So you’re saying use the pixels as ones and zeros, so if I had just for example, a Microsoft Word document, I don’t use Microsoft Word but that’s beside the point, you’re thinking include the document into a video? I don’t know if that’s actually possible but damn, that would be kind of cool. Because I also have pictures and documents. But be cool if you could do that.
litchralee@sh.itjust.works 6 hours ago
Steganography is one possible way to store a message “hidden in plain sight”, and video does often make a seemingly innocuous manner to store a steganographic payload, but in that endeavor, the point is to have two messages: a video that raises no suspicions whatsoever, and a hidden text document with instructions for the secret agent.
Encoding only the hidden message as a video would: 1) make it really obvious that there’s an encoded message, and 2) would not be compatible with modern video compression, which would destroy the hidden message anyway, if encoded directly as black and white pixels.
When video compression is being used, the available bandwidth to store steganographic messages is much lower, due to having to be “coarse” enough to survive the compression scheme. And video compression is designed around how human vision works, so shades of color are the least likely to be faithfully reproduced – most people wouldn’t notice if a light green is portrayed slightly darker than it ought to be. The good news is that with today’s high resolution video streams, the raw video bandwidth is huge and so having even just one-thousandth of that available for encoding hidden data is probably sufficient.
That said, hidden messages != encrypted messages: anyone who notices that there may be a hidden message can try to analyze the suspicious video and retrieve the payload. Encoding, say, English text in a video would still leave patterns, because some English letters (and thus ASCII-encoded bit patterns) will show up more frequently. But fortunately, one can encrypt data and then hide it using steganography. Encrypted data tends to approximate random noise, making it much harder to notice when hidden within the naturally-noisy video data. But bandwidth will be cut some more due to encryption.
TL;DR: it’s very real to hide messages in plain sight, in places people wouldn’t even think of looking closely at. Have you thought about the Roman Empire today?
vrek@programming.dev 4 hours ago
You know if somebody did this with an encrypted data but only like every 50th frame on a video like that old “smiling man” videos or they “watch Ian sleep” videos then someone saw it and posted it either here or Mastodon or one of the forbidden companies it could generate a new conspiracy.
Would be great if a bunch of people did like a full deep dive. Getting the raw videos, figuring out the frames to extract, determining the encryption method, find a method to decrypt it, combine all the fragments back together, figure out its in some obscure latex format, write a custom interpreter for that format which hasn’t been used since the early 90s…people online thinking it’s some state actor or like a command and control module for some bot net… They finally crack it all and access the raw data and it’s just a backup of pornographic slashfic from usenet back in the day.
Especially if breaking the encryption was like a novel and new approach. Imagine like a talk at defcon “how I found about Harry Potter’s sex with Gwen from Spiderman…and broke aes256 in the process”