Two and Three-Dimensional Sound Source Localization with Beamforming and Several Deconvolution Techniques

Microphone array associated to signal processing technique such as beamforming is an efficient tool to localize acoustic sources. In an open wind-tunnel, the microphone array can be set below, above or on the side of the flow. The microphones are spatially distributed over a two-dimensional (2D) plane and the geometry may be regular, circular or a spiral shape. For each case, the scan zone where the potential source positions are searched is a plane parallel to the microphone array. In this study, beamforming and several microphone array techniques (Clean-SC, L1-GIB, DAMAS and SC-DAMAS) are compared in the case of two and three-dimensional (3D) scan zones. The microphone array techniques are compared in terms of position detection, source level estimation and computation time. The cases of a 2D scan zone, either parallel or perpendicular to the microphone array, are investigated. Then, the performances of each technique are studied in the case of 3D scan zone with a single or two microphone arrays. SC-DAMAS provides the best noise source map at the expense of a large computation time if the number of scan point is large. Clean-SC is a good tradeoff between fast computation and correct source detection. The second microphone array improves the noise source map provided by L1-GIB with a reasonable computation time.