Directional filtering of recordings in spherical harmonic domain for auralization of simulated noise barriers

Abstract

The effects of traffic noise on humans can cause significant health problems and can impair the quality of everyday life. Current research and legal requirements mainly refer to physical sound pressure measurements of noise. However, the human perception of noise within its acoustic and visual context is not fully considered. With the development of growing population that is exposed to noise caused by roads, rails and planes, the need of an evaluation method that includes the perception of the noise within its environment gains in importance. At the same time a financially efficient tool is needed to reduce costs of construction projects for noise protection measures. Therefore, an evaluation tool should be evolved, that makes it possible to experience and rate noise protection measures based on a psychoacoustic model in virtual reality. Supported by findings from the virtual reality model, suitable noise protection measures can be built in real-life subsequently. The aim of this thesis is to process spatial audio recordings in spherical harmonics domain in such a way that the resulting signal contains the spatial sound field of the recording including the effects of a synthesized noise barrier. Through this, the auralization of varying noise barriers can be realized for real life scenarios in a virtual environment. The processing is implemented by applying filter functions to directed sectors of a plane wave decomposed form of the signal. The filter functions approximate the transfer functions of simple single screen noise barriers, based on a secondary source diffraction calculation model. With this methodology it is possible to establish directional frequency filtering, which only affects components of the sound field that approach the receiver from certain directions. The model is used for filtering complex as well as synthetic generated sound fields and is subsequently analyzed thoroughly. Limitations of the processing are discussed and improvements are suggested.