Abstract
The first wide-ranging attempt to characterize the visual process in terms of the transfer of visual signals defined in the spatial frequency domain was that of Otto Schade (1956). Schade's interest was in being able to incorporate the human visual observer into the complete chain of devices comprising a television system (camera, transmitter, display, observer) in order to predict the overall performance and hence optimize the engineered components. In this context Schade made the first comprehensive measurements of the minimum contrast required to detect the bars of sinusoidal gratings, an extension into the spatial domain of the popular linear systems approach to system characterization in the temporal domain based on signals modulated sinusoidally in time. Although I had become familiar with linear systems methods in the course of my PhD work analyzing human muscle control mechanisms, it took the publication of nonsense results by DePalma and Lowry in 1962 (they claimed that sinusoidal gratings were more visible than square wave gratings) to convince me to help Fergus Campbell to see if it was possible to understand the relationship of the contrast thresholds for gratings of different waveform in terms of linear systems (Fourier) theory. It was possible. Our early success (though our results were not published until 1968) encouraged me to extend the psychophysical observations to the combined spatial and temporal domains (1966) and to collaborate with Christina Enroth in applying similar techniques to characterizing cat retinal ganglion cells (1966).
Classic paper: Campbell, F. W., & Robson, J. G. (1968). Application of Fourier analysis to visibility of gratings. Journal of Physiology, London, 197, 551–566.