Early vision models attempt to capture how visual information is encoded and represented in the early stages of the visual system. As with most models, the improvement of early vision models depends on suitable, high-quality experimental data to constrain their parameters.

Campbell and Robson’s (1968) foundational work, introducing the notion of independent spatial frequency channels in early vision, is one of the most influential studies in this field. Despite its significance, their study has methodological limitations, however, namely the experiment design using the method of adjustment and a very limited number of participants consisting of the two authors only. These limitations make the original data unsuitable for constraining current computational early vision models.

Here we replicate and extend Campbell and Robson by employing a two-interval forced-choice (2IFC) design with a sufficient number of participants and trials to ensure statistical power. Pilot data reproduce the contrast sensitivity function (CSF) as well as the expected contrast sensitivity ratio of 4/π for square-over-sine sensitivity. Additionally, we also replicate the seminal study by Graham and Nachmias (1971), to directly test the putative phase-independence of the multiple-channels hypothesis. Given that similar methodological limitations exist in the latter study, the experimental design again will be 2IFC with the method of constant stimuli with adequate participant numbers and trials.

Together, our experiments provide a modern, much extended replication of two seminal studies in pattern perception, allowing the results to be used to constrain the parameters of early vision models.