Psychoakustische Modellierung der menschlichen Hörwahrnehmung

In complex acoustical environments we constantly have to deal with a mixture of sounds from different sources. The normal auditory system is able to subdivide this mixture of sounds into different auditory ob-jects in order to distinguish important sounds from other sounds. The aim of this project is to investigate the underlying object binding mechanisms using psychoacoustics in humans in combination with modelling, taking into account physiological results with stimuli comparable to those used in psychoacoustics. The long-term goal is the development of a model with a realistic internal representation of auditory objects in complex acoustical environments. To this end, the previous funding period focussed on different object binding cues and their relative importance in the formation of auditory objects. Among the important cues are coherent envelope fluctuations across frequency which are a common quality of natural sounds, and the spatial location of the sound source as reflected in binaural information. The underlying mechanisms were, so far, mainly investigated close to threshold and thus models were primarily designed to predict perception at threshold. In the forthcoming funding period, the investigation will be extended towards supra-threshold perception in conditions of release from masking due to object binding cues. This suprathreshold perception will be characterised by means of discrimination thresholds and by asking for sensations. For example, it will be investigated how perception of modulation depth (roughness, fluctuation strength) changes when a signal is added to the masker. The results are an important test for the hypothesis that modulation cues are used for the detection of signals disrupting the envelope coherence across-frequency.
Normal hearing subjects as well as subjects with a hearing loss and cochlear implant users will participate in the experiments to disentangle peripheral (cochlear) from higher processes and to investigate if the cues comodulation and interaural disparities are reduced in their effectiveness for the latter two groups of sub-jects. In addition to psychoacoustics and modelling, EEG measurements will be used to compare the results with the psychoacoustical results and model predictions with a special focus on the perception at supra-threshold levels.