Methodology for development of airborne sound insulation descriptor valid for light-weight and masonry walls

Evaluating the adequacy of single number quantities that are used to assess the sound insulation performance of walls by listening tests is hampered by the difficulty to perform large numbers of subjective tests, due to the need for combining many stimuli, each representing different source/wall combinations. In this article, this problem is tackled by an alternative approach that makes use of Loudness based validation of proposed single number quantities. The adequacy of the newly proposed method was assessed by comparing wall rankings based on Loudness on one hand, and on laboratory listening tests on the other hand. Perceptual comparison of sounds passing through different types of walls between dwellings was performed by a listening test procedure that focused on differences in perception of sound insulation of a set of real-life heavy weight (HW) and light weight walls (LW), and synthetic variants of those. The latter were upshifted or downshifted versions of real wall insulation spectra. The listening tests were based on a three-alternative forced-choice testing method. Altogether, 40 test subjects with normal hearing ability were asked to rate the loudness of sounds transmitted through different walls, which were simulated by filtering a set of real-life sound stimuli. The resulting ranking of walls in order of isolation performance was compared with rankings based on the Loudness of the transmitted sound, and on the value of single-number quantities, that are commonly used for objective evaluation of the sound insulation i.e., Rw and Rw + C50–5000. Special attention was given to the weight of low frequency isolation values in the subjective assessment and in the calculation of the single number quantities. Two new single-number quantities, Rmod and Rmod,2 are proposed, which are found to fit better to subjective perception of sound insulation than the investigated existing quantities.