Acoustic comfort in highly attended museums: A dynamical model

High-attendance exhibitions are often uncomfortable, due to too loud background noise. This is due to several factors, primarily the human voice. Moreover, most of the largest museums are housed in historical buildings which cannot be acoustically corrected, therefore the control of human noise is the only way to manage acoustic comfort. Human noise due to visitors is treated as a non-stationary phenomenon, related to flux of visitors/talkers. In the present study, a predictive model is proposed. Markov-chain theory is proposed to study the temporal behaviour of occupancy and generative algorithms are used to analyse the spatial distribution of visitors. The model has been validated by means of measurements in a highly attended museum in Florence during a free-entrance day. Measurement results agree with the predictive model, returning useful information on vocal effort and feedback processes. Results also show that a maximum number of visitors and a maximum visit time should be established in order to keep a proper acoustic comfort. Excess number of visitors or visiting time, may result in an almost two-fold increase in human noise level.