Tagungsband - DAGA 2024
Article
Exploring the Impact of Varying Number of Ambisonics Channels on a Perceptually-Based Audio Reproduction in Reverberant Rooms
* Presenting author
Abstract:
Ambisonics has emerged as a prominent technique for capturing and rendering complex spatial sound fields, providing an immersive reproduction of the original recorded audio in free-field playback scenarios. The theoretical foundation of this method assumes an anechoic playback environment, and using ambisonics in a reverberant playback room can cause perceptible modifications of the recorded sound field, creating a reduction in sound quality. This study deals with this challenge by employing a perceptually-based optimization approach. For this, the direct and reverberant sound field is separately captured, optimized, and rendered on a spherical array of 50 loudspeakers. The direct sound field is rendered using VBAP, while the reverberant sound field is equalized to compensate for the effect of the playback room acoustics, ensuring that spectral energy, direct-to-reverberant ratio, and interaural coherence (IC) of the original recorded audio are reproduced in the reverberant playback room. For rendering of the reverberant part, the effect of using varying numbers of first-order Ambisonics channels (2, 3 or 4 channels) is investigated. A quality-assessment experiment will evaluate the timbral and spatial aspects of the rendered audio for the compensated signals in comparison to conventional Ambisonics.