Broadly speaking, embodiments of the present invention provide a wearable audio device including one or a plurality of microphones, a sound recognition systems and a controller to control the device based on one or more recognized sounds or classes of sound. Embodiments use stored sound models.