Adaptive Mixing of Excessively Directive and Robust Beamformers for Reproduction of Spatial Sound

The spatial sound reproduction methods that utilize a parameter‐based model describing some properties of the sound field reproduce sound scenes with a high perceptual accuracy when compared with non‐parametric techniques. In the method, two types of beamforming signals are generated: a set of narrow and potentially noisy beamformers that approximate with high accuracy the target reproduction setup, and a set of more noise‐robust beamformers that spatially approximate the same setup only loosely. This chapter presents an alternative parametrization of the sound field that circumvents the modeling of the sound field with parameters such as direction of arrival (DOA) and deals with the aforementioned issues related to the perceived quality of the individual channels while retaining the high overall perceived quality of reproduction. It provides an overview of the adaptive time‐frequency processing. The chapter also provides details of the method and listening test results for loudspeaker and headphone implementations.