Revisiting the loudness of sounds with asymmetric attack and decay

Abstract

Stecker and Hafter [J. Acoust. Soc. Am. 107, 3358–3368 (2000)] compared the loudness of sounds whose envelopes had a fast attack and a slow decay (designated F-S) and a slow attack and a fast decay (designated S-F). They found that, for sinusoidal and broadband noise carriers, S-F stimuli were louder than F-S stimuli of equal energy. They argued that this effect could not be explained by current models of loudness and that the loudness effect may be related to the parsing of auditory input into direct and reverberant sound. Subsequent work has shown that the differences in loudness between F-S and S-F stimuli can be partially accounted for by the loudness model of Glasberg and Moore [J. Audio Eng. Soc. 50, 331–342 (2002)], which incorporates a form of temporal averaging that is asymmetric in time. However, the model does not account for the context effect found by Stecker and Hafter. The largest differences in loudness between F- S and S-F stimuli occurred after pre-exposure to a F-S stimulus. This may happen because, when successive sounds have similarly slow decays, the decaying part is attributed to room reverberation and contributes less to loudness.