Pulsed fluoroscopy (hereafter called pulsed) at reduced acquisition rates, typically 15 acq/s (pulsed-15), is proposed to reduce x-ray dose in interventional procedures. However, since the human visual system (HVS) acts as a temporal low-pass filter that interacts with such acquisitions, the proper dose for pulsed must be obtained in perception experiments. We determine the dose for low-frame-rate pulsed that gives visualization equivalent to that of conventional 30 acq/s fluoroscopy, hereafter called continuous. Computer-generated phantoms are used. They consist of stationary, low-contrast disks on a flat background containing Poisson noise that mimics quantum noise in fluoroscopy. Image sequences are displayed on the video tachistoscope, a device with considerable display flexibility. Three experimental paradigms are used. (1) In a paired-comparison study, pulsed and continuous are displayed side-by-side on the same monitor, and the visibility of a contrast detail phantom is compared. (2) Using this same display, subjects record the minimally detectable disk contrast (the min-contrast measurement). (3) In a four-alternative forced-choice experiment, a disk is placed in one of four positions, and the subject determines the position of the disk. The methods are complementary--the forced-choice experiment properly eliminates the subjectivity of the observer threshold while the paired-comparison study is much more time efficient. With regard to pulsed and continuous comparisons, remarkable similarity is found between the supra-threshold experiments (1 and 2) and the detectability experiment (3); i.e., the average absolute differences in the equivalent-perception dose as determined by the three measures is approximately 3%. No difference is found between interlaced and noninterlaced display. A relatively small dependence of dose savings on disk size is found with larger disks giving increased dose savings. Average dose savings of 22%, 38%, and 49% are found for pulsed-15, pulsed-10, and pulsed-7.5, respectively.