Oculus and their Kickstarter couple years back really brought back and in full force virtual reality goggles. Numerous attempts have been made in the past and never before has it stuck and advanced so much. Even tho Oculus fell far behind their production schedule and bought out by Facebook, they've definitely ignited a new age of virtual reality fury. Dozens of devices are now available, some tethered to your computer and some simply uses a smartphone. Two primary market segments are the gaming industry and film/video industry. While there is quite an excitement about it with the advent of new technologies, I personally still feel it is very much an investors market at the moment. Sure, the consumer market for 360 VR is larger than ever before, but most of the money in this industry is still being spent by investors who have yet to make any real return from the consumer market. Given that 3D movies and glasses was obviously full of hot air before the bubble had even burst, I think the investors are failing to see the same thing for VR. It's great for short tid bits here and there, but far from comfortable for a feature length narrative. Sure some people can game on it for hours, and Frontline and Conan has put it to good use, but the numbers are not impressive. I would venture that most users that view VR goggles over extended period of time will just end up looking forward majority of the time. That said, I have invested a large portion of a year working on the technology. There's only so much information I can divulge as much of the development is still under wraps.
I designed several rigs with various configurations. Most popular rigs involve using GoPros. The biggest downside to that is image quality and most important the inability to sync and manually control the camera. Rigs I worked designed and built mostly revolved around the Red Epic Dragon. No other rig at the time could even come close to the image quality especially in challenging environments such as Linkin Park's concert at the Hollywood Bowl. Sorry, I'm not allowed to post the footage. I made a custom power distribution box that powered 6-8 Red Epics off of a car battery. Unfortunately powering that many Epic's (26Amps) cause the voltage to drop too low too fast even though the battery still had plenty of power. One car battery without any step-up regulator was able to power 5-6 cameras consistently. Machining of base plates and such were sent off to ProtoLabs, who were very quick in production and had very reasonable pricing. I love their quoting system, once they setup the quote for your part you can from their website select various options and materials to see how much each option would cost instantly. I was able to find a Red Scarlet CAD model on GrabCad which I had to then modify to fix some proportions as well as added a canon mount. I drew up the monitor myself and have made it available for download on my GrabCad account (https://grabcad.com/library/red-5-touch-screen-1) as well as the SSD card (https://grabcad.com/library/red-ssd-mag-1). And yes, the rig was extremely heavy.
Various lenses and focal lenses were tested to see which gave best results once stitched.
Early stitching software all had the same tearing issue. Where the software can not match up all the pixels between individual cameras causing a break in what should be a seamless image. It didn't take long after stitching together the first video that a crucial information in the stitching process was absolutely necessary but no one had implemented at the time, depth. It was easy to see that the real money maker is not in the camera rig, but whoever can patent the best stitching method using depth information, be it extracted from the images or mapped along side the cameras. The depth information basically generates a 3D environment in which you can view the images/video, some calling it volumetric video. Today, I'm seeing more and more companies use various methods to extract that information and more and more I'm finally seeing seamless stitching.
You can view one test footage with old stitching technique done a long time ago (footage shot in 2014). On a phone or tablet you can physically move around to look around, otherwise on a desktop you can click and drag.
You can also check out my old office setup, stitched together with pictures taken by a DSLR... yea this is how I work hehehe.
Below is shot with 8 ReplayXD cameras as a very rough and quick 3D test back in 2014. Much have improved since.
Audio in VR
This is something very few people in this space give more than a minute to consider. While you can look around visually most VR video players do not rotate any spacial surround sound audio respectively. Even worse I've seen many productions turn to using ear shaped microphones by http://3diosound.com/, biggest waste of money I've seen for audio. One set is useless in itself already much less those that use 4 pairs to get more surround sound.... There's so many things wrong with using those mics and it's more than I care to write about in this post, you'll need come one over and buy me a case of beers for me to explain.
Currently, there is no real subsitute to standing in the middle of a physical surround sound system to get real positional audio. Sure there are filters and DSP processes that simulate 3D audio for headphones and in ear pieces, but it's just does not work as well. In addition to that very few people realize that on-ear and in-ear headphones have to be processed with different algorithms to get proper 3D audio simulation and very few software offer it.
Either way, much like many 2D low budget productions, audio in VR is mostly an afterthought at the moment. My recommendation, record the surround audio with 6+ pencil mics appropriate to your budget and separetly mix in the talent either on lav or preferably somehow shotgun the talent. You are better off recording clean audio and process it with the vast amount of techologly available now and to come, rather than audio from over priced ear and head molds.
Where does it go from here?
360/VR production has it's uses, but I do not see it making any more gains than 3D did in the entertainment industry, definitely not in any short of long form narrative/feature films. The most useful place and potential for advancement I can see for this technology would be things like geological/civil surveying, environmental mapping (very useful for special effect plates too), virtual tours for the likes of realtors, things of that nature. For quite some time before VR became the latest fad, Google already had a massive start on the competition with their google maps/earth.
