Digital camera: Examining sharpness and resolution

Digicam-Test 1

A section of the test image. Click on the image to open another view with the complete image. Determine the correct exposure by spot metering on the gray circles labeled "Spot".

The test image above serves as a target. Display it in 100% view on a monitor and photograph it full-frame with your digital camera. The result permits an estimation of sharpness and resolution capacity. By these means you can assess and compare the reproduction capabilities of different cameras and lenses or different camera or lens settings (zoom, aperture) in a rather objective manner.

The test image measures 1002 by 668 pixels, i.e. far below 1 megapixel. Normally with a full-frame shot it should be no problem, for example with a 6 megapixel camera, even to reproduce the partly only 1 pixel wide details sharply. More so, shouldn't the structures of the pixels themselves be visible in the photo? On the screen the test image is displayed by more or less 2 million colored LCD light dots. At least theoretically there is a lot of reserve up to 6 megapixels.

Practically though, 6 megapixels (and even higher resolutions) are not enough to reproduce the LCD light dots discreetly. The colored bands and dots in the figure below to the right are above all interpolation artifacts. These bands are considerably wider than the LCD light dots.


The test image in 100% view on an LCD monitor, one time a close-up was taken (left), the other is the according section of a full frame shot of the same image, taken with a Dynax 7D (right).

In order to display the test image without the display structure, download the TIFF version, open the image in an image processor or viewer and view it at 400% or more. The distance between camera and monitor (considering the distance at which the image in 100% view is exactly full frame on the camera's display) must be increased by the same factor. Presuming such an enlargement you can use the test image on CRTs, too. The larger the test image's pixels are in comparison to the monitor's light dots, the less is the influence by the monitor's own blurring and interpolation effects.

This way you can not only compare the sharpness and reproduction quality of different cameras and lenses, but also find out how zoom and aperture settings influence image sharpness. Very large and very small apertures create new blurring due to optical reasons, which you can see clearly on a test photo. With a range of shots featuring different apertures you can determine rapidly which aperture is the sharpest on your camera/lens combination.

Moreover, a full frame shot of the test image reveals possible lens distortions (barrel or pincushion distortions). Also, chromatic aberration and other defects in color can be judged. For that purpose though, you should print the test image on an A4 sheet of paper and photograph the print. A black-and-white print with a good laser printer is not only fully sufficient for that purpose, but even better than a color print.


You can also find this test in the book Fotobearbeitung mit Paint Shop Pro X (German only)

Download test image

digicam-test1.tif (180 KB)

Download the test image in TIF format with embedded Adobe RGB color profile.

Displaying the test image

Take notice that for displaying the test image on your monitor, the latter composes the image pixels out of colored light dots. This structure is possibly visible on the photo you take, and is of course not an element of the test image. It can be rendered almost invisible by viewing the test image enlarged. To eliminate these effects completely, you can of course also print the test image.

For monitor display I recommend an LCD display, because only these work pixel accurately if you show the image in a 100% (or a multiple of 100%) view. In a 100% view one LCD "pixel" - consisting of three light dots in the colors red, green and blue that lie in parallel - represent exactly one image pixel. The rasters of CRTs however never match exactly the pixel patterns, introducing additional interpolation effects.