This investigation was based on the development of the handheld Lexis array microphone. From that study, the routines for optimising the array processing were readily available.
The two main challenges were:
1. To replace the simple free-field model of sound propagation with a model that included the head (modelled as a rigid sphere) and which allowed for a sound source on the sphere, in order to model the target sound source (the user’s mouth).
2. To replace the standard Directivity Index and the noise sensitivity metric with measures relevant for the headset case.