This paper describes the results of current research at DREA in which techniques of optimum array processing are being applied to active sonar. We are presenting these results at the Advanced Study Institute in order to illustrate some actual applications for such processing and to point out some of the practical considerations which arise in real systems. In particular, the paper concerns the problems which arise when the individual sensor elements have a complicated directivity pattern themselves. This is a common phenomenon in active systems where the receiving sensors are complex resonant structures and are housed in a dome or towed body presenting various baffling and diffraction effects. Most treatments of array processing consider ideal elements which have well behaved directivity properties and are transparent to the field. The results of this paper show that where these properties are not met, careful in situ array measurements are required, and even with such measurements practical array gains may not be as good as predictions based on ideal sensors.