Audioman

New features
- `Audio.save()` now has a `ffmpeg_options` parameter. This can be used for custom output, as well as adding effects that are not supported by audioman.
- `Audio.unload()` now saves the current audio in a temporary file, which will be loaded when accessing the `.samples` property.

Bug fixes
- There was an error when putting a `dict` or a `list` in the `.setdefault()` method on an `AudioTags` object.

First release. I copied the code from my `python-soundtrack-creator`, as well as edited a few things. I split this into a separate module and am uploading it to pypi so other people can use it without all the soundtrack stuff.