Scaling is one of the most efficient methods for assessing perceptual image quality and its underlying dimensions (sharpness, brightness, colourfulness etc.) . Experiments with simple stimuli such as squares, circles and dot patterns have shown, however, that the outcome of a scaling experiment is susceptible to contextual effects [2, 3]. That is, the response to a stimulus depends not only on the stimulus itself but also on the other stimuli to be judged in a session. Contextual effects due to stimulus spacing or frequency of occurrence of stimuli have been found to have a substantial influence on the results of a single stimulus (or ‘direct’) scaling experiment . The present study examined whether contextual effects are also present when complex stimuli, in particular images of natural scenes, are evaluated. To this end, two experiments were carried out measuring the possible influence of stimulus spacing on the evaluation of perceived sharpness for the following three scaling techniques: single stimulus scaling, double stimulus scaling and a scaling procedure based on difference judgements. Note that these techniques represent the three kinds of evalu ation methods recommended by the International Telecommunication Union (ITU), formerly CCIR . The first experiment concerned the use of three different response scales during single stimulus scaling. The second experiment was a direct comparison between the three above-mentioned scaling techniques using a 10-point numerical category scale only. In both experiments the stimulus set consisted of blurred versions of one static image of a terrace scene.
|Title of host publication||Advanced methods for the evaluation of television picture quality : proceedings of the MOSAIC workshop, Eindhoven, 18-19 September 1995|
|Editors||R. Hamberg, H. Ridder, de|
|Place of Publication||Eindhoven|
|Publication status||Published - 1995|