Modeling time-varying loudness requires emulating temporal integration and post-masking.

Evidence

So far we have reviewed the perception of loudness caused by tone bursts. Speech and music are composed of sounds whose spectrum changes all the time. The perceived loudness can thus be expected to change all the time. However, it has been found that the loudness of such timevarying sounds seems to be the result of a short-time integration of sound power followed by a processing of the peak values by another mechanism. The perceived loudness is thus defined by loudness peak values. For example, a sound, in which 200-ms sound bursts alternate with 200-ms pauses is perceived to have loudness similar to a corresponding continuous sound. — Ville Pulkki, Matti Karjalainen, Communication Acoustics- An…, 1177 (#complicates, via linked claim)

Categorization

• XREF: connects to: Loudness models need 100-200 ms time constants to capture temporal integration and post-masking. — Ville Pulkki, Matti Karjalainen, Communication Acoustics- An…, 1627; Static loudness models fail on time-variant sounds, motivating frame-based temporal integration models; Masking thresholds can be transformed into time functions of specific loudness (+9 more)

Broader Connections

• Perceived loudness of time-varying sounds tracks near-peak levels, not averages #complicates (0.78)
• Time-varying sound loudness is set by integrated loudness peaks, not moment-to-moment level #complicates (0.75) — Ville Pulkki, Matti Karjalainen, Communication Acoustics- An…, 1177