Score:0

hiding bits inside random bits, is this some sort of encryption?

jp flag

Let's say that I uniformly sample a 1000 bits randomly. Then, I hide a secret of 10 bits on those 1000 bits, randomly. That is, I pick each of the 10 bits and randomly replace them inside the 1000 bits, in an order that just I know. And then I call this order my key. Knowing where each bit is, gives me the hability to recover it.

How easy it is to guess those bits from the random ones surrounding them?

Eugene Styer avatar
dz flag
Hiding one message inside another is called Steganography (https://en.wikipedia.org/wiki/Steganography)
cn flag
What are you trying to achieve? A well hidden key doesn't do anything. If you use that key in AES, the receiver still needs to have the full key to decrypt anything (which could be you in case this is used for encrypted storage). And it doesn't matter if a few bits are pre determined or fully random. Worst case: you weaken the system so much, that an attacker has an easier time for a brute force attack.
kelalaka avatar
in flag
Are you trying to achieve [plausible deniability](https://en.wikipedia.org/wiki/Plausible_deniability)? If not, use dicewire like passwords instead of storing the locations then use Password-Based Key Derivation functions like Scrypt, Argon2.
Score:2
in flag

If you just hide the information, hoping no one knows it's even there it's steganography not cryptography. However if you call the selection of bits(ordered?) a key this can be evaluated as cryptosystem as well.

It's a problematic cryptosystem, as the key is larger than the plain text and the ciphertext much larger.

Security wise if the key is only used once it would be very difficult to find the plain text. However we could still learn some stuff about it, e.g if the ciphertext has more zeros in it, it is slightly more likely so does the plain text.

If we tried using the same key repeatedly for multiple plaintexts, especially if they are known we will quickly be able to figure out the key. Each ciphertext pair eliminates many possible keys. In fact each known plaintext bit eliminates half the possible keys, And though the key space is large we can quickly figure them out one with a bunch of known plain texts.

If you want to use the key only once,.this is clearly inferior to a one time pad. The key is larger and the ciphertext much larger.

Score:-1
ng flag

While what you talk about is most obviously related to Steganography, it is also somewhat related to Big-Key Cryptography. The vague idea here is to protect against key exfiltration by having a ridiculously large key. There is a constraint that on each message you encrypt, you only use a small number of bits of the key (for efficiency reasons). But by the key itself being very large, it is harder for an adversary to exfiltrate it --- it is easier to notice 10GB of network traffic than 256 bytes.

Meir Maor avatar
in flag
In the proposed scheme hiding K bits in a ciphertext of size N, the ciphetext is the big part not the key. The Key would be $k*log(n)$ while the cipher text is $n$ while we assume k<<n.
mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.