Infinite-Storage-Glitch
Infinite-Storage-Glitch copied to clipboard
DvorakDwarf
Audio too?
Seems like there's an additional storage medium here waiting to be explored in the audio tracks. Can surround sound be used for more individual tracks on a single video? I'd be interested in seeing just how much more data can be stored in the same amount of video.
The problem is the compression and degradations in bitrate, so some sort of forward error correction would be needed, which would reduce how much could be transmitted via audio. That said, these videos are trippy enough, imagine coming across one that also had an audio track, all of it noise? And it turns out it all encodes a file? lol, would blow people's minds. "What is this witchcraft!"
Also, it would require additional audio tools; this already integrates OpenCV, so it's kind of complex enough already. Plus, I haven't found a lot of good FEC algorithm crates already packaged for Rust, like wirehair, so that might increase the difficulty of this ask, since it'd require several dependencies; FEC encoding to protect against lossy data in streams in advance, a library that can turn audio into a useful byte stream, and then a library to convert bytes into an audio stream of a specific codec, plus the integration with the video container like MPEG.
"What is this witchcraft!"
You mean like a modem audio codec? Try piping the data through : http://www.whence.com/minimodem/ The data rate will be a little slow, but it will be more than zero.
I was thinking about the same thing recently, maybe try using audio using the same logic as videos in binary mode? As in sound for 1 and no sound for 0 with a minuscule interval between? Although I doubt a lot of data can be encoded this way. This might help result in a final video size that's less as well, as in x% of data encoded through video and 100-x% in audio and then merged through ffmpeg?
Actually I was wandering whether something like QAM used for both audio and video could improve entropy :)
Adding audio would make it more data-dense but that doesn't change much since the video would just get larger and length of video is not a huge problem
Maybe it would be a nice idea to add some sort of error resistance trough audio, so instead of using it as a whole medium, as it seems it doesn't contribute to it useful anyway, use it to send some sort of checksum? or maybe some other error correction so it get#s more reliable :D
