Fig 1.
Discretized versions of letters were displayed on an array of LEDs with brief single or multiple flashes or with steady emission of light.
Four examples are provided here; the letter K shows how a letter would be positioned within the array.
Fig 2.
In Experiment 1 respondents were asked to identify letters that were displayed with a single flash with intensity being varied.
Statistical models showing the probability of recognition against flash intensities are plotted in the two panels. Individual models for the eight respondents are shown on the left. Each showed a monotonic rise in recognition as intensity of the flash was increased. The right panel shows the group model along with a 95% confidence band.
Fig 3.
In Experiment 2 letters were displayed with multiple flashes across a duration of 750 ms, varying flash frequency.
Letters could be reliable identified irrespective of frequency, but the key question here was at what frequencies would the flashes be seen as fused, i.e., appearing as steady emission. The probability of flicker detection is given on the ordinate against flash frequency on the abscissa of each panel. The individual models (left panel) shows that respondents saw the sequence of flashes as flickering at frequencies below 10 Hz and the flashes appeared as steady non-flickering emission at frequencies above 20 Hz. The right panel shows the group model and its 95% confidence band.
Fig 4.
In Experiment 3 each letter was displayed with a perceptually fused sequence of flashes at 24 Hz for 750 ms, and intensity was varied to determine how the probability of recognition changed.
The left panel shows the individual models of respondents; each showed a monotonic rise in hit rate with increasing flash intensity. The right panel shows the group model plus the 95% confidence interval.
Fig 5.
On each trial of Experiment 4 a letter was shown twice, once as a 24 Hz flash sequence and also as a steady display, each for 750 ms.
The intensity of each flash was fixed, and the intensity used for steady displays was varied. [Note that scale values are in nanowatts/sr rather than microwatts/sr.] Respondents judged the two display conditions to be different in brightness at each end of the range and as the same brightness in the middle of the range. The Ferry-Porter law specifies that the average intensity of the flashes should equal the intensity of steady displays. They were not quite the same in this experiment, but the degree of match is impressive given the extreme brevity of flash durations.
Fig 6.
This is an adaptation of a figure provided by Broca & Sulzer [22], wherein the perceived brightness of brief flashes was compared to steady levels of light.
The two were seen as approximately equal in brightness at very long intervals where the flash judgments were becoming asymptotic. At short flash durations there was a very large difference in the perceived brightness at high flash intensities. Additionally, there was a shift in where the peak was manifested, being seen at progressively shorter flash durations as flash intensity was increased.
Table 1.
Letter Attributes.