Master Theses - Details

Real-Time Own Voice Estimation for Hearables and Headphones utilizing Multiple Microphone Signals

Supervisor:Christoph Weyer, Stefan Liebich

Area: User Interface Design, Machine Learning

Tools: Matlab, Python, Estimation Theory, Machine Learning

Categories: Bachelor Thesis, Master Thesis

Status: Open


In recent years, wireless bluetooth earbuds finally arrived in market for consumer electronics. These so called hearables have the potential to improve the hearing and listening experience of the wearer. The possibilities include selective listening, augmented reality, active noise cancellation and many more. However, these features require a multitude of different algorithms to be employed in real time in a computation and power constrained environment, thus posing challenging signal processing problems.

For multiple applications of hearables and headphones, an estimate of the user's own voice is of great importance. These include improved sound quality during phone calls, better speech recognition, but also a more natural perception of one's own voice while using the hearable, utilizing active occlusion cancelation.

Modern consumer headphones, as well as hearables, are often equipped with multiple microphones. The question arises: How can we use these microphones in conjunction to get a good estimate of the user's own voice? Existing speech enhancement techniques utilizing multiple microphones could be used to solve the problem. These techniques include, among others, beamforming, as well as multi-microphone noise reduction.

As part of this thesis, different speech enhancement approaches shall be researched and analyzed for their suitability to solve the own voice estimation problem using multiple microphones. One or multiple promising approaches shall be implemented in a real-time capable fashion. The implementation should reflect the importance of computational constraints of the hearable platform. Their performance shall be evaluated based on real world example signals.