Evaluating Live Loudness Meters from Engineers' Actions and Resulting Output Levels

摘要

Qualities of loudness meters were investigated by running a simulated live broadcast show, mixed by subjects with different degrees of audio engineering experience. The subject-induced fader movements and the resulting output levels were analyzed using a general linear mixed model and then interpreted according to a proposed set of quality criteria for live loudness meters. The results show that, for subparts of the program, meters with slower integration time caused less dispersion of the resulting output levels and a smaller error to a given target level. The varying integration times of the meters did not cause any significant differences in reaction time. It is concluded that this type of data may be used in evaluation of loudness meters and the specific study provides insight into the effect of time scales and meter implementation for EBU R 128-compliant meters.