Broadband maskers overshoot-effect short tones but narrowband maskers do not

Evidence

This effect (sometimes called the overshoot effect) also depends on the spectral content of masker and test sound. The effect disappears if the two sounds have similar spectra. Figure 4.21 gives examples for two different bandwidths of the masker. First, the masker is white noise (as used in Fig. 4.20), and the results are indicated by the dotted line. Second, the solid curve shows the results for a masker centred at 5 kHz and having the width of the critical band. The delay time, Llt, between the onset of the masker and the onset of the 2-ms test-tone burst, is the abscissa. The overshoot effect for broad-band maskers is largest at very small Llt. It diminishes for larger values of Llt and disappears completely after about 200 ms. The overshoot effect does not appear for narrow-band maskers. — Professor Dr.-Ing. Hugo Fastl, Professor Dr.-Ing. Eberhard, Psychoacoustics- Facts and Models, 512

Categorization

• DEFINE: Names and delimits the overshoot effect: broadband-only, largest at tiny onset delays, gone by ~200 ms.
• QUESTION: Why does overshoot require a broadband masker specifically? Mechanism likely involves auditory transient processing differences, worth chasing in the model discussion.
• corroborated: 2 independent passages support this claim — Professor Dr.-Ing. Hugo Fastl, Professor Dr.-Ing. Eberhard, Psychoacoustics- Facts and Models, 512
• XREF: connects to: Short test tones are masked up to 10 dB more right after masker onset; Short maskers yield significantly larger suppression estimates than long ones — Dijk, Pim van, Physiology, Psychoacoustics and…, 940; Fully masked spectral components contribute nothing to perceived sound