The Validity of CCIR Quality Indicators Along a Graphical Scale

Several scaling methods are used to quantify the perceptual quality of television images. One method used is the double-stimulus continuous quality scale recommended in CCIR Rec. 500-4. In this method a graphical scale is divided into five intervals with which quality terms are associated. The aim of the research discussed in this article was to quantify the perceived distances between a number of Dutch-language terms, including the translated CCIR terms, that are used to describe the perceived quality of television images. It is shown that the translated CCIR quality terms do not divide the quality scale into intervals of equal sizes. Also, in view of the large standard deviations of the average positions of the quality terms, it is concluded that a graphical scale divided into intervals with which quality terms are associated cannot be used as an interval scale. This implies that numerical processing of scores (such as averaging) is highly questionable. A 10-term ordinal scale is proposed when an indication of the absolute quality of display systems is needed.