Visual stimuli are judged to be shorter than auditory ones. We used this well-known phenomenon to understand how auditory and visual signals contribute to perception of duration. Specifically, what duration is perceived when auditory and visual stimuli coincide (the visual or the auditory perceived duration?), and to what degree does selective attention modulate the influence from each modality? Visual and auditory stimuli of different durations (0.2 to 1 s) were either presented coincidentally or separately, and observers judged their durations in a 2AFC temporal-bisection task (typically used in the time-perception literature). Regardless of whether the auditory stimulus was stronger, subjectively matched, or weaker compared to the visual stimulus, perceived durations of auditory-visual stimuli were always the same as those of auditory stimuli presented alone. This refutes the hypothesis that auditory stimuli are perceived to be longer because they are more salient than visual stimuli, and also indicates that auditory signals dominate in time perception when a stimulus is redundantly presented through auditory and visual modalities. Can selective attention overcome this auditory dominance? We made the durations of the auditory and visual stimuli different in the bimodal condition and asked observers to judge durations of the visual stimuli. Only when the auditory stimuli were weaker (although strong enough to dominate time perception in the absence of selective attention) and were temporally embedded within the visual stimuli were observers able to filter out the auditory modality and correctly judge visual durations. Our results thus suggest that time perception predominantly depends on auditory signals regardless of the relative salience of the auditory and visual signals, attention to the visual modality is only effective under limited conditions. These limitations may be accounted for by differential use of auditory and visual modalities in naturalistic object perception.