An overview of issues concerning loudness measurement versus time.

Abstract

An overview of present methodologies for loudness measurement, especially considering time-varying sounds, is given. An ideal loudness-measuring method would mimic both the spectral and temporal acuities of human hearing including the relationship of those acuities at all audible frequencies. Due in part to practical considerations in their development history, none of the present set of loudness-measurement tools offers fully accurate representation of subjective time-frequency impressions. The most general limitations are insufficient spectral resolution and frequency-dependent time resolution. A recently standardized method well matches the temporal behavior of subjective loudness perception but only above a certain frequency due to lower-frequency time-resolution limits imposed by prior standards necessitating compliance. Examples will be shown of trade-offs and varying balances among several critical factors: time and frequency resolution at all audible frequencies, representation of temporal masking and growth/release effects, and appropriate spectral representations over varying combinations of broadband and narrowband signals.