When evaluating monitors a controlled environment is essential to ensure that consistent and repeatable results are obtained. For this group test, our test bed consisted of a single PC running Windows XP Professional, fitted with a 64MB ATi Radeon 7500 graphics card with a RAMDAC of 350MHz. This has both a DVI and D-SUB port allowing us to test monitors that are equipped with either or both of these interfaces. In every case, the dual head function of the graphics card was disabled and each monitor was tested with only one connection at a time using the cables supplied by the manufacturer. All monitors were positioned in the same orientation on our purpose-built test bench, separated by at least one metre from other equipment that may produce electromagnetic fields and/or interference. In addition, the monitors were left to warm up for at least one hour before assessment under controlled lighting conditions. Before analysis we also installed the appropriate drivers supplied by the manufacturers as well as any software associated with each model.
DisplayMate for Windows Tests
All CRT tests were conducted at a resolution of 1,280 x 1,024 at a refresh rate of 85Hz, and a 32bit colour depth. Furthermore, each monitor was set to its respective default colour temperature and degaussed before testing commenced. After the units were warmed up, we initially used DisplayMate for Windows (available from www.displaymate.com) to standardise each CRTs black level and white level in order to optimise contrast and greyscale reproduction. The intensity range was also checked, ensuring that each CRT could simultaneously reproduce the extremes of the greyscale.
Once calibrated, we ran through DisplayMatesâ€™ comprehensive suite of test screens in order to evaluate overall image quality. These tests are briefly described below under four generalised sections, where each CRT was given a fail, average, good, or excellent performance rating:
Geometry and Distortion
Tilt â€“ used to assess and correct any signs of image rotation in relation to the bezel.
Pincushion/barrel distortion â€“ a test screen composed of parallel lines running around the perimeter of the picture. Used to check for sides that curve inwards or bulge outwards
Horizontal and Vertical curvature â€“ A simple series of parallel lines that run in both planes across the screen, used to check for curvature.
Fine geometric linearity â€“ A fine crosshatch pattern used to check for small-scale non-linearities. The boxes should appear square, and equally sized.
Power Regulation â€“ A series of blinking white and black areas that reveal signs of image expansion and contraction over the whole screen or in localised areas.
Sharpness and Resolution
Horizontal and vertical line, bar and wedge resolution â€“ tightly spaced vertical and horizontal lines, bars, and wedges labelled by fractions of full resolution are used to determine how well the alternate black and white lines are resolved.
Video bandwidth and transient response â€“ A set of test screens used to assess how well the monitorâ€™s electronics handle the fast transitions in the video signal required to display a pixel.
Corner resolution â€“ a test used to assess the relative sharpness of tightly spaced vertical and horizontal lines in the corners of the screen, in comparison to the screenâ€™s centre.
Fine focus matrix â€“ a graphical pattern for determining any variations in focus over the entire display.
Colour and Greyscale
Horizontal and vertical colour registration â€“ a series of test screens with multi-coloured, segmented lines that reveal how well the red, green, and blue beams converge together in both planes.
Streaking and ghosting â€“ A number of test screens designed to expose how well the monitorsâ€™ electronics switch off/on the beams when there are large changes in intensity between adjoining areas of the screen.
Colour purity â€“ Red, green, and blue test screens are used to check for intensity variations and patchiness across the screen, indicating each beamâ€™s accuracy in exciting the relevant phosphor elements.
Colour scales and greyscales â€“ three test screens are chiefly used here, comprising a 16-step greyscale, a 64-step greyscale, with equivalent red, green, and blue scales, plus a 25-step scale for 10 principal colours. For each we assess how uniformly graduated the scales are, while making sure the specified number of steps is distinguishable. We also look out for changes in hue and tint, and whether the ends of the scales are compressed or expanded.
Background interference â€“ this smooth background is used to search for any fluctuations in intensity brought about by magnetic and electrical fields.
Dark screen â€“ Used to look for any spurious reflections and heightened glare from the face of the monitor.
Defocusing, blooming and halos â€“ This comprises of a series of dark grey, light grey, and white pixels, lines, and zeros on a black background. Used to see whether high intensities cause the beam to defocus, resulting in halos around pixels, and line edges and the centre of the zeros becoming broader and fuzzy.
Text focus â€“ short rows of text spread around the screen allow us to scrutinize each monitorâ€™s ability to render text in varying fonts on different coloured backgrounds. Also useful for inspecting colour convergence and anti-aliasing.
TFT Flat Panel Tests
All TFT flat panel tests were carried out at the native resolution of 1,280 x 1,024 at 60Hz, with a 32bit colour depth. Each unit was also set to its default colour temperature, and the contrast and brightness adjusted for optimised greyscale reproduction. However, as DisplayMate is geared towards testing CRTs, only a selection of its tests were employed. From those already described, the following are considered most appropriate. Text focus, colour scales and greyscales, colour purity and background interference.
If the flat panels were fitted with an analogue D-SUB port we also checked and adjusted the pixel phase and clock using DisplayMateâ€™s pixel-tracking and timing-lock test screen. If digital noise persisted it was noted. Our final technical test involved assessing the viewing angles and comparing them with those stated by the manufacturer. If these were much less than stated the scores were lowered, whereas units that demonstrated particularly impressive angles, both horizontally and vertically, gained higher ratings.
In addition, each flat panel was tested using real world scenarios. First up we looked at how well photographic images were rendered, paying specific attention to flesh tones and large areas of colour exhibiting slight variations in tone. We then rated DVD playback colour, brightness, contrast and detail by replaying the 16th chapter of The Matrix while also looking out for signs that response time was contributing to lag and motion smearing. Lastly but not least we took into account the aesthetics, build quality and functionality of each unit.