> and **straight up call this one steganography** i concur. the annoying part, is that without the tool used, you cant even determine if there is steganography used in a good implementation, so no way to even start using brute force guesses on passphrase or even start using a rainbow table. Real spooks know all about rainbow tables though. Real spooks would not use merely three word passphrases because even clearnet rainbow tables available have 3 word/digits now in 2021, and some rainbow tables have all typeable glyphs up to 12 letters long. 7 letter open value rainbow tables take 744 TB according to disinformation on wikileaks and other sites : see : https://www.google.com/search?q=biggest+rainbow+table+digits Google is all too eager to take over main search result to spread disinfo of 744 TB, as seen if you do that link. 7 bytes is (256^8) ore a total combination of bits equating to (256*256*256*256*256*256*256) or (7.2057594e+16) byte combinations implicit. 72,000 terabytes, but the people curating them tend to use only standard ASCII, and ascii high bit is "undefined" (not zero) and is 7 bits, and the people making and trading rainbow lookups tend to ignore control characters other than space : https://project-rainbowcrack.com/table.htm so on that page it shows a 10 character passphrase comprised of "loweralpha-numeric#1-10" being 316 GB, per assumed hash algorithm and computations for three famous hashes shown, all being the same 316 GB size naturally. you can even download the 316 GB rainbow tables and use them to immediately get password that will work for a windows laptop with a lost password. I used "L0phtCrack" a few times since 2000 to many grateful scatterbrained users. Nowadays you would use freeware versions ( ophcrack, Cain and Abel, or John the Ripper according to this page : https://sectools.org/tool/l0phtcrack/ ) cryptome admins would not hide in the light so brazenly, so I can only assume he is baiting nation states to waste money processing those blatant steganography-payload files online. The payload might just be a fortune cookie like : **"Which department reimbursed you NSA boys this time. $$$ !"** As for your find of an obvious stenography test : > Also, check out this creepy video from 5 years ago. I'm going to go ahead and straight up call this one steganography: https://www.youtube.com/watch?v=NP3UMcTcAdY He is employing white and pink noise mixed with loud and amplitude-clipped "organic derived pinknoise" not CPU algorithmic, to blind out the YouTube anti-steganography filters, and the constant scene changes in your link are to mimic ROCK MUSIC VIDEOS to also be whitelisted and spared from more aggressive CIA/FBI anti-steganography which evidently spares pop music videos perhaps. Very amusing to a cryptopunk to watch that. We know he succeeded because all failures would have been downloaded and reuploaded with less data payload. You are looking at success of the most possible "signal to noise" stenography state of the art for the year 2015, assumedly. Bravo for him. pink noise : https://en.wikipedia.org/wiki/Pink_noise I would have left audio alone, and only gone after using an exotic legal mp4 file format that allows lots of razzle dazzle in the h264 file format, including separate channels and such. Some funny engineers put ENTIRE TOKEN streams into mp4 and youtube left the nonstandard data alone and entire mini trucrypt hard drive images got inserted! https://appliedtech.iit.edu/school-applied-technology/projects/mp4-steganography That "TCSteg" guy thinks like me. Why stick out with accoustic noise unless delivery meant as rapid data insertion into a phone app by CIA that uses ***microphone only**, and meant for a "coffee shop" rented machine. Avoid white and pink noise to be more covert, unless needed to perversely disable youtube reprocessing.