Experiments in translating a bytestream into audio and back, play it over speakers, record it with a microphone and decode it to bytes again.

Experiments where made using Paul Austers voice reading a book aloud.

255 samples of his reading (2 second duration per sample) where chosen and a script takes a bytestream piped into it (for example a bitmap) and starts speaking the bytestream aloud using the Paul Auster samples as a representation of the bytes

Another script listens to the sound, looks for audio fingerprints and looks them up to get the associated byte value. It then outputs a bytestream to stdout and another script reconstructs in this case a bitmap from the bytestream.

The transmission of data is lossy but the fun of listening to it live with Paul Austers voice is making up for it.

Watch the video here.

Get the code here