In this article we analyze the visual ways of perception - seeing, observing, and gazing -, and discuss the double dependence of the process of perception on theory. In doing so we argue that open methodological and epistemological questions can be answered by differentiating the process of perception into three steps: pure vision, cognitive representation, and transformation of perception into words. In addition, the processes of perception have to be further differentiated in conscious and unconscious levels of awareness. As existing empirical analyses of the influence of diverse variables on the results of perception indicate, the process of visual perception is highly sensible and includes various sources of methodological artifacts. The suggested two-dimensional "visual difference", consisting of the difference between image and language as well as of the difference between presentative and discursive symbols, should sensitize researchers towards diverse processes of screening and selection. This concept should therefore be systematically integrated in the interpretation of all processes of perception. The question of the methodological quality of visual studies remains in the focus of the analyses as it becomes clear that any process of perception is always subject to the individual observer's multiple mechanisms of screening and selection. This article therefore constitutes a first attempt to outline consistent "quality criteria" for visual studies.