Audio and Video became first-class citizens on the Web with HTML5 the same way that other media types like images did in the past. Through their new APIs you can access, control and manipulate timeline data and network states of the files. With the coming additions to the APIs you will be able to read and write raw data to audio files (Audio Data API) or manipulate captions in videos (Timed Track API). But the real power of these new HTML elements stands out when they are combined with the other technologies of the web stack, be it Canvas, SVG, CSS or even WebGL.
For instance, you can build a DJ mixer table app where you show several controls to fade and mix audio tracks. Canvas lets you read pixel data from videos so you could add a video visualizer with a scene selection feature and, on top of that, use SVG filters á la Photoshop to distort the movie while the videos are playing.