Concert Hall Acoustic Design: An Alternative Approach

Abstract

A neural network analysis was undertaken to relate the acoustic quality of concert halls to Beranek's six acoustical parameters: IACC, Gmid, EDT, BR, SDI, and TI. The results of the analysis are difficult to generalise but the present work indicates that there is still no reliable method of predicting the acoustic quality of halls. The limitations of the data and the methodology are explained and the results of the analysis are applied to the Concertgebouw, in Amsterdam, to see how changes in the acoustic parameters might change the acoustic quality of this hall. It is also shown that several different combinations of parameters can give reasonable predictions of acoustic quality. Finally, it is shown that there may be a better way of predicting the acoustic quality of halls which uses some of Beranek's acoustical parameters and the number of seats in the hall.