ABSTRACT
To enable the design and characterization of a diffusing surface, it is necessary to be able to predict the reflected pressure from the surface. Currently, this is usually done by considering the scattering from the surface in isolation of other objects and boundaries. The prediction techniques could also be used as part of a whole space prediction algorithm, where all surfaces in a room are simultaneously modelled. At the moment, however, long computation times and storage limitations mean that whole space prediction algorithms are forced to use relatively crude representations of the actual scattering processes. Consequently, when predicting the responses in rooms and semi-enclosed spaces such as street canyons or pavilions, it is more common to use geometric models. The issue of modelling scattering in geometric models is discussed in Chapter 12. Therefore, the issue for this chapter is predicting the scattering from isolated
surfaces. There is a range of choices of models, from the numerically exact but computationally slow, to the approximate but fast models. The prediction methods can also be differentiated as either time or frequency domain models. In diffuser design, frequency domain methods have dominated the development of the modern diffuser. For this reason, this chapter will concentrate on these methods. Table 8.1 summarizes the prediction models which will be considered in this chapter, along with their key characteristics. The next section will start with the most accurate model, a boundary element
method (BEM) based on the Helmholtz-Kirchhoff integral equation. It will then be demonstrated how the more approximate models can be derived from this integral equation, and the relative merits and limitations of the techniques will be discussed. To round off the chapter, an overview of less commonly used techniques will be given.
8.1 Boundary element methods
