Open source software for numerical computation

Skip to main content

dBEuler project: statistical pass-by rolling noise measurements in Scilab

By Guillaume Dutilleux, CETE

The so-called statistical pass-by (SPB) measurement remains the reference method when it comes to classifying road pavements with respect to traffic-induced noise emission. Its principle is to record vehicle speed $v$ and maximum sound pressure levels $L_{Amax}$ for a large set of vehicle pass-bys in different categories like Light Vehicles or HGV. From a regression analysis where $L_{Amax}$ is the dependent variable one obtains a reference noise level for the investigated pavement section.

dBEuler has been developed mostly in Scilab as a free software solution for SPB measurement campaigns. Combined with a radar cinemometer and an acoustic pressure sensor, it provides all functionalities required to record calibrated traffic sounds, compute time- and frequency-weighted acoustic descriptors, extract pass-bys from recordings, tell invalid pass-bys from correct ones, perform robust regression analysis, and generate a pre-press quality measurement report.  dBEuler also uses Scilab as a "glue-language" to rely on dedicated C, XSLT, and Tcl/Tk components plus a few third-party free software tools like LaTeX, SoX or ImageMagick. The audio channels can be brought either by an audio interface or an external audio recorder.