On using ‘‘loudness-weighted’’ SEL, LEQ, and DNL to assess noise environments

LEQ and DNL are commonly used to assess the long-term noise environments around airports, roads, etc. The fundamental building block to these assessments is the A-weighted sound level or SEL. These are used to compute LEQ or DNL. But many problems have surfaced over the simple use of A-weighting. The A-weighting has been shown to be deficient in assessing sound with strong low-frequency content, and other characteristics, like impulsiveness, are not properly accounted for by the A-weighting. However, no other simple filter has been shown to be superior to the A-weighting filter. One can consider building a more complicated filter that dynamically changes with level and frequency to better reflect human response than does the A-weighting. The equal loudness contours expressed in phons offers a set of curves that can be used to design such a filter. Like A-weighted sound level or A-weighted SEL, signals processed with such a new ‘‘filter’’ would be termed ‘‘loudness-weighted’’ sound level or ‘‘loudness-weighted’’ SEL. This author has analyzed many sounds to illustrate, test, and evaluate the concept of loudness-weighted SEL. This paper discusses the concept of using ‘‘loudness-weighting’’ in place of A-weighting and the results indicated by the analysis of many common sounds.