Contribution
Comparative Evaluation of Room Compensation Approaches in Ambisonics Reproduction
* Presenting author
Abstract:
Ambisonics is a widely used technique for capturing and rendering spatial sound fields, delivering immersive audio experiences in free-field playback. However, its effectiveness diminishes in reverberant rooms, causing noticeable sound quality degradation. To address this, two types of approaches have been proposed. The first use multichannel room equalization methods in the spherical harmonics domain, employing inverse filtering of the re-encoding matrix. While effective, these methods face challenges related to robustness, causal filter realization, and the inherent limitations of pure Ambisonics recording and rendering. The second approach is perceptually based, in which the direct and reverberant sound fields are separated from an Ambisonics recording and then rendered after perceptually-based filtering. The direct sound is usually handled using the Vector Base Amplitude Panning (VBAP) method, while the reverberant components can be rendered by Ambisonics or other approaches. In the second approach, the original structure of rendering using conventional Ambisonics is altered. In this study, both approaches are compared against a reference dummy head recording using a MUSHRA-like test, evaluating their performance in compensating for the reverberation of the reproduction room.