Estimation of the background noise levels in large atria with known room acoustic properties and function

One of the major identified acoustic problems in large halls such as atria or shopping malls is continuous unpleasant background noise, based on typically long reverberation of sound. Levels of background noise caused by internal sound sources depend on activity performed in the place on one hand and on room acoustic conditions on the other hand. Present noise can be measured relatively easily, by means of sound level meters and evaluated through statistical noise analysis, expressed e.g. by histogram of noise levels. However, the prediction of statistical noise levels, such as L₉₅, L₉₀ or even L_eq, in rooms is very difficult, as it doesn’t depend only on the overall sound power spectrum of sound source (typical sound source definition in ray-based algorithms), but also on the time varying character of sound. Function and thus activity performed in an atrium plays therefore an important role in terms of acoustic comfort analysis and without the information on time variation in source signal, parameters such as L₉₀ or L₅ cannot be successfully predicted. 

In this paper several typical signal sounds (such as human steps, speech, music, ventilation etc.) are synthesized and auralized in Odeon software and assessed in parametric study of an atrium, followed by statistical analysis of predicted auralized wave files. Correlation between the activity and room acoustic parameters is calculated and estimation of background noise levels L₉₉ is shown together with equivalent sound pressure level L_A,eq.