Print Quality Metrics for Grayscale Text

The effect of printer resolution and grayscale upon text print quality was evaluated both empirically and analytically. Empirical data on subjective print quality was obtained in two ways. First, people were asked to rate print quality on a scale of 1 to 10 where 10 was considered to be optimal. Second, people were presented with pairwise presentations of different print samples and asked to indicate which sample they preferred. Both methods indicate that text print quality increases with printer resolution and number of gray levels. The relationship between subjective ratings of text print quality and several attribute-based and image-based machine vision metrics was evaluated. Subjective print quality ratings are monotonically related t o measurements of character jaggedness and unrelated to measurements of edgr sharpness. Subjective ratings are also monotonically related to an image-based metric that attempts to quantify the perceived difference between any particular print sample and an ideal high-resolution print sample (approximated, in this case, by an offset print sample). The metric is computed by filtering the difference between digitized images of a character printed with offset technology and the same character printed with inkjet technology using published contrast sensitivity functions. Empirical print quality ratings are monotonically related to the visually-weighted difference betulern an offset print sample and the lower-resolution inkjet print samples.