Each month, the lab proposes invited talks for its partners.

This month, we will have the pleasure to welcome Fabian-Robert Stöter, head of research at Audioshake, who will give a presentation entitled:

“Music Source Separation: Is it solved yet?”


In this talk, we present the research task of music source separation, a cornerstone problem for many applications in the audio industry. We review methods that were proposed in the past years which lead to current state-of-the-art deep-learning-based systems. In this talk, we then raise the question of whether the task has been sufficiently solved. To guide future research, we present the pros and cons concerning applications in academia and industry. The particularity of this talk will be to approach the topic from both, a theoretical perspective, and an interactive demonstration regarding how to implement and assess ideas in practice using existing open-source ecosystems for datasets, models and metrics/evaluation. The talk will conclude with an introduction to the Sound Demixing Challenge 2023 and guidelines on how to participate in that challenge.


Since 2021, Fabian-Robert Stöter is the head of research at Audioshake where he is working on machine learning models for music processing. Before, he was a postdoctoral researcher at the Inria and University of Montpellier France. He received his PhD (Dr.-Ing.) at the International Audio Laboratories Erlangen (a joint institution of Fraunhofer IIS and FAU Erlangen-Nürnberg in Germany supervised by Bernd Edler titled « Separation and Count Estimation for Audio Sources Overlapping in Time and Frequency ». Before, he graduated in electrical engineering/communication engineering from the University of Hannover, Germany.

His research interests include supervised and unsupervised methods for music source separation and signal analysis of highly overlapped sources and sound source counting.

More on the speaker’s website.