Low frequency modeling - experimental study on the influence of source database in mitigated scenarios

European Noise Directive requires to draw up action plans to mitigate exposure: nowadays, most frequent noise mitigation measure are still barriers. In fact, they are very effective reducing traffic noise, especially along motorways and extra-urban roads. However, some studies evidenced that, after mitigation, annoyance rises: in fact, the installation of a barrier decreases the low frequencies masking effect due to presence of high ones. Therefore, annoyance and complaints may rise despite having spent money for mitigation. There should be methods allowing noise action planners to take into account this effect so that they can estimate with spectral detail the mitigation available solutions, included barrier installation. In this paper, we will illustrate how low frequency estimation after the mitigation measure may differ between some main noise emission models applied in national standards; we will assess the difference between methods according both C-weighted and A-weighted levels. Each national standard method, used for noise mapping, has its own database with different vehicle reference spectra, which takes to different estimated scenarios. In addition to these methods based on standard databases, we will show results obtained using measured local traffic spectra as input of a simplified propagation model. At last, we will evaluate how traffic speed, percentage of heavy vehicles and obstacles influence models differences