MP3 uses a pyschoacoustic model in its lossy compression so that it doesn't bother to reproduce sounds that can't be heard. This is the same, except for text: it replaces long words with shorter ones, sometimes even the shortest, most expressive 'words' available in Unicode: emojis.

Links

Making-of blog post

GitHub Repo

Intermittently world-accessible live version

Share this project:
×

Updates