We conducted a subjective survey on image quality using a set of photographs printed on commercially available inkjet printers. The purpose of the survey was to gain first hand insight into customer preferences in digital photographic printing.The observers in this survey were participants at the IS&T NIP16 Conference and the Diamond Research Corporation's Digital Imaging Conference. Both conferences took place in October 2000. In this survey, an image of a newlywed couple was used and six attributes were studied including: image defects such as blur (unsharpness), noise (graininess), and banding (effect of a clogged nozzle); personal preferences such as color rendition and tone reproduction; and finally, printer type. The observers were asked to rank order four images of different quality levels for each of the six attributes.The results show that: a) the human visual system is very acute at detecting blurriness in an image, b) the presence of image noise in luminance is much more detectable than in the color channels, and c) banding due to missing cyan ink and yellow ink appears to be more readily detectable than banding due to missing magenta. In terms of color rendition and tone reproduction, a greenish cast is objectionable to most survey participants and darker images are preferable to lighter ones. In terms of printer type, the results suggest that the participants have a consistent preference for images from certain brands of printers over others. In this paper, the design of the experiment and the subjective analysis results will be discussed in detail.The significance of this subjective survey is in the unusually large number of participants (close to 130) and also the worldwide representation of the participants (12 countries). We believe that the methodology used in this study and the survey results should be of interest to most in the business of digital imaging.
Gretchen Gast, Ming-Kai Tse, "A Report on a Subjective Print Quality Survey Conducted at NIP16" in Proc. IS&T Int'l Conf. on Digital Printing Technologies (NIP17), 2001, pp 723 - 727, https://doi.org/10.2352/ISSN.2169-4451.2001.17.1.art00063_2