Wwise does not provide any mechanism to interleave streams in order to reduce seeking time on disk, so you may quickly run out of I/O bandwidth if you use several streamed tracks in the same segments.
As a rule of thumb, you should render as much audio content in the same music track as you can. Create separate tracks for specific instruments only when you need to change their properties at run-time, with States or RTPC.
You may also wish to reduce bandwidth by sending silent tracks to virtual voices. Beware, however, that when a streaming voice changes from virtual to physical, it has to wait some time before playback actually starts. This wait time is defined by the track's "look-ahead time". Playback will be synchronized with the position in the segment (and other tracks), but volume change will occur later than expected. Thus, it doesn't work well with delayed state changes (refer to http://kb.gowwise.com/questions/177/Using+States+with+the+Interactive+Music+Hierarchy for details about delayed state changes).
Streaming multiple tracks simultaneously and changing their individual volume/presence is the classic way of making music interactive. We call it the "layering" approach. On the other hand, Wwise favors another approach to interactive music, which consists of connecting segments together at run-time, based on transition rules. We call it the "connecting" approach.
If you are concerned about I/O, you may use "Sync To Same Time as Playing Segment" transitions in a connecting design in order to emulate classic, layering designs. If you have a track that is heard in State A, but not in State B, render the music twice, with and without this track. Then, assign them to separate segments, put these segments in a switch container, and set the transition rule between them to synchronize to "Same Time".
Of course, this technique may not scale well to a great number of interactive tracks, as you will end up with a lot of duplicated music structures. Note that the media data is not duplicated unless it is explicitly included in several soundbanks.
