Exploring 360 video

360 Video, is it really VR?

I’ve heard many a debate over whether 360 video can really be classified as VR.  Even though the video can be viewed in a VR headset, in 360 degrees, some believe that it can’t truly be classified as “Virtual Reality” because you can’t engage with the video.  This is a good point, so what does it take to make 360 video classified as VR? This will most likely be an on going discussion.

360 video has been around for a while now, but it still has yet to find its feet in mainstream content creation.  There are a lot of cool projects and interesting experiences out there utilizing 360 videography, but it’s still almost considered a fringe tool.  

Quality of 360 cameras

Over the recent years there have been more cameras becoming widely available from companies like Kandao, GoPro, Insta360, and Ricoh that have 4K and even up to 12K resolution.  However, with an increase in resolution there is a steep increase in cost.  Cameras in the 8K and 12K resolution ranges are thousands of dollars and are primarily angled towards professional content makers.

I got my hands on the first release of the Ricoh Theta back in 2016.  It was definitely a lot of fun trying out new ways of capturing music and video.  So I brought out my recording and video gear to a reunion jam night by a friends band, Pangea.  I quickly found out that the technology for 360 video had a long way to go.  Even in the world of streaming and audio tools.  (read my blog about ambisonics when you have the time)

It’s unfortunate that the camera’s quality is so subpar, BUT it’s still a very enjoyable experience to get inside Pangea’s songs none the less!

Below is a clip from the recording session with Pangea using the first iteration of the Ricoh Theta. As you can see due to the dim lit room and the Theta’s limited resolution the video severely lacks definition.  If you watch the video on a a larger screen than a phone, the softness of the video is quite prominent.

When the video is outputted from the camera in a 1080 viewable video it doesn’t look too bad on a phone, but once it’s viewed in 360 on a larger screen like on a tablet or computer it is obvious that it lacks much needed resolution.  This is the problem with many early 360 cameras. 

If you stretch one 1080 video “around” the viewable space you’re essentially getting less than SD quality.  That and add in issues with loading quality and battles with full screen/landscape mode in 360 YouTube players the experience -is more often than not- less than desirable.

Since the quality of the video was a difficult hurdle to surpass, it forced me to develop a creative solution.  The name of the following song is “Paralyzed Eyes.”  It felt appropriate to try a split perspective approach to showing 360 video in the 2D plane.  You can see the outcome of some interesting experimentation with editing techniques in the following video.

I wouldn’t recommend watching either of those videos on a VR headset, but on a mobile device it is still passable(?)  I’m looking forward to getting my hands on something like the GoPro MAX, or the Kandao QooCam to get a chance at making higher quality content.  Only thing is they’re fairly expensive just for demoing purposes.  

After seeing how poor the quality of the Ricoh Theta was I tried out multi Gopro set ups, which was a costly endeavour.  Again I was less than happy with the end result as action cameras have poor low light video quality.  The resulting video fell short of quality expectations, although I was able to capture a much higher resolution video quality.  I hope to get that project online to share soon.

Because the cost of getting a quality 360 camera is definitely a detriment for me, I had an idea.  There are great open source communities online for getting your hands dirty and building your own tech.  You can buy camera sensor modules from places like ArduCam -a website that sells parts for simple open source cameras- and build your own camera based on the Raspberry Pi or NVIDIA Jetson.

The Raspberry Pi and NVIDIA Jetson Nano are extremely popular in the open source tech crowds

Now, one of my next projects will involve ordering a couple 4K sensors, a PTZ kit, and some fisheye lenses.  Put it all together to run off of a Raspberry Pi, and I can customize my own 360 camera for a fraction of the cost of an 8K camera on the market.   And since I’m already building a track based AI robot for my son this year for his birthday I thought why not try mounting it to his robot so you can drive it around and view the camera through a VR headset!  Much fun will be had.

Now imagine driving a robot around an art gallery and being able to view everything from the comfort of your own home.  Would that count as “engage-able” “virtual reality” 360 video?  I want to find out.