Edge diffraction in computer modeling of room acoustics

Abstract

Listen

Translate article

Computer modeling in room acoustics is typically based on geometrical acoustics techniques. Limitations with such methods include, among other things, a lack of diffraction modeling, which primarily leads to inaccuracies at low frequencies. The inclusion of diffraction modeling is quite straightforward for first-order diffraction, which can be combined with specular reflections of any order. One impractical aspect, however, is that the number of diffraction components can be extremely high, and grows faster (with the reflection order) than the number of specular reflections does. At the same time, the importance, or magnitude, of the diffraction components will differ over an immense range. This variation can be exploited by estimating the importance of each diffraction contribution by the magnitude of its onset, and skipping the remainder of the calculations for those that are deemed too weak. This will be demonstrated for some typical geometrical cases including a set of overhead reflectors, an orchestra pit, and a convex hall shape for which diffraction is less critical. Results indicate that ‘‘diffraction culling’’ can provide a significant reduction in computation time with only small effects on the overall responses for the tested geometries. [This research has been supported by the Research Council of Norway].