Creation and handling of sounds via a straight-forward, clear and coherent syntax. For example, to create a sound, play it and change its volume:

var track:Track = new Track("path/to/track.mp3");
track.start();
track.volume = .5;

Whereas with the standard API, you would write:

var urlRequest:URLRequest = new URLRequest("path/to/track.mp3");
var sound:Sound = new Sound(urlRequest);
var soundChannel:SoundChannel = sound.play();
var soundTrans:SoundTransform = soundChannel.soundTransform;
soundTrans.volume = .5;
soundChannel.soundTransform = soundTrans;

Hierarchical sound structures. You can put a bunch of sounds into a Group or a Playlist object and control them globally. For example:

var group:Group = new Group( [new Track("path/to/track1.mp3"), new Track("path/to/track2.mp3"), new Track("path/to/track3.mp3")] );
group.pan = -1;
group.addChild(new Track("path/to/track4.mp3"));
group.start(); 

You can also put groups into other groups, and thus create complex hierarchical structures. The 4 main structures (Track, Group, Playlist and Mixer) implements a common interface, so for many methods and properties, you can treat every element the same way.

Each navigation/sound transform methods (start(), stop(), mute(), left/right/center(), etc.) has a "fade" option to make smooth sound transitions (tiny internal tween engine, with a single timer)

Efficient memory management.

M3U/XSPF playlists import/export.

Event-driven objects.

And many more to come: BPM detection, complex ID3 data (covers...), last.fm metadata, cue points...

ASaudio online documentation