2

Suppose I have a rather small file, containing.. a photograph of a... llama. And I am very paranoid of people copying my llama photo, so I decide to store it in something inconspicuous like a cassette tape. I trust people wouldn't be able to open my file because they would assume the cassette has music in it and if they try to listen to it, it will be some noise.

How would I be able to do this without compromising data integrity? Can I simply dump some binary data into the tape? Should I care about the speed at which I read the tape later on?

And finally, what tools in the real world would be necessary for such a project? Are there concise applications, frameworks or other tools for cassette recording that I can use, for example, in a Unix environment? Do I need to code my own tool?

Thank you!

Apollo
  • 141
  • 4
  • 5
    This is most certainly not an easy task. Please see [this](https://www.instructables.com/id/Images-On-An-Audio-Cassette/) link for further reading. – NiallUK Jun 17 '19 at 14:19
  • Is the important bit here the cassette tape or that it is innocuous? Because many PCs have DVD burners you can do this a lot more secretly, securely and reliably using a CD, and possibly steganography and extra redundancy. Also, what OS. This is probably easier to customise in Linux by chaining existing tools. (You might struggle to get a jpg photo onto a cassette with reliable playback without non-trivial custom.coding) – davidgo Jun 17 '19 at 19:31
  • One audio encoding is called "Kansas City Standard" https://en.wikipedia.org/wiki/Kansas_City_standard?wprov=sfla1 – Hans Hubert Vogts Jun 17 '19 at 19:39
  • @davidgo The question is about cassette tapes specifically. I'm curious about how that problem could be solved. – Apollo Jun 24 '19 at 15:17

1 Answers1

2

Actually you can. This was a very normal way to store data for a few years in the 1980's. What you do is translate the binary image of the file into a sequence of higher and lower sounds, which you record onto the audio tape. You can read them back later and translate the sequence of frequencies into a sequence of bits.

What actually works quite well (and is much easier to implement) is to translate the nibbles of data into DTFM sounds as used on a telephone network, then read them back via DTFM recognition.

Eugen Rieck
  • 19,950
  • 5
  • 51
  • 46
  • 1
    Yup, this is how I saved stuff for my TRS-80. Was all excited when 45 minute tapes came out, but they were too stretchy for data. And note - NEVER forget which side has Led Zeppelin and which side has your program. Led Zepplin won't run under a BASIC interpreter, and programs sound bad and tend to ruin speakers – ivanivan Jun 17 '19 at 14:59