really?
My guess is that this high-end camera will allow studios, production companies, and broadcasters to build their own high end camera to shoot production-quality work at a fraction of the cost of other cameras. This one costs around $30K while other solutions of a similar caliber might cost $60Kor more.
The software that has been built for this camera is not to be underestimated. It is probably far more impressive than the hardware itself, the result of years/decades of research in computer vision, and I expect that the output of this camera, using the image processing pipeline that is also provided open-source with the camera, will really be impressive.
While this 17-lens camera is large and fairly unwieldy for some types of shooting — you can’t wear this thing on your head, or shoot inside a car with it — I would expect that this initial hardware configuration, where many basic issues around hardware architecture and specifications have been addressed, that it will allow the DIY community to adapt this setup with more or fewer lenses to accomodate different scenes and different types of shooting. We can expect other hardware configurations — 6 cameras, 8 cameras, and so on — with adapted software processing to still produce workable stereo — to emerge from this platform.
This is a fantastic day for the world of VR Video and I’m even more excited about our future than before. Facebook is showing they are all-in on the VR ecosystem, and want to help the community of VR video creators evolve and develop this exciting field as quickly as possible.
These are exciting times.