Room Acoustics Model Calibration: A Case Study with Measurements

Abstract

Due to complaints of poor speech intelligibility in a wine-tasting classroom, an acoustical investigation was conducted. Impulse response measurements were performed and a ray-tracing model was created using Odeon. Since the room's surface material properties were not definitively known, initial estimates were made and the model was calibrated based on the impulse response measurements using a recursive genetic algorithm. The proposed retrofit acoustical materials were evaluated by modifying the calibrated model based on manufacturer's data. The prediction results demonstrated a significant improvement in speech intelligibility should be expected. After the retrofit materials were installed, post-construction impulse response measurements were performed and good agreement with the modelling results was found. Subjectively, the predicted improvement in speech intelligibility was confirmed by end-user observations.