Issuse encrypting a gzip compressed raw text: value change to base64 - encryption

Question

I have the following raw text :
.
It represents a gz compressed value. Here is the raw text uncompressed :
.
Now I would like to encrypt it using AES in CBC mode with a 128 bits long key. Here are the key and the inizialization vector I'm using:
key: asdfgh0123456789
iv: 0123456789asdfgh
Anyway when I try to decrypt the cypertext obtained from encryption I got base64 encoded raw text as my input.
[Here][1] is the website service I'm using to encrypt and decrypt.
Why does my input change automatically to base64 ? Whats wrong ?
Screenshot :
Screenshot: [![enter image description here][2]][2]

Answer 1

The problem with a sequence of bytes is that they cannot simply be shown on a screen because you can have 256 different values on a single byte but the alphabet only has 26 different letters.
Given a sequence of bytes it can be represented in letters and numbers when converted to base64.
1) The text TEST is represented in bytes as 54 45 53 54 and as base64 as VEVTVA==
2) GZ of TEST is represented in bytes as 1f 8b 08 00 00 00 00 00 00 ff 0b 71 0d 0e e1 e2 02 00 cf 1b f9 1e 06 00 00 00 and as base64 as H4sIAAAAAAAA/wtxDQ7h4gIAzxv5HgYAAAA=
Now you can encrypt the bytes of the gz or the base64 of the gz - they are different.
When using a web page that takes text as input you better use the base64 of the gz. Then, when you decrypt you get what you used.

Answer 2

First and foremost you have to understand that you are not encrypting a raw text gzip. You are actually encrypting a base64 form of your raw text gizp.
Then symmetric encryption algorithms like AES are useful because given a cleartext and key they allows ciphering and deciphering to transform respectively your in input first in ciphertext and then cleartext.
As shown in your screenshot you correctly achieved that! So really nothing wrong. To return to original form of raw gzip text you just have to decode the output of decryption using the base64 encoding scheme and you'll get wath you are looking for.
Moreover as you already knows not all bytes values can be represented by visible symbols, this is the main reason often encrypted text or binary data is represented using base64. Over other enconding scheme, it has less overhead. For instance hex encoding double the input size; while instead base64 makes it (on average) 33% bigger.
As a bonus: Here is useful tools for this kind of transformations Cyberchef.