Modelling the performance of speaker arrays in domestic listening environments.

Abstract

The evaluation of multi-speaker arrays is often completed in ideal anechoic conditions to better enable the observation of speaker interaction at a known listening position. Real-room testing of speaker arrays is often time-consuming and has practical limitations, which can have an impact on analysis and also the repeatability of testing. This research will investigate the performance of multiple simulated speaker arrays such as 5.1, WFS and Ambisonics in modelled domestic environments. This will be implemented using acoustic modelling software and measured speaker directivity which will allow for subjective analysis using binaural auralisations; where modelled living room impulse responses will be generated for each speaker type and position to give several ‘virtual loudspeaker arrays’. Objective evaluation of each system's spatial reproduction and localisation ability will also be considered using techniques such as velocity and energy vector analysis as examples.