Abstract
Object-based audio systems have been getting more attention in the recent years. The object-based approach redefines the audio production and reproduction chain. Instead of optimizing the individual signals for every loudspeaker of a specific audio system (channel based approach), the object-based production delivers a scene description that contains a set of sound source signals along with some metadata. This scene description is processed in the renderer together with a set of input parameters, which are related to the local installation, to create the corresponding optimized output signal for each loudspeaker. As a result, any object-based production can be adapted to any listening room by just modifying an input parameter. This paper describes the scene adaption process (upscaling) carried out in two different rooms from a technical and artistic point of view. The research starts with an object-based production mixed in a studio with 20 loudspeakers that are distributed into 3 layers. Then, the same scene is adapted and reproduced in a dome setup with 48 loudspeakers arranged into 5 layers. Finally, both scenarios are compared in terms of spatial impression, localization accuracy of movements in vertical and horizontal planes, sweet spot dimension, and overall audio quality.