Assessing the Visual Impact of Development Proposals: The Validity of Computer Simulations

Abstract

Determining the contribution of a single variable, or single landscape element, to visual preference is made simple if scenes with variation in only a single variable can be compared. Computer technology based on the capture and editing of video signals provides the opportunity for such comparison, provided the computer output is a sufficiently good approximation of the scene as recorded in a photograph. The degree of approximation depends to a large extent on the pixel and color resolution of the computer system. Systems offering more than 256 simultaneous colors are still quite expensive. This paper finds a moderate correlation (0.76) between original photographs and corresponding images reduced to 256 colors. Preference for the digitized scenes is generally lower than for original slides. Subsequent analysis based on multi-dimensional scaling techniques, image analysis, and scene rating for special features suggests that a better approximation is obtained when (a) the greenness level of the original scene is well reproduced, (b) a prominent cultural modification (e.g. transmission tower) is present, (c) the scene is confined to fore- and middle-ground, and (d) topographic relief is low.