Acoustic source localization often relies on the free-space/far-field model. Recent work exploiting spatio-temporal sparsity promises to go beyond these scenarios, however, it requires the knowledge of the transfer functions from each possible source location to each microphone. We propose a method for indoor acoustic source localization in which the physical modeling is implicit. By approximating the wave equation with the finite element method (FEM), we naturally get a sparse recovery formulation of the source localization. We demonstrate how exploiting the bandwidth leads to improved performance and surprising results, such as localization of multiple sources with one microphone, or hearing around corners. Numerical simulation results show the feasibility of such schemes.