Why does an encrypted byte array have a different length from its char[] representation?

Question

I was working on some encryption/decryption algorithms and I noticed that the encrypted byte[] arrays always had a length of 33, and the char[] arrays always had a length of 44. Does anyone know why this is?

(I'm using Rijndael encryption.)

Answer 1

Padding and text encoding. Most encryption algorithms have a block size, and input needs to be padded up to a multiple of that block size. Also, turning binary data into text usually involves the Base64 algorithm, which expands 3 bytes into 4 characters.

Answer 2

That's certainly not true for all encryption algorithms, it must just be a property of the particular one you're using. Without knowing what algorithm it is, I can only guess, but the ratio 33/44 suggests that the algorithm might be compressing each character into 6 bits in the output byte array. That probably means it's making the assumption that no more than 64 distinct characters are used, which is a good assumption for plain text (in fact, that's how base64 decoding works).

But again, without knowing what algorithm you're using, this is all guesswork.

Answer 3

Without knowing the encryption you're using, its a little tough to determine the exact cause. To start, here's an article on How to Calculate the Size of Encrypted Data. It sounds like you might be using a hash of your plaintext, which is why the result is shorter.

Edit: Heres the source for a Rijndael Implementation. It looks like the ciphertext output is initially the same length as the plaintext input, and then they do a base64 on it, which, as the previous poster mentioned, would reduce your final output to 3/4 of the original input.

Answer 4

No, no idea at all, but my first thought would be that your encryption algorithm is built such that it removes 1 bit per 10 from the output data.

Only you can know for sure since we cannot see you code from out here :-)

Answer 5

It would be a pretty lousy encryption algorithm if it was just replacing bytes one-for-one. That was state of the art 50 years ago, and it didn't work very well even then. :)