Interesting. I don't think it's a numbers station-style system. Generating videos and uploading them via relatively complex protocol (where you have the potential to fuck up and leak info about yourself) to a server you don't control seems like a very roundabout way of broadcasting information to someone. Once you bring computers and the internet into the mix there's a huge number of simpler and safer ways to broadcast encrypted information without requiring particular equipment or expertise of the receiver.
Great podcast! I didn't actually learn anything new but the producer has that great talent of making it interesting anyway. It reminds me of the better documentaries we sometimes get on BBC Radio 4 in the UK.
Glad you liked it! Podcasts are finally giving US creators the chance to match, or at least approach the quality of BBC content. There's a joke going around, two guys are sitting at a table recording a podcast and one of them goes, "Hey, remind me, are we recording your podcast or mine?"
From the description, it sounds like a Webdriver Torso-style account that was accidentally set to public. Muddied images like from an out of focus camera, items that last only a few frames, distorted audio. All things that would let someone tune codec settings over difficult corner cases. Once the account was discovered and publicized, they deleted the account, probably with someone being chided before the test corpus was set back up on a properly locked down channel.
The last time I heard about something like this, it was the automated output of a machine learning project. Even if there is an interesting purpose, you have to be really bored and have an active imagination to worry about it.
There are apparently two videos being uploaded per minute, so it's almost certainly something automated. No human being would sit there and upload videos every 30 seconds all day.
> It's probably a channel dedicated to a paint-2-sound generator. Example (that's very old software: Coagula). Pixels actually represent the spectrum over time (X), and the color intensity the magnitude at frequency (Y)
>
> But what is its purpose?
>
> Synthesize mouth-like sounds. Anyway that's just an hypothesis. But the pseudo of the guy matches to something in the domain, aka the sampling theorem (unfavorable semi sircle: full circle = full spectrum = 2PI, but in sound over 1PI = nyquist frequency, there's the folding phenomena).
Man I can't get into this stuff at all because the curiosity and uncertainty alone would kill me. This thing, the cicada thing, that reddit number subreddit thing. Ugh.
The curiosity is what makes it interesting! People have been spending tons of effort analyzing the Voynich manuscript for a hundred years. If someone suddenly solved it, it would quickly become uninteresting.
