Room acoustic behaviour estimation with machine learning

Abstract

In today’s world, machine learning and neural networks have taken a firm place in the everyday life of the average person. Moreover, deep learning networks are now an essential and integral tool in every scientific discipline for solving multidimensional and diverse problems. Inevitably, therefore, the field of room acoustics has also incorporated machine learning into its problem-solving techniques for all kinds of problems. In this thesis, we explore the use of deep learning networks to estimate acoustic parameters of a space, and provide a comparative analysis between the architectures used and similar publications.