Topic: A project for steganographers and cryptographers (Read 1659 times)

It's pretty clear from the context that specific information is being accessed--what I'm pointing out is that the method of access is irrelevant.

That seems to be oversimplifying things too much, anyone with a pencil and paper can have "access" to any information, anyone that can go to a beach can have "access" to any information (a turbulent gust of wind blows by and arranges the sand in such a way that the information is easilly recognizable), anyone with a TV receptor can, just untune it and stare at the snow long enough,  anyone with an internet connection has access to for example kiddie porn, just use a search engine to find, and so on and so forth. Just having access to some information can't make you a criminal.

there is no such thing as magic algorithm that can change one file into anything
you cant just find some movie because it simply isn't there
to put information there you would have to use XOR - basically treat block as white card and prepare file with differences between block and video
still, as block data are completely different than any data you may want, you won't have any significant size reduction - you can't just ignore entropy and hoping for the best

if it would be that simple all of us would already have identical dvd with random data and zip files in the internet would be 1KB long
it just don't work that way

I think the algorithm you are looking for is called XOR

abstraction, many intelligent people go through a phase of this type of thinking when they are learning about cryptography, steganography, information theory, and so on, or shortly afterward.  I did so myself around 6 years ago.  The problems with systems such as you describe become readily apparent as you study further, which is why we've had a bit of fun at your expense.  MarkM is satirising one of the fundamental underlying problems with such a scheme, and it is probably the most helpful one for you to understand:

Any two datasets which have no systemic reason to share fundamental similarities in structure and content will be essentially random with regards to each other, and thus attempting to encode one against the other will be no more efficient than an optimised internal compression method of the set in question.

In the case of Bitcoin, this is true at a very deep level due to the intentional randomness of the hash function and encryption in general.  Essentially, the use of any such encoding system would be a "red herring" to the actual data transfer, merely additional work that is performed at both ends, yet yields no benefit for the wasted time and electricity.  Therefore the only reason to do such a thing would be to somehow disguise the data.  The blockchain is a poor candidate for such a use because, as you've pointed out, it's freely available to everyone.  Using the blockchain to disguise data would be like painting the password to your computer on the side of your house in spraypaint.  So, the only remaining reason for a scheme like this would be to provide plausible deniability to the person transferring/holding the data.  "Your honour, I do not possess [copyrighted content/top-secret memos/child pornography].  I only possess a key which is clearly not [copyrighted content/top-secret memos/child pornography] and this publicly available blockchain."  Arguments like this make sense to geeks like us because they point out the inherent freedom of form that information offers.  In real life courts, however, nobody cares.  You have some way to access the content in question, and whether it involves a satanic ritual or pressing a button, that's all a judge will care about--my point 1).  The statement in the preceding paragraph that you should understand is essentially my point 2), that no matter how cool your encoding scheme it won't ever yield benefits for key size.  Point 3) is just a nod to the fact that projects like this even as described would take up far more labour than their proposers ever imagine, for negligible benefits (even if the simplistic problems are addressed it turns out you end up having to solve intractable problems first, like defining consciousness and constructing a human-equivalent AI).  And the p.s. is just a final reminder that all information has to be somewhere.  If the keys for millions of web videos are smaller than the videos, there's certainly not millions of videos' worth of patterns in the blockchain.  And if you've discovered some way to reduce video sizes generally, you should just use that knowledge to build a better codec.


Chin up, though.  Information theory, encryption, and steganography are exciting.  They just hold their exciting discoveries wrapped a little bit deeper in the folds of mystery than this--you'll find them if you stick at it!  After all, someone invented Bitcoin.  Maybe you'll come up with the next big thing--there's a lot of untapped potential in these areas.


sincerely,
eMansipater

So the target audience is to have the same stale old videos that you do eh?

To fit all this into the existing already stored block chain I re-propose my long sought "decode anything into anything" application.

Want porn? No problem, it can be decoded as porn. Would you like a and A to stand for that and *that*, or does decoding them as this and *this* look better to you?

Please specify pixel by pixel the video you wish to decode from this blockchain and we will try to find an algorithm that does indeed churn out that set of pixels when used to "decrypt" that same old blockchain...

And for the censors, hey, this app can make anything as obscene or even kiddieporn as your eyes wish to perceive it as! Please specify a pixel to change and how to change it that will to your eyes seem more prurient...

-MarkM-

Alright, everybody--on three!

1)Obfuscation != indemnification
2)Video-sized public keys = no bandwidth savings and terrible codec performance
3)"Build a web browser"

Need I say more?

No hate from here, I've had many a nerdgasm on cryptography/steganography before.  Just thought I'd save everyone the trouble of trying to explain why this won't work.

p.s. the blockchain is around 150mb right now

Use the Bitcoin block chain for steganography. All people who use the Bitcoin client have the whole block chain downloaded, so it's safe to assume the target audience has the same information you do. Then, figure out an algorithm that can "decode" the block chain into a file based on a public key. Next, build a web browser that simply manages key exchanges. When you visit a web site, it would only transfer a key to your browser. The browser would "unpack" the website using the key. On the page displayed, there would be more keys - keys which would decode into text, script, images, ...media. Stale video would not need to be streamed anymore, just need the key to it and decode it locally. "Live" video would still need to stream the keys at some rate.

I like the idea of having an encyclopedia and its changelog presented as a set of keys, instant video on demand without hogging bandwidth, and a tip if you like this idea... 17dBSzDq3SXJzoDLx3PDnqJ2MRqCecYPY4