IMPROVING SPEECH INTELLIGIBILITY IN UNDERGROUND STATIONS

Underground stations are typically long enclosures, for which classical room acoustics is inappropriate. A ray tracing computer model has been developed specifically for the prediction of the sound field in long enclosures. The aim of the model is to enable the acoustics of long enclosures, particularly underground stations, to be studied. In particular the model has been designed to predict the acoustic parameters associated with speech intelligibility, including sound propagation, clarity index, and speech transmission index. This paper briefly describes the development and validation of the model. It shows how the model has been used to study the sound field in long enclosures with multiple sources, of which underground stations are particular examples, and discusses the effects predicted by the model of various factors upon speech intelligibility. (A) For the covering abstract see IRRD E101083.