Avaliação de modelos estatísticos para simulação da cauda reverberante em acústica de salas

Abstract

The ray tracing algorithm, since its creation in 1968, provides a simple yet reliable solution to the room acoustic simulation problem. In certain rooms, however, properly estimating the reverberant tail using the ray tracing technique can be very time consuming. In this work, we describe a way to deal with this constraint, by definig a statistical model for the tail of room impulse responses. The statistical synthesis is computationally more competitive than properly estimating the reverberant tail via ray tracing. We evaluate the proposed method by comparing simulated and real acoustic descriptors for different rooms.