Two-channel noise reduction algorithms motivated by models of binaural interaction

In this thesis, signal processing strategies for the reduction of undesired interfering signals in binaurally recorded signals are derived and described. The properties of the different processing strategies are discussed and the processing performance is investigated using artificial signals. Two hearing aid algorithms are described that combine different noise reduction strategies and provide a complete processing of a simulated, digital hearing aid. Furthermore, a method is described and applied that allows for the optimization of 'critical' processing parameters with respect to the subjectively perceived signal quality. Finally, particular audiological properties of the algorithms are investigated and compared, i.e., the influence of the processing on signal quality and on speech intelligibility in noise is measured with hearing impaired subjects.

Level thesis:degree_level

thesis.doctoral

Grantor dc:publisher

Universität Oldenburg

Year

2001

Author dc:creator

• Wittkop, Thomas

• [Keine Schlagwörter von Autor/in vergeben.]

Repository record source_url

http://oops.uni-oldenburg.de/356

OAI identifier oai:identifier

oai:oops.uni-oldenburg.de:356