Auscultation sounds offer a rich source of diagnostic information that could potentially be used to detect a vast number of heart and lung pathologies non-invasively. Recent times have seen efforts directed towards the possibility of using an array of acoustic sensors for the localization of sounds, thereby leading to improved diagnosis. However, a majority of the existing array processing and inverse-source solution schemes proposed to date rely on simplifying assumptions, and many of the factors that contribute to violation of the assumptions, such as heterogeneity of the thoracic cavity, shear wave contributions, extreme near-field conditions, to name a few, are not taken into account. This work presents a test bed capable of simulating the sound distribution around the thorax. Specifically, a finite-difference time-domain (FDTD) forward model of the human thorax that solves the viscoelastic wave equations while taking the intrinsic anatomy and the associated viscoelastic properties of the various tissues and structures into account is presented. The test bed can be employed to develop and validate various array signal processing algorithms and inverse problem techniques used for localizing sound. Test results demonstrating the utility as well as effectiveness of the test bed are presented.