Machine learning approach to design acoustic experiences

Machine learning in acoustics is an emerging field with great potential yet to be explored. Previous research includes predicting acoustic properties, such as the impulse response (sound signature) from visual features that include an image of a room. However, one application that can significantly transform the acoustic environment design process is reconstructing the room geometry based on its acoustic properties. This project explores applying machine learning in predicting space geometry from an impulse response and aims to provide a tool for designers to design a space based on the desired acoustic experience.

This project is in collaboration with Nikhil Singh—a PhD student at the Opera of the Future group.