Acoustic properties of environments are difficult to model, due to their dependence not only on geometry, but also an the internal properties of materials (density, etc.). An acoustically accurate model that can be quantitatively validated is extremely useful for the evaluation of improved enclosures, silencers, acoustically absorbent materials, and human factors studies. We present a new parallel algorithm that generates full 3D simulations of speech intelligibility and sound pressure levels which include sound diffraction, sophisticated external noise and surface absorption models. The simulation environment interface is decoupled from the algorithm using a C++ framework running under X11R5 and Motif. The sequential version of this algorithm produced sound pressure levels that agreed with measured values on the STS-40 Space Shuttle mission to within 4%. We also present the results of some experimental studies performed using ParaSoft Express on the 56-node Intel Paragon at the University of South Carolina.
展开▼
