Speech Signal Processing Group

The group for Speech Signal Processing at the Universität Oldenburg was founded in December 2011. Our research focuses on speech enhancement frameworks with one and many microphones and modeling of speech signals for hearing instruments.

In speech enhancement, we aim at deriving statistically optimal methods to increase the intelligibility of speech, which is disturbed by noise and/or reverberation. At the same time, the enhanced speech should be of high quality and exhibit as few processing artifacts as possible. 

In speech modeling, we aim at finding and analyzing compact speech representations. Further, we derive optimal estimators for the parameters of speech models, such as the speech fundamental frequency.