A Technique and a Scale for Loudness Measurement

TLDR

The paper proposes a method for loudness scaling that yields a true ratio scale. The method employs two loudness functions—one derived from equisection judgments and another from fractionation judgments—whose interrelation determines the constants required for a ratio scale. Experiments confirm the method’s validity, reduce observer differences, produce a flatter loudness function, and demonstrate better agreement with equal‑loudness intervals, monaural‑binaural data, and multicomponent tone measurements.

Abstract

A method of loudness scaling is presented for obtaining a true ratio scale of loudness. Two separate loudness functions are required, one based on equisection judgments and the other based on fractionation judgments. Interrelation of these two scales allows determination of all constants necessary for a ratio scale. The method provides validation of the assumptions underlying its use. Experimental data are also presented which show that (1) the method is valid for individual observers; (2) differences between loudness scales of different observers are considerably reduced with the method; and (3) the loudness function obtained with this method is much flatter than the standard scale now used. It is then shown that this new loudness scale is in better agreement with other data on equal loudness intervals than the standard scale. It is also shown that this scale is in agreement with monaural-binaural loudness data and data on loudness of multicomponent tones.