C4DM Seminar: Alberto Bernardini: Differentiable Physical Models of Acoustic Systems for Audio Signal Processing
QMUL, School of Electronic Engineering and Computer Science
EECS Seminar & Centre for Digital Music Seminar Series
Seminar by: Alberto Bernardini
Date/time: Thursday, 29th May 2025, 11am
Location: GC204, Graduate Centre Building, Mile End Campus, Queen Mary University of London, E1 4NS
Zoom: https://qmul-ac-uk.zoom.us/j/9798452959
Title: Differentiable Physical Models of Acoustic Systems for Audio Signal Processing
Abstract: Accurate simulations of acoustic systems—such as analog audio circuits, transducers, and acoustic reverberant environments—are often computationally demanding, which limits their use in real-time or online audio signal processing contexts. To overcome this challenge, there is a growing need for physically motivated models that are both computationally efficient and sufficiently accurate to be embedded in audio processing pipelines. In recent years, automatic differentiation techniques have opened new possibilities by enabling the direct optimization of physical model parameters using measured data. This development has given rise to hybrid modeling approaches that combine the interpretability and robustness of physics-based models with the adaptability and accuracy of data-driven methods. In this seminar, we will explore the emerging field of differentiable physical modeling for audio applications. We will present concrete examples including models of analog audio circuits for Virtual Analog Modeling, physical models of acoustic transducers such as loudspeakers and microphones, and artificial reverberation systems based on Delay Networks. These case studies will demonstrate how differentiable models can be leveraged to bridge simulation and machine learning in the design of modern audio processing systems.
Bio: Alberto Bernardini received the B.S. degree in computer engineering from the University of Bologna, Bologna, Italy, in 2012, and the M.S. degree (cum laude) in computer engineering and the Ph.D. degree (cum laude) in information engineering from the Politecnico di Milano, Milan, Italy, in 2015 and 2019, respectively. From 2019 to 2021, he was a Postdoctoral Researcher with the Politecnico di Milano. He is currently an Assistant Professor with the Dipartimento di Elettronica, Informazione e Bioingegneria (DEIB), Politecnico di Milano. He authored more than 80 publications in international journals and proceedings of international conferences. He is coauthor of three international patents. His research interests mainly include audio signal processing, computational acoustics, and modeling of nonlinear systems. He was a recipient of the Dimitris N. Chorafas Award in 2019. He is an IEEE Senior Member. He is an Associate Member of the Digital Signal Processing Technical Committee and a Regular Member of the Nonlinear Circuits and Systems Technical Committee, both of the IEEE Circuits and Systems Society. He is a Member of the EURASIP Signal and Data Analytics for Machine Learning Technical Area Committee. He was an Associate Editor for the IEEE Transactions on Circuits and Systems I: Regular Papers. He is now an Associate Editor for the IEEE/ACM Transactions on Audio, Speech, and Language Processing, the Elsevier Digital Signal Processing journal, and the EURASIP Journal on Audio, Speech, and Music Processing.