This dataset contains the data underpinning the paper "Too many shades of grey: photometrically and spectrally mismatched targets and backgrounds in printed acuity tests for infants and young children" and is released under a CC-BY-4.0 license as per https://creativecommons.org/licenses/by/4.0/ . For attribution and citation, please refer to the latest version of the paper, as reported in the dataset metadata. Full data acquisition details are reported in the paper above. A summary is given below, and data format details are reported in the readme fields in each folder of this dataset ============================= Data acquisition conditions ============================= --- Luminance and spectra --- Luminance and spectral properties of the target patterns and backgrounds are reported for four printed tests and four digital displays. The four printed tests assessed are: Teller Acuity Cards (TAC); Keeler Acuity Cards for Infants (KACI); Lea Paddles (LP) and Cardiff Acuity Cards (CAC). The digital displays assessed are: iPhone 6 and iPad 3 (Apple Inc, California, USA); a laptop screen (MSI GL62M 7RD; Micro-Star Int'l Co. Ltd., New Taipei, Taiwan) and a 4k HD monitor (Philips BDM4350; Koninklijke Philips N.V., Amsterdam, The Netherlands). The displays were turned on for at least 15 minutes prior to measurements, set to a screen brightness of 50% with auto-brightness turned off. 5 measurement points are located a set proportion away from the screen edges in each case to account for different screen sizes. Because the printed tests had different target shapes due to the nature of the tests, measurement locations differed by test. For details, refer to the open-access publication above. Luminance measurements a performed with a luminance meter (Minolta LS-100, Konica Minolta Sensing, Europe B.V.) with a 1° aperture, close-up lens (No. 110) and calibration traceable to the Japanese national primary standard. The targets were placed on a horizontal surface with the luminance meter lens perpendicularly above at a distance of 71 ± 5 mm. Three illumination conditions are examined. Ambient light uses fluorescent bulbs (Sylvania CF-LE 40W, LEDVANCE, North Carolina, USA) to represent a typical clinic room. Two LED studio lights with tuneable colour temperature and high colour rendition index (Aputure Amaran AL-H198C; Aputure, Shenzhen, China) were positioned 45° to the target surface, illuminating the target from two sides and generating cold light at 5500K colour temperature, and warm light at 3200K colour temperature. Illuminance was measured with a light meter (ISO-TECH ILM-01; RS Components, Corby, UK) at 0, 15 and 30 minutes after starting the luminance readings. Studio lights were switched on 30 minutes prior to the measurements to allow stabilisation. Average illuminance under ambient, cold and warm conditions were 450, 3682 and 3432 lux respectively, with variations of approximately 1% or less of the average value over the course of the measurements. For the four printed tests, five luminance measures are given under each of the three lighting conditions: white of coarsest grating; black of coarsest grating; grey background; average (over field of view) of second finest grating; average (over field of view) of finest grating. There were two exceptions to this: 1) It was not possible to measure average (over field of view) of second finest or finest gratings for the CAC test due to the small size of the target details, i.e. three fine lines rather than an extended grating. 2) It was not possible to measure average (over field of view) of second finest grating of the LP test with the close-up lens because the spatial frequency was too low for consistent measurement. It could be measured without the close-up lens, at 106 cm (± 1 cm), normalising subsequently to the luminance of the close-up lens condition. For the four digital displays, five luminance measures were made under two lighting conditions (ambient lights on and off), but not under warm or cold conditions as ambient illumination does not affect emission spectra. The five measures were white of full white screen, black of full black screen, and three greys: a checkerboard pattern, a vertical grating, and a horizontal grating, at the maximum resolution of each device i.e. each half-cycle of the gratings was 1 pixel, combining white (RGB 255, 255, 255) and black (RGB 0, 0, 0) pixels. Spectral measurements were made using a compact spectrometer (FLAME-S-XR1, Ocean Optics, Florida, USA), range 200–1025 nm, with an optic fibre cable (QP600-1-VIS-NIR, Ocean Optics, Florida, USA) of 600 μm core diameter. Spectral measurements of three of the printed tests (TAC, KACI and LP) were done under the same three lighting conditions used for the luminance measurements (ambient, cold, and warm), but illuminance of the ambient light was slightly lower (347 vs 450 lux). Spectral measurements of CAC were not made for the same reasons given in the luminance section. --- Psychophysical tests --- Five young (22–28 years) adult subjects, two male. The experiment was approved by the University of Strathclyde Research Ethics Committee in accordance with the Declaration of Helsinki (application number DEC/BioMed/2019/267). Each subject performed all five acuity tests (TAC, KACI, LP, CAC and a tablet-based digital acuity test, Peekaboo Vision (PV - Scottish Health Innovations Ltd, Clydebank, UK, iPad 3 only), with test order pseudorandomised and equally balanced, repeated under three lighting conditions. Subjects were tested with both eyes open, wearing habitual refractive correction as needed. Test distance was 10 m. Detection thresholds for each test were determined using a descending method of limits (coarse to fine), with lower spatial frequency targets presented only once. When subjects first incorrectly identified a target, that card and subsequent higher spatial frequency cards were each presented ten times until a subject identified ≤ 5 of the 10 correctly (chance level performance). Testers instructed subjects to indicate “on which side do you see the pattern?”, mandated a “best guess” if the subject did not know, and did not disclose whether choices were correct. A time limit of 10 s per card/level was set. Threshold was defined as 8 correct results out of 10 presentations. For the CAC test, correctly naming the object was assumed to mean a subject could see that level, and the next level was tested. If the target could not be correctly named, a forced choice method was used, with subjects indicating whether the target was at the top or bottom of the card. For the PV test, a 2-target protocol was selected.