Purchase this article with an account.
Johan Wagemans, Andrea van Doorn, Huib de Ridder, Mieke Leyssen, Jan Koenderink; Measuring pictorial space in paintings: Converging operations. Journal of Vision 2011;11(11):66. doi: 10.1167/11.11.66.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Although a picture is a flat object covered with pigments, human observers often experience a 3D “pictorial space” when looking “into” the picture. This space is a mental entity, apparently triggered by so-called pictorial cues. In the case of paintings or drawings, the visual artist has put patterns of pigment on the canvas to trigger certain experiences in prospective viewers. Several experiments were carried out with the same stimulus, a copy of a wash drawing by Francesco Guardi, representing an imaginary landscape with much pictorial depth but no “ground truth”. In a first series of experiments, we selected 5 points that were well-localized in pictorial space. On each trial, one of these points was designated as target and another one contained a 3D pointer, whose slant and tilt values could be adjusted by the observer until it appeared to point to the target, in the perceived 3D space. In a second series of experiments, 49 localizable points were selected and on each trial two points were marked. Here, observers simply had to indicate which of these points appeared closer in depth relative to the viewer. In a third series of experiments, 10 of these points were selected and presented in pairs. Now, observers could adjust the relative sizes of two discs on these locations, making use of size constancy to indicate the perceived relative distances between them. We have explored various methods to operationalize the geometrical properties of the 3D space as perceived by the viewer. Examples include pictorial depth, either in a metrical or a mere ordinal sense. In general, we find that different observers tend to agree remarkably well on ordinal relations, but show substantial differences in metrical relations (scaling). Moreover, when comparing the results of the different methods tested on the same observers, we find considerable consistencies.
This PDF is available to Subscribers Only