Measurement of surface reflection properties : concepts and uncertainties

Although the quality of room acoustic simulations has increased significantly in recent years, an entirely realistic result is seldom achieved in complex scenarios. Among the factors influencing the degree of realism of such simulations, the boundary conditions concerning sound reflection are considered most important as they determine the sound field to a great extent. Standardized measurement methods exist but they contain inherent uncertainties or do not always yield enough information for a correct modeling of the sound field. To ameliorate the situation, acoustic measurement techniques related to the absorbing as well as the scattering properties of architectural surfaces are investigated in this thesis.The research is divided into two parts: the first part consists of determining the most relevant causes of uncertainty for the standardized measurement methods of random-incidence absorption and scattering coefficients. The difficulties of obtaining accurate results that are often encountered in practice are explained by analytically relating the variation of the input quantities --- such as sample surface area or reverberation time --- to the variation of the absorption and scattering coefficient. Special focus is set on the spatial variation of reverberation times as the primary uncertainty factor. The predicted uncertainty is successfully validated with measurements in both full-scale and small-scale reverberation chambers. Based on the uncertainty analysis, a method is developed to determine the necessary minimum number of source-receiver combinations in the sound field to ensure a specified precision of the absorption or scattering coefficient.The second part of the thesis focuses on signal processing steps related to the measurement of angle-dependent reflection properties in the free-field. For this purpose a hemispherical microphone array is described and validated in this thesis. Improvements to the subtraction method are presented that allow to include the source and receiver directivity. Sound reflection models of different accuracy and calculation complexity are considered to deduce the surface impedance from measured reflection factors. Array processing techniques are investigated as an alternative method to obtain a source reference signal in-situ and to process the spatial response of the reflection measurement.Measurements show that the array setup can be used to obtain the angle-dependent absorbing properties of samples with few source positions. The results indicate that for receivers close to the surface the simplified plane wave model should not be used as it leads to large errors, especially at low frequencies. Some uncertainty remains in the phase angle of the complex reflection factor, which is due to incomplete knowledge of the source and receiver positions. Nonetheless, relatively stable results can be obtained even for samples of finite extent. With the help of array-processing methods, the setup can also be used to determine the directional diffusion and scattering coefficient of small samples, yielding the same result as established far-field methods