Detection of silent temporal gaps in sinusoidal markers

Area of Science:

• Auditory perception
• Psychoacoustics
• Signal processing

Background:

• Gap detection is crucial for speech and music perception.
• Previous studies were confounded by gating and duration cues.

Purpose of the Study:

• To measure gap detection thresholds under controlled conditions.
• To minimize confounding factors like gating transients and duration cues.
• To model gap detection performance using auditory filter characteristics.

Main Methods:

• Adaptive forced-choice procedure for measuring gap detection thresholds.
• Sinusoidal markers with varying frequencies (F1, F2) and controlled gap durations.
• Minimized extraneous cues using continuous noise masker and randomized marker durations.

Main Results:

• Gap detection thresholds increased with increasing frequency separation between markers.
• Thresholds ranged from 5 to 90 ms across tested conditions.
• Data closely fit a roex (generalized power law) filter model.
• Listener performance was modeled by a constant-percentage (7%) bandwidth filter centered on F1.

Conclusions:

• Auditory gap detection is frequency-dependent.
• The roex filter model effectively describes gap detection performance.
• A constant-percentage bandwidth filter provides a good approximation of auditory processing for gap detection.