Sound-Insulation Ratings from Third-Octave-Versus Octave-Band Measurements

Abstract

Methods proposed internationally for the rating of impact-sound and of airborne-sound insulation allow use of measurements made with either 12- or 1-oct bands of noise; observed values are compared with a standard sequence conveniently described by a curve that is shifted until there is an average unfavorable deviation between 1 and 2 dB; the value of the shifted reference curve at 500 Hz is then the single-number rating. Peaks and valleys are delineated nmch less clearly by 1-oct-band measurements than by 13-oct-band measurements, and in recognition of this fact the proposed international standard allows a maximum deviation from the rating curve of 8 dB for 13-oct bands but only 5 dB for 1-oct bands. Ratings obtained with the two different bandwidths were investigated with a variety of source spectra and families of sound-insulation curves having coincidence dips at different frequencies. Extreme examples of impact-sound produced by the standard tapping machine were selected for calculation. In general, a rating obtained from 1-oct-band data was like that obtained from the corresponding 14-oct-band data within 1 dB. Moreover, the analysis showed that nearly the same rating of sound insulation could also be obtained simply as the difference between the sound level A in the source and receiving rooms.