So a few years back while I was working as Senior Producer for West End Studios, I started producing VR content and 360 video. It was still a very new and unexplored format at the time and we had to invest in some quite pricey kit. On top of that, there was the expensive software required in order to stitch together the video from our bundles of GoPro Hero cameras (I think they were Hero 4). All rigged using a custom 3D printed housing and mounted atop a light stand or magic arm. It looked something like this...
It was quite a fragile, incredibly awkward and somewhat unreliable solution but we got to grips with it and it was exciting to be breaking new ground. Besides, it was kind of fun testing out our results through the Oculus Dev kit or the Samsung headset using a phone, seeing how sick we could make our colleagues.
I remember having to hit the record button on all of the GoPros manually because the remote was a pile of crap and then clapping all around the field of view for synch. Oh, and then there was the battery life and overheating!
Despite all of the challenges, I managed to produce a couple of what I believe to be extremely niche world firsts, so that's something to be proud of. Partnering with Dock10 in Manchester, we recorded the first 360 footage of a music review showcase for local bands and the second was the first hand drawn and animated 360 content to support a global rebrand effort. Actually, you can still see some early experiments on my YouTube here.
Why am I telling you all this? Well, it has been nearly 10 years since I recorded any 360 video content but then, out of the blue, a few months ago I was asked to assist with a pilot research project by the University of Southampton on assessing the impact of coastal erosion on the British Isles. The idea was to follow in the footsteps of BBC's 'The Walk That Made Me' by Chris Packham. We film researchers as they talk to local people gathering different perspectives on the subject of coastal erosion during a short walk around immediate and familiar surroundings. The idea of using an unobtrusive camera mounted on a selfie stick, operated by the interviewer out in the open, simultaneously providing both an intimate and expansive recording, mobile and untethered, with minimum kit and crew was an exciting prospect.
The first thing I needed to do was sensibly price up a two day location job on an impossible budget. This resulted in a crew of 1 (me) using kit already owned by the University and volunteer talent from the researchers. Luckily the university had an old GoPro Fusion camera and a few 32GB cards. There was only one battery which was useful for about 50 minutes of recording 5.2k footage when it was new five years ago and an old, first generation Zoom H4N that took about 5 minutes to boot but functioning, although we did experience interference which I think was actually originating from the cheap XLRs that came with it.
I managed to convince the project lead to invest in extra batteries (unbelievably you can't buy them new anymore but, by a complete stroke of luck, I found a guy local on eBay who had a stack of unused Fusion batteries for sale, never used, still sealed in plastic from the factory but completely flat, obviously) and a couple of 256GB micro SD cards from Amazon (you need two, one for each onboard camera). I knew we would be outside for much of the time and conducting over a dozen interviews (some of which ran to over half an hour) so battery and record time could be a limiting factor.
A few words about this GoPro Fusion. I wanted to use the Insta 360X3 but the budget wouldn't stretch to a new purchase so it was a case of mend and make do. Granny would be proud. The Fusion came out in 2018 I think and it consists of two, back to back 180° cameras recording simultaneously to independent micro SD cards, capable of producing a 5.2k equirectangular 2:1 final image using the accompanying Gopro Fusion Studio software. The big problem however is that the GoPro Fusion Studio software is obsolete and unsupported. It will not run on silicon Macs at all and the only way to make it run on a modern PC is to delete a DLL file from the application library (libeay32.DLL). It took a bit of research to discover this workaround and probably not a solution that most consumers would be aware of. If they do not have access to an old Intel Mac or a PC it would be useless anyway.
So, after hacking the startup library, I discovered that the software will run and operate as intended. This is nuts and a complete insult by GoPro who should be ashamed of themselves for abandoning customers who invested over £500 in this camera while it was available on the shelves only a few years ago. Bastards. this is a shameless and lazy act of corporate greed, forcing a consumers to upgrade to the GoPro Max and new software.
If you are considering a GoPro for an imminent purchase, caveat emptor.
Anyway, apart from facing a live shoot, exposed to the elements, using an obsolete and unsupported camera and stitching software, lacking any real confidence that I could easily or reliably produce useable, equirectangular footage, I also knew that audio was going to be a challenge. As luck would have it, the University also had a couple of old Sennheisser G2 transmitter and wireless receiver pairs (although it turned out the wind jammers were missing, not helpful on a beach in Norfolk).
So, embracing the challenge, we set off and after a 5 hour drive, I could finally get my hands on all the borrowed kit for the first time. An evening to set up, test and charge batteries (oh yeah, no battery charger was supplied, so the only way to charge the batteries was in the camera). This didn't leave a lot of time to run a technical researsal before the first recording at 0900 the next morning.
The Fusion actually delivered for us on the day, I was very relieved when I pulled the footage off the cards. The interviewer found it unobtrusive and manageable on the selfie stick. It recorded reliably but I found the file naming convention and 7 minute record chaptering slightly off-putting when I was browsing the rushes (exactly 7 minutes, not even 4GB, they were 2.8GB) but still, I managed to navigate it. The file naming convention was such that the chaptering prevented sequential listing by file name, which I have never seen before on a camera. For example, recording 1 of 21 minutes would be appended with the numbers 03345, 13345 & 23345. The second clip would be 03446, 13346 & 23346. So, if your folder is organised by name, your chaptering will not be visible together sequentially in a group, this at first induced a panic because I thought the recordings were corrupted, it took me a few minutes to work out. By-the-way, I couldn't use Fusion Studio because I was on location with a Silicon MacBook Pro.
Back in my studio I deleted that DLL file and forced Fusion Studio to run on my PC (pathetic work around GoPro). It did a good job, I was very impressed. So much simpler and faster than the stitching software I remember from all those years ago (not surprising I suppose, it's only two cameras). I didn't time it but I batched up about 15 interviews of varying length from a few minutes to 49 minutes and the software dealt with the chaptering automatically. The stabilisation and auto horizon did a pretty good job and will provide a good starting point for the edit. All the equirectangular files output as 5.2k ProRes422 successfully without crashing overnight from my RTX PC directly to my NAS via 10Gbe. I think it probably took about 2-3 hours.
Once the files were in the edit, the next problem I experienced was audio drift. 5 minutes into a recording the onboard GoPro audio is noticeably out of synch. Not the end of the world as I had recorded audio independently using the Zoom.
All in all, I think what we recorded will hit the spot and the client will be very happy. Subject to clearance, I'll post some of the results after I've delivered.
Ultimately a successful shoot on a very limited budget but not for the feint hearted and a little too much by the seat of my pants for my liking. However, it did prove that a great deal can be achieved on a limited budget, enthusiastic talent and a bit of experience.
If this pilot project is received well and considered a success, I hope I'm asked back to assist with a wider roll out. It is a really interesting and important project, I can't wait to see how our part of it will be woven into the final delivery.
Comments