Demonstrator D3 relates communication specific user behavior such as head movement, gaze direction, and torso posture, to the acoustic context, delivered by computational auditory scene analysis (CASA) methods.

Based on this comparison, the algorithm can distinguish between relevant and irrelevant sources, and a spatial filter algorithm can be controlled. In this way, it implements the core concept of the CRC for hearing devices: Control of source-selective reproduction of the acoustic environment by a device-internal model of the communication loop based on behavioral and scene analysis.

Core functionality was demonstrated by combining an analysis of gaze direction with a direction-of-arrival estimator to steer the spatial filtering towards an attended sound source. Currently, the focus is on developing environments and tasks to capture ecologically valid movement behavior under laboratory conditions.

Bitte tragen Sie beim Ansehen und Anhören der Videos einen Kopfhörer.