Say a song is recorded somewhere on the planet with a computer, and that stream of data is stored.
If this is stored using Audacity, you can set it to .aup format, which is a very simple XML file.
Why do we need anything else than a few data points in a file? Why would we use .mp3 or anything? In fact, encoding this increases the size for a few local samples.
I think the issue could be size for images for example, but I can't say that for audio files.
.aupfiles aren’t self-contained, they refer to other files which contain the actual audio.