"Les Mercredis de STMS" welcomes Professor Lauri SAVIOJA for a presentation entitled "Machine Learning in Room Acoustic Modeling".
Machine learning is revolutionizing research in numerous fields. This presentation reviews how machine learning has been utilized in room acoustic modeling ranging from modeling diffraction from a single edge to prediction of impulse responses in formerly unseen room geometries. In addition to just geometry data, some models also use photos for audiovisual learning. The range of applied models is very wide going from small two-layer neural networks to deep learning with GANs (generative adversarial networks) and transformer models. The achieved results are impressive, and it is easy to envision a future in which these models are used in AR/VR applications although there is still a long way to reach the level that would be needed for more serious use of room acoustic modeling.
Lauri Savioja works as a professor of virtual reality technology at Aalto University, Finland. He used to serve the university as the head of the Department of Computer Science, but at the moment he is having a sabbatical at IRCAM. His research has focused on room acoustic modeling since early 90’s although in more general terms his research interests include room acoustics and virtual reality. Prof. Savioja is a fellow of the Acoustical Society of America (ASA) and of the Audio Engineering Society (AES). From 2018 he has been an Associate Editor for the Journal of the Acoustical Society of America.