Multiple reference impairment scale subjective assessment method for digital video

This paper proposes a novel subjective impairment testing method called multiple references impairment scale (MRIS) method in order to eliminate or minimise the contextual effects and large variance/standard deviation existing in the current subjective video impairment test methods. In MRIS, references using distorted pictures for each impairment level are provided and the subjects are asked to determine the impairment level by comparing the processed picture with the predetermined reference distorted pictures. Therefore the conventional categorical rating task is turned into several pair comparisons, which human subjects are accustomed to. Experiments have been conducted to compare the proposed MRIS method with the DSIS II method; the results reveal that the proposed method has much less variations in the subjective data, does not suffer from contextual effects and is good at differentiating small changes in impairment levels.