A long, techy post about the 360 videos I’ve been accumulating…a departure from the quick updates that I’ve been trying to confine them to.
Along with posting these blog entries, I’ve been uploading time-lapse videos of days working in the woods shot using a 360 camera. The videos shown there are usually “reframed” so that they’re viewed as standard 2D videos pointing in a direction that I set while editing the 360 video after the fact. They’re also typically sped up faster (than the 30x speedup at which they’re initially recorded) by another 2-8x because at even 30x, a 8 hour work day results in a 16 minute video. I barely find them entertaining enough to be worth watching for 2 to 8 minutes, so that’s how much I speed them up.
Having a barely-fun-enough-to-be–worth-watching video has been a secondary objective of these recordings. My main objective is to keep track of what I’m doing, where I did it, and when. So I’ve also been recording my location with the GPS on my phone. I sync the start and stop of a 360 video recording with the start and stop of the corresponding GPS track recording. I upload the video to YouTube, and then serve it up via a web-app that I’ve been working on for various similar purposes, along side an interactive map showing the track traced out by my phone as of the point-in-time in the video. It’s kinda hard to describe in words.
Here’s a screen recording of it in action.
Actually, it’s still hard to get a sense of the interactivity of the web app in a screen recording.
Here’s a FB Post describing a publicly deployed version of the app, recording a recent bike ride.
For the Frankenberry geo-video digest version of the app, I’m going to add a way of browsing or searching over past recordings based on the GPS tracks. Then, in the future, when I’m evaluating the outcome of some work I did, e.g. the technique I used in a specific location for combating Holly and Ivy, I can hopefully easily find footage of what I did. It’ll be fun to figure out how to present the tracks in a way that helps quickly visualize how much time I spent in an area (heat map?), and to play with ways of searching by choosing a location and ranking track-day results based on time spent on a that day near that location (and how old that day is?). Hopefully it will also prove useful.
At 30x speedup…I’ve settled on a standard of 1 frame shot per second, played back at 30 frames per second…I won’t get a lot of fine grained detail into, say, exactly how I tackled holly root extraction that day, but my hope is it’ll show enough to help jog my memory of this. Regardless of memory-jog success, it’ll at least give me some idea of how much time I invested in that area.
Even at 30x speedup, the footage takes a lot of storage space. Each day’s 360 video is ~15GB, and storing and serving up many days of that from Firebase (which is what I’m using for the web app) would be a bit costly. Even just storing it in Google Drive would quickly eat into the 2TB for which I currently pay $100/year. Here, by using YouTube for hosting the videos, and accepting the confines of YouTube’s IFrame Player API I stay almost completely in the “free” tier of cloud usage and incur zero cost for hosting the video content. YouTube allowing me to store and serve up 100s of GB of videos for free honestly seems almost too good to be true…they’re “Unlisted” and I don’t see how they’ll make any money off of them to pay for their storage and bandwidth costs. Maybe it’s really just so insignificant to them that they don’t care, or maybe I’ll find out that there is a limit.
As with how useful any of this will turn out to be (or if I even actually end up building any of it), time will tell…
If you’ve gotten this far, thank you. If you have any thoughts, I’m happy to hear from you.