Calibration of "pokes": psychophysical measurement of computer-generated temporal intervals

To answer questions raised about the accuracy of temporal intervals used to produce perception of apparent motion, interstimulus intervals generated by a computer were measured using physical and psychophysical procedures. Calibration of such devices was assumed initially to be routine but search of the literature indicated that calibration has been ignored in studies of motion and masking. Direct/physical methods of calibration proved unfeasible for assessing the accuracy of a computer programmed through "poke" procedures because signals for temporal intervals could not be differentiated from the output controlled by the machine's "hardwired" program. The results of psychophysical measurement of temporal intervals indicated that, on the average, differences, "constant errors" between expected/programmed intervals and matches of those intervals generally were less than six milliseconds.