The calm before the storm: Silence compresses subsequent visual time without reducing temporal precision

Silence is often conceptualized as the mere absence of auditory input; however, accumulating evidence suggests that the omission and offset of sound are processed as psychologically salient perceptual events. This study investigated whether an abrupt transition from sound to silence biases the perceived duration of subsequently presented visual stimuli. Across four psychophysical experiments, participants performed forced-choice temporal discrimination tasks in which a brief silent interval (sound offset) was inserted between a standard and a comparison stimulus. Visual stimuli presented immediately following the offset were consistently perceived as shorter in duration than physically identical stimuli presented under continuous background noise. This duration compression effect was robust across simple geometric stimuli and generalized to dynamic video content. Critically, sound offset significantly shifted the point of subjective equality (PSE) while leaving temporal sensitivity measures, including the difference limen (DL) and Weber fraction, unaffected. We also conducted additional experiments demonstrating that the effect could not be attributed to graded changes in sound intensity and persisted across various interstimulus intervals (ISIs). These findings indicate that sound offset selectively compresses perceived visual duration without degrading temporal discrimination precision, supporting the view that silence functions as a categorical contextual event that recalibrates the temporal reference framework for subsequent duration judgments. Collectively, the results highlight the role of auditory context transitions in shaping multisensory time perception. OPEN PRACTICES STATEMENT: The trial-level data for all experiments are publicly available on the Open Science Framework (OSF) at https://osf.io/3ksa8/overview . The experimental scripts and analysis code are available upon request. None of the experiments were preregistered.