WHITENESS ASSESSMENT OF PAPER SAMPLES AT THE VICINITY OF THE UPPER CIE WHITENESS LIMIT

摘要

The perceived whiteness of commercial paper samples was evaluated under three different illumination conditions. Several samples were found to be outside the region of validity of the CIE whiteness although they were perceived as acceptably white by most observers. If the limits are disregarded, these samples are assigned high whiteness values, leading to a low correlation between instrumental and visual whiteness since these samples are not perceived as the most white. Non-linear whiteness models defined in the whole colour space have recently been proposed by Aksoy and Uchida. These were tested together with two models proposed by the authors. These models are modifications of the Aksoy model. In the first model, Wmw, the maximum whiteness is at a lower b~* value than the CIE whiteness. It correlated well with ranking and scaling [R~2 > 0,7). The higher correlation obtained with this model indicates that the upper whiteness limit in the CIE whiteness could be extended, as was proposed by Uchida. The second model, W_(eCIE). keeps the CIE whiteness formula within its region of validity and applies a penalty function similar to the one used by Aksoy outside this region. In this case the maximum whiteness is determined by the CIE whiteness upper limit. The performance of this model was slightly inferior (Rank correlation R~2 = 0.56 for a first set of samples and 0.69 for a second set) but its main advantage is that all samples within the CIE whiteness limits are given the same whiteness values as with the CIE whiteness formula. The results showed that the introduction of a penalty function replacing the CIE whiteness limits is a promising way to handle white materials at the vicinity of the upper CIE whiteness limit.