Quality grading of impaired television pictures

Subjective tests of pictures impaired by distortion and noise provide basic data for the design of television broadcasting systems but their complex nature is a source of difficulty. The results are affected by the many arbitrary factors involved in choosing such features as the type of psychometric test, the class of observers, and the test pictures and viewing conditions. Because these tests are inevitably costly and time consuming, it is the more regrettable that different workers have generally chosen different methods, yielding results that cannot be compared satisfactorily. As a possible basis for some degree of standardisation, a general-purpose quality-grading method is proposed. This is of the type, suitable for nonexpert observers, in which impaired pictures are presented in random sequence and the observers give their opinions in terms of a 5 grade scale of quality. Also described are statistical procedures for analysing the primary data and methods of presenting the results.