My work is concerned with the microphone array-based separation and localization of sound sources, particularly speech. I also look into ways to combine both by exploiting the knowledge of the source locations (in the form of prior information) within the separation. To perform direction of arrival estimation, I have considered deep learning and classical model-based methods. To then approach the problem of separating multiple speakers using time-frequency masking as well as beamforming, I have extended my previous work on localization, but also considered spatial clustering as an alternative method. Among my current topics of interest are the development of better ways to integrate prior knowledge within the (neural network-based) source separation, addressing the special case where the sources are close in space, and improving the synergy between beamformer and directly applied time-frequency masks in a combined system.
Outside of work, I play handball and go running, but I also watch sports including football, handball, and F1 on TV (or live when I get the chance). Other things I enjoy are food & drink, playing poker with friends, quizzing, and seeing the world. I’m originally from Germany, although I’ve actually spent more than half of my life abroad.
Academic background: Since 2018: PhD student at Ghent University 2016 – 2018: Master in electrical engineering at RWTH Aachen University (specialization: information and communications technology) 2013 – 2016: Bachelor in electrical engineering at RWTH Aachen University