An angular frequency domain metric for the evaluation of wave field rendering techniques

Spatial sound rendering techniques are based on established theoretical frameworks. The theoretical wave field predicted by the underlying of a specific rendering technique is an approximation of the desired ideal wave field. Moreover the practical implementations require often compromises which cause an additional deviation of the rendered wave field from the desired one. To assess these deviations, metrics are devised which quantify the difference of two wave fields. Typically they rely on the comparison of the wave fields at a discrete set of positions in the area of interest (discrete-space metric). In this paper a novel metric is proposed which is formulated in the angular frequency domain. It expresses the difference of two wave fields in terms of their Fourier series angular coefficients. It is efficient since it avoids the spatial summation over a discrete grid in the area of interest, moreover it is flexible since it considers a weighting function to focus the evaluation on specific regions of spatial and temporal frequencies. After a theoretical treatment such a metric is computed for simulated and measured data and compared with a discrete-space metric. The comparison shows that the new metric yields results which allow a flexible assessment of the characteristics of the rendering techniques.