Morgan M J, Ward R M
J Opt Soc Am A. 1985 Jul;2(7):1205-10. doi: 10.1364/josaa.2.001205.
Thresholds for spatial-interval discrimination were determined under conditions designed to introduce randomness in the spatial-frequency content of the stimuli from trial to trial. Neither a random scaling of the display nor the addition of flanking bars at varying distances affected the observer's ability to judge the relative interval between pairs of bars, provided that the flanking bars were clearly resolved from the targets. Performance deteriorated only if the flanking bars were too close (less than about 2 arc min) or if the display was optically blurred. We conclude that the results pose difficulties for spatial-frequency theories of interval discrimination.
