Assessment of adequacy of single number quantities for wall insulation performance based on the Zwicker’s loudness of transmitted indoor noise

Abstract

Previous work suggested a method that makes use of the Zwicker’s loudness value of transmitted sound to validate single number quantities (SNQs) determining the airborne sound insulation of walls in dwellings. The correlation between the ranking of walls based on a SNQ, and the one based on transmitted sound loudness was used to qualify the adequacy of that SNQ. The underlying idea is that Zwicker’s loudness adequately represents people’s perception of loudness and that evaluations based on calculated loudness can substitute listening tests. The current work compares 16 SNQs: eight standardized or proposed in past studies, and eight new ones. The results are more robust, since more (56) walls were included, among which also highly isolating ones. SNQs are assessed for the indoor noise stimulus recently suggested by CSTB, France, and also for four filtered variants, and for two levels, the ‘usual’ or ‘normal’ level (as reported by CSTB), and that level increased by 10 dB, expressing the less common but psychologically and socially relevant situation of excessive neighbour noise. The adequacy of SNQs is analysed per wall category, (1) lightweight, (2) heavyweight, (3) CLT-based and (4) combined walls and per stimulus type. Three approaches are used to quantify the correlation between SNQ-based wall insulation performance rankings and Loudness based ones: (i) Spearman’s correlation coefficient, (ii) correlation and (iii) inverse RMS residual of fitting with a parametrized function that has been empirically conceived so as to describe as good as possible the correlation SNQ-Loudness correlation behaviour for the considered population of walls.