Abstract:

One way of improving speech intelligibility in acoustically difficult environments is the application of highly directional loudspeaker arrays in order to increase the ratio of direct to reverberant sound. Separate processing of the signals applied to the individual transducers results in a nearly frequency-independent directional behavior of the array. The directional characteristics (expressed in terms of beamwidth and beam-angle) may be changed on-the-fly by digitally implementing the required transfer functions. Basic directional aspects of one-dimensional controlled arrays will be outlined and general constraints regarding the required hardware will be derived. Simulations will indicate the advantages which might be achieved by the application of digitally-controlled loudspeaker arrays over more traditional methods of speech reinforcement. A hardware setup will be described and the results of some real-life implementations will be discussed and compared to the simulations.