Test Images

To aid the best viewing of the images on this site or indeed anything you might want to view on your computer, a few test images are provided.

Monitor Gamma

If your monitor and system colour settings are adjusted correctly, the image below should appear as a uniform greyscale wedge, white at the top, black at the bottom, and in the middle, half intensity (but much lighter than half brightness1). The image assumes a monitor gamma of 2.5.2

[Gamma 2.5 Test Image]

Adjust your monitor's brightness and contrast settings until the monitor is comfortable to look at and displays the image in shades of grey, with no hint of colour bars.

On a CRT monitor, the brightness control adjusts the black level (which should be no more and no less than black). Once that is done, the contrast control may be adjusted to brighten or dim the display for comfort without fear of spoiling the calibration.

On an LCD monitor, the brightness control adjusts the brightness of the backlighting and does exactly what the contrast control does on CRTs. While the brightness control may be freely adjusted without side-effects, the contrast control just makes a terrible mess of the image if it's at any but one particular setting. If your LCD monitor isn't too cheap and nasty, the test pattern should appear very close to what would be displayed on a good CRT.

You may find the larger gamma test image more pleasant to use.

PNG Transparency

The PNG specification's section on alpha channel processing requires partially transparent pixels to be blended to the background in the linear light domain. The following composite images assume the monitor gamma is 2.2, which is the default to use when the program displaying the image knows no better.

Each image is a link to a reference image for a display with a gamma of 2.2.

[B&W Fan Image]
[Colour Fan Image]
[Bars Image]

The left (reference) and right (test) halves of the fan images should appear mirrored but otherwise identical. The radial lines should appear smooth and uniform in width regardless of their orientation. There should be no colour tinges.

The coloured pattern reminiscent of frequency bars on television test patterns should appear as smooth white, cyan, magenta and yellow bar patterns on bands of black, red, green and blue, respectively. The top and bottom edges of each band are overlaid with a reference image. On a system on which a monitor gamma of 2.2 is assumed, the middle sections of each band should closely match the edge sections.

1The non-linear perception of brightness by humans is why two 100 Watt bulbs will not brighten a small room much more than a single 100 Watt bulb would, even though the two bulbs emit twice as much light as the one. It is also why chequerboard patterns of black and white have an average brightness much brighter than mid-grey.

2Real CRTs and LCDs contrived to emulate CRTs act if they have a gamma that is close to 2.5, but video recording standards impose a rendering intent of a net gamma of 1.14 by applying a gamma under-correction of 1/2.2 in the camera (which effectively darkens the midtones). This rendering intent caters for a display device traditionally capable of a limited maximum brightness and which must be viewed in a dimly lit room. The dilation of the viewer's pupils effectively scales the intensities of all parts of the displayed image by the same factor. Human perception of the illumination of the eye's retina is non-linear, though: The simple linear scaling of intensity effected by the pupil's dilation would cause the displayed image's midtones appear too bright compared to the image's highlights thus causing the image to appear washed out, if this effect wasn't countered by the deliberate under-correction of gamma applied at the camera. Perhaps in the future, it will be common to use cameras and displays with extremely high dynamic ranges3 to produce images that are intended to be viewed under normal lighting conditions. In that case, the end-to-end gamma may be standardised to 1.0.

3A well-designed EHDR television displaying an image of a class 4 laser beam, nevertheless, should not risk blinding the viewer. Likewise, an image of a nuclear detonation at close range should not cause the viewer to be instantly incinerated.