Body image distortion in photography

Harper, Bernard (2006) Body image distortion in photography. PhD thesis, University of Liverpool.

Text 433029.pdf - Unspecified Download (51MB) | Preview

Abstract

This thesis investigates the theory that photography is, in terms of body image perception, an intrinsically distorting and often fattening medium. In the professional practice of photography, film and television, there is a widely held belief that the camera "adds 10lbs" to the portrayed weight of actors and presenters. The primary questions addressed here relate to the true extent of the fattening effect, to what perceptual mechanisms it can be ascribed and if it can be counteracted in common practice. Current theories in the perception of photographic images rarely, if ever discuss the medium's perceptual accuracy in recording the original scene. It is assumed by many users that most photographs convey essentially the same information they would have seen had they been present when they were taken. Further, it is generally accepted that photographs are an accurate, veridical and scientific method of record and their content should be trusted unless there is evidence of a technical failure, editing or deliberate tampering. This thesis investigates whether this level of trust is appropriate, specifically by examining the reliability of photography in relation to reproducing the face and form of human subjects. Body Image Distortion (B.I.D.) is a term normally used to describe the primary diagnostic symptom of the slimming disease, anorexia nervosa. However, it is demonstrated here that people viewing 2D photographic portraits often make very significant overestimations of size when comparing otherwise identical stereoscopic images. The conclusion is that losing stereoscopic information in conventional 2D photography will cause distortions of perceived body image, and that this is often seen as a distinct flattening and fattening effect. A second fattening effect was also identified in the use of telephoto lenses. It is demonstrated, using psychophysical experiments and geometry that these 2D images cannot convey the same spatial or volumetric information that normal human orthostereoscopic perception will give. The evidence gathered suggests that the Human Visual System requires images to be orthostereoscopic, and be captured using two cameras that mimic as closely as possible the natural vergences, angle of view, depth of field, magnification, brightness, contrast and colour to reproduce scenes as accurately as possible. The experiments reported use three different size estimation methodologies: stereoscopic versus monocular comparisons of human and virtual targets, bodyweight estimations in portraits taken at differing camera to subject distances and synoptic versus direct viewing comparisons. The three techniques were used because photographic images are typically made without disparity and accommodation/vergence information, but with magnifications that are greater than found with direct viewing of a target. By separately analysing the effects of disparity, magnification and accommodation/vergence the reported experiments show how changes in each condition can effect size estimation in photographs. The data suggest that photographs made without orthostereoscopic information will lead to predictably distorted perception and that conventional 2D imaging will almost always cause a significant flattening and fattening effect. In addition, it is argued that the conveyed jaw size, in relation to neck width is an important factor in body-weight perception and this will lead to sexually dimorphic perception: disproportionately larger estimations of bodyweight are made for female faces than male faces under the same photographic conditions.

Item Type:	Thesis (PhD)
Depositing User:	Symplectic Admin
Date Deposited:	20 Oct 2023 12:40
Last Modified:	20 Oct 2023 12:56
DOI:	10.17638/03174677
Copyright Statement:	Copyright © and Moral Rights for this thesis and any accompanying data (where applicable) are retained by the author and/or other copyright owners. A copy can be downloaded for personal non-commercial research or study, without prior permission or charge.
URI:	https://livrepository.liverpool.ac.uk/id/eprint/3174677