Author (s)
Zotter Franz 1, Frank Matthias 1, Schörkhuber Christian 2, Höldrich Robert 1
Affiliation
1 Institute of Electronic Music and Acoustics, University of Music and Perfrming Arts Graz, Austria
2 sonible GmbHm, Graz, Austria
Publication date
2020
Abstract
Six-Degrees-of-Freedom audio recording and rendering approaches have been recently proposed to enable a variable-perspective playback for a listener. There are works pre-senting spatially sampled BRIR sets [1, 2] to prepare forthe variable-position and variable-orientation binaural rendering based on either linearly interpolated binaural (dummy-head) signals [3, 4] or parametrically interpolated ones [5]. Moreover, some works present spatially distributed measurements of perspective room impulse responses [6] and we find works about projecting directionally localized sound objects in single-perspective recordings onto an outer convex hull of the room [7, 8, 9, 10], and works and patents about the interpolation from perspective recordings synchronously taken at multiple perspectives in the room, with parametric concepts to extractand render the sources detected therein and the diffuseor unlocalized parts [11, 12, 13, 14, 15, 16, 17, 18]. Some of the works avoid or at least partly avoid any short-termtime-frequency-filtering based processing to get artifact-free baseline rendering [19, 20, 21, 22, 23, 16, 24], which, however, may stay limited in spatial precision. Anyway, any kind of broadband baseline rendering methodis valuable to either conceal annoying time-frequency-based processing artifacts or just as a standalone solutionof decent audio quality. This contribution presents a simple and signal-independent strategy that was outlined in [19, 20] and tested in [21], but has never been written up in its most simplistic form that proved useful in practical demonstrations. It is based on recordings with distributed Oktava MK-4012 4D A-Format microphone arrays. Our demos typically used 12. . . 16 tetrahedralmicrophones covering a walkable area of 25. 200 m2. In the virtual space, for each array, the 4 array signals are routed to 4 virtual loudspeaker objects (VLOs), and an additional objects to get a good extrapolation when the listener is located between wall and the area covered with the distributed arrays. For this particular setup, the processing steps of the rendering method are outlined in greater detail below.
Full paper
https://pub.dega-akustik.de/DAGA_2020/data/articles/000458.pdf
Keywords
6DoF audio rendering, binaural room impulse responses (BRIRs), variable-perspective playback, distributed microphone arrays, signal-independent rendering, spatial audio interpolation, virtual loudspeaker objects (VLOs)