Victor Zheleznov

The combination of traditional physics-based simulation methods and modern machine learning approaches is an emerging research field in computational physics. This project aims to develop specialised machine learning algorithms for the unique task of physics-based sound synthesis of acoustic musical instruments. By combining a natural acoustic character of a physical model with expressive power of machine learning, the motivation is to create next-generation digital musical instruments. Musicians could obtain access to highly accurate simulations of acoustic musical instruments that can be experienced in real time, with the scope to extend designs to novel natural-sounding digital instruments without a real-world counterpart.