Data-driven estimation of sound absorption coefficients considering the positions of microphones

The sound absorption coefficient of walls is important to simulate the sound reflections in a room. However, measuring the sound absorption coefficients in detail requires many measurement points along the wall surface. Recently, machine-learning-based estimation methods have been proposed for the sound absorption coefficients using the room impulse responses as training data. In this study, to obtain the sound absorption coefficients of each wall surface with free microphone placement, we proposed the machine learning-based estimation with the room transfer functions and its measurement positions. The positional information of microphones is provided to the network in the mask data format. In the two-dimensional simulation experiments, we evaluate the estimation accuracy of the proposed method compared to the fixed microphone arrangement.