Over lunch yesterday, my friend Richard was expanding my limited knowledge of spherical cameras: cameras which can take an all-round image – or even video – of the world; an immersive view which you can then explore in a VR headset, or by holding up your phone and moving it around, or (suboptimally) in a regular browser window by dragging around with your mouse, StreetView-style.
Richard’s done a nice blog post about this, from which I’m shamelessly cribbing! Thanks, Richard!
The normal way to get high-resolution all-round footage is to use a special rig like this one:
These are actually not that expensive… assuming that you already have ten GoPro cameras lying around ready to go in them. If not… it’s a different story.
But there are now more compact and affordable options if you don’t mind sacrificing some resolution – things like the Ricoh Theta V. For more information about how it works, Richard pointed me at this great video:
Now, we all know that you can get interesting feedback effects when, say, standing in a lift or changing room with mirrors on opposite walls, or by pointing live cameras at their own output.
What happens when you explore the same ideas in a spherical world? Cool things:
Oh, and if you’re like me, the next question is, “Can you do this kind of thing with live footage?” And the answer is yes. Here’s… ahem… a recording… to prove it…
A final thought.
In the old days, some comedy shows would emphasise that they were “recorded in front of a live audience”, so that you knew that the laugh track wasn’t canned. Soon, we’ll be able to recreate the effect of a live audience so realistically that it won’t be sufficient to persuade you that the recording was actually done in front of real people. Even if you have multiple spherical recordings.
So here’s a challenge to ponder over the weekend: How could you prove that a YouTube video of an event was originally a real event, shown to real people?
While working on my FCPXchange utility, I did some experiments to see how well Final Cut Pro kept track of files if you moved them around under its feet. I was quite impressed with the results:
Now, exploring a bit further, later, I realised that it can’t be using Spotlight, at least not exclusively, to track the file, because it could still find it even when I put it in a folder explicitly excluded from Spotlight indexing.
And then I realised that all my changes had been moves not copies. If I did anything which involved copying the media and then deleting the original, FCP could no longer track it, whereas, if it were using Spotlight metadata, presumably it could track that in the copy too.
Now, I don’t know much about the deeper workings of HFS+, but I’m guessing that it’s effectively tracking ‘inodes’ here, which means that the same bit of content can be found, whatever name it may have in folders. This, however, will only work with the original copy on the same disk. If you start shifting files around between disks or servers, you’ll thwart it!
One of the reasons I am sometimes envious of design/media companies is that they can get away with names that, in other sectors, would cause people at least to snigger, if not positively guffaw.
Can you imagine a law firm, or a steel manufacturing plant, deciding to name itself The Marmalade? Even in the technology world that I tend to inhabit, where many companies, let’s face it, have some pretty silly names, I’m still impressed.
But you can get away with such names if you have other ways to make people take you seriously. And Seb Wills pointed me at this Fast Company post which suggests that The Marmalade may not find that too hard. The embedded video clip, showcasing some of their work, contains some very impressive sequences.