Autogenous Spatialization for Arbitrary Loudspeaker Setups

Autogenous spatialization is an experimental approach for distributing audio signals to loudspeaker systems, based on the signals’ properties and the configuration of the loudspeaker setup. Time-domain and frequency-domain features of incoming audio streams are extracted and serve as control signals. The audio streams are then distributed to individual loudspeakers, based on a mapping between these control signals and properties of the loudspeakers, such as position and orientation. Results of the approach can range from a localizable panning between loudspeakers to alienating modulation effects and spatial sound synthesis applications. The proposed system is implemented as a C++ toolbox to allow the integration into various ecosystems for music production and sound design. External objects are provided for Pure Data, which allows the flexible testing and exploration of the spatialization approach.