Anatomy of a flash. 2. The 'width' of a temporal edge

In a previous study it has been found by using the Crawford masking paradigm with spatially coextensive stimuli that the end points of even a weak flash have a strong masking effect on the probe pulse, but no threshold elevation was observed when the probe was 100 ms or longer from flash onset and offset. The relationship between the strength of a temporal edge (eg the amplitude of a sudden increase in field luminance) and the duration of masking has not been systematically investigated. In the present experiments the duration of masking produced by a step of luminance as a function of step size was measured. With steps whose amplitude was between ten and five hundred times detection threshold, the duration of masking was about constant, ie 70 and 50 ms for on-responses and off-responses, respectively. This refractory period defined the 'width' of a temporal edge. Unless two suprathreshold unidirectional luminance changes are separated by an interval exceeding the width of a temporal edge, they are perceived as a single change in brightness. A linear systems analysis led to predictions for ramp stimuli that were confirmed by experiment.