Binaural Beamforming Algorithms and Parameter Estimation Methods Exploiting External Microphones

In everyday speech communication situations undesired acoustic sources, such as competing speakers and background noise, frequently lead to a decreased speech intelligibility. Over the last decades, hearing devices have evolved from simple sound amplification devices to more sophisticated devices with complex functionalities such as multi-microphone speech enhancement. Binaural beamforming algorithms are spatial filters that exploit the information captured by multiple microphones on both sides of the head of the listener. Besides reducing the undesired sources, another important objective of a binaural beamforming algorithm is the preservation of the binaural cues of all sound sources to preserve the listener's spatial impression of the acoustic scene.

The aim of this thesis is to develop and evaluate advanced binaural beamforming algorithms and to incorporate one or more external microphones in a binaural hearing device configuration. The first focus is to improve state-of-the-art binaural beamforming algorithms, more in particular to develop a novel algorithm that jointly preserves the binaural cues of a desired source, interfering sources and background noise. The second focus is the incorporation of one or more external microphones to improve the noise reduction and binaural cue preservation performance of binaural beamforming algorithms, without assuming any a-priori knowledge about the position of the external microphones.