Added sound model implementor UUID.
Added capture preamble duration in recognition event.
Replaced number of users by list of user IDs in key phrase
descriptor.
Report confidence level for each user with user ID in recognition
event.
Added more configuration data to start_recognition() grouped in a
recognition configuration structure.
- List of keyphrases to listen to
- Recognition mode
- List of users with min confidence level
Added definitions for sound model load callback event.
Bug: 12378680.
Change-Id: Ia8c12ab20448b702d13db964bdd90ac4117a58d7
135dd2ee3d