Architectural treatments for improving sound fields for public address announcements in underground station platforms

This study investigates the effects of subway platform design elements, such as finishing materials, on the temporal and spectral characteristics of public address (PA) announcements. To reproduce the acoustic conditions of the simulation model, acoustic-fitting using ray-tracing models was performed for two existing subway stations with island platforms using field measurement data. A systematic investigation using a computer station model was then carried out to clarify variations in the sound field characteristics when modifying station dimension and replacing finishing materials. It was found that smaller station dimensions yielded a higher speech transmission index (STI) and sound pressure level (SPL) with low early decay time (EDT), whereas higher absorption materials yielded a higher STI with lower SPL and EDT in the platform sound fields. In particular, a smaller station cross-sectional area obtained by modifying the station width and corridor height was the most effective for improving STI. More absorption on the floor and ceiling surfaces in the platform area had a marked influence on improving STI. In addition the auralized PA sounds with anechoic male and female announcements were analyzed using spectrogram and auto-correlation/interaural cross-correlation factors. It was found that the absorptive treatments contributed to clearer temporal characteristics of the PA sounds with decreased binaural dissimilarity and without any changes in spectral components.