Green’s function retrieval by crosscorrelation and multidimensional deconvolution: Application to atmospheric acoustics

In this work, the Green’s function of an atmospheric acoustic propagation channel is extracted from experimental measurements. Of particular interest, is the accuracy of Green’s function retrieval methods in ideal and non-ideal situations. To this end, an emitting acoustic source (impulsive and audible) is placed at a distance of more than 90 meters from receiving tri-axis microphone arrays in open (ideal) and wooded areas (non-ideal) in southern Maryland. Green’s function retrieval methods are employed and investigated on the collected data. Green’s function retrieval by crosscorrelation has been successful in various applications despite the limitations of its lossless medium and equipartitioned wavefield assumptions. To overcome the violation of these assumptions, the multidimensional deconvolution has been proposed. Comparisons of the results between impulsive and audible sources using the two methods will be presented. The effects of multipath, source-receiver distance, temperature, and wind will be discussed.