This paper investigates the effect of prior knowledge of a room`s geometry on the accuracy of predicting its acoustic properties. The motivation comes from the need for the creation of a framework that reduces the complexity of modeling real spaces. The project aims to model cultural heritage spaces for film-making applications, like location scouting. Preliminary investigation is performed on an existing dataset, where reverberation and early decay time are estimated in several octave bands through audio feature-driven regression. The room volume is treated as a feature, and performance is calculated with and without including it in training. Results indicate a slight improvement in performance in most bands. This leads to the conclusion that additional geometrical information should be provided during dataset creation for the significant improvement of robustness.
(See document for exact affiliation information.)
https://aes2.org/publications/elibrary-page/?id=22115
(364KB)
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member Join the AES. If you need to check your member status, login to the Member Portal.
