A model for predicting the masked thresholds of the voiceless plosive bursts /k,t,p/ in background noise is proposed. Because plosive bursts are brief, are generated by a noise source, and have different spectral characteristics, the modeling approach accounts for duration, center frequency, and signal bandwidth. Noise-in-noise masking experiments are conducted using a broadband masker and bandpass noise signals of varying bandwidth (100-5483 Hz), duration (10-300 ms), and center frequency (0.4-4 kHz). Data from these experiments are used to parametrize an auditory filter model in which the effective bandwidth and the signal-to-noise ratio at threshold for each filter are duration dependent. The duration-dependent filter model is then used to predict the thresholds of synthetic and naturally spoken plosive bursts in background noise. Finally, results from pilot notched-noise experiments are presented which support duration-dependent frequency selectivity.