Back to list

Prediction of the listening area for surround playback systems based on the energy vector model

Authors Kurz, E.
Year 2017
Thesis Type Audio Engineering project
Topic Psychoacoustics
Abstract For some time, Higher-Order Ambisonics is a research topic in reproducing three-dimensional sound fields. Among others objectives, research aims at sound field reproduction over a large listening area. For prediction of a sound field the so-called energy vector (rE-vector) is a commonly used tool. At the central listening position, the energy vector provides good results for predicting position and width of a phantom source. However, at off-center listening positions, the performance of the energy vector becomes worse. This project work compares different extensions of the energy vector model. The extensions try to include time and level differences, as well as the precedence effect in dependence of the listening position. Existing data from listening experiments is used to evaluate the extensions and to optimize some parameters. Simulations investigate the loudspeaker array of the IEM CUBE for Ambisonic playback of 1st to 5th order. Extended energy vectors for discrete points within the whole listening area are computed. The resulting localization accuracy of the extended energy vectors at off-center listening positions can be compared to further predict areas with sufficient localization accuracy.
Supervisors Frank, M.