You're right, it's Git's. Github did set a good precedent, though, by allowing you to see the diffs of images in a few different ways. That's a really clever feature that gives Git new functionality.

Image diffs make sense, though, because a lot of projects have images (Web sites, resources for UIs, and so on). Audio is probably present in far fewer projects, so I'd expend much less development on it.

it would be a good match if we had a language to describe the music. somebody must be working in that space right?

/me is ignorant of such things