Does your camera have BIOS (better instead of smaller) Pixel? You do not know? This test chart provides the opportunity to visually analyze the image quality of your digital camera.

Capture
To do this you take a picture of this test chart using your digital camera. Adjust the zoom position or the distance of your camera in a way that the gray background of the chart fills the complete image height. Depending on the aspect ration of your camera the borders will be cropped (also the logos and title will not be visible in the image). To avoid any motion blur you should use a tripod and the self timer of the camera.

Please ensure a uniform illumination of the chart and to take pictures that are comparable to the ones on the BIOS Pixel website that we took using several cameras, the illumination level should be app. 1/15 seconds at f-stop 5.6 and ISO 100 (equals 1/60 at f2.8; 1/30 at f4; 1/8 at f8). It may also be interesting to decrease the light level in order to see how the image quality varies at low light conditions.
Take a couple of pictures using different ISO speed settings (ISO 100, 400 800, 1600, 3200) and check the exposure using an eyedrop tool of your image processing software. The digital value of the gray background should be 120 (+/- 15). It is also important, that the white background is not clipped (digital value > 250). If necessary take another series of images and use the appropriate exposure compensation.

Evaluation
After you have taken the images you can visually check them. In order to do that you have to zoom into the displayed image to at least the 100% level. This means that one pixel in the image equals one pixel on the monitor. You can scroll through the image in order to see the different sections.

1. The gray patches are used to check the noise. Noise describes faults in an image that are more visible at high sensitivity levels (e.g. ISO 1600 or 3200) Usually noise is more visible in the mid and dark gray areas in comparison to black and the brighter regions.

2. The hyperbolic structures in the center of the chart are used to check the resolution of the camera. Until which number can you clearly distinguish the lines? The number x100 (e.g. 9 x 100 = 900 Line width/picture height) is the measure how many lines per picture height can be differentiated by the camera. (For the experts: the number has to be divided by 2 in order to get the often used line pairs per picture height). These days the resolution does not have much to do with the pixel count of a camera. You will clearly notice from these tests that typical consumer cameras with larger pixel count often provide worse image quality.

3. In addition it is of interest to look at different resolution and sharpness levels in the center and in the corners of the image. Does the resolution decrease in the corners significantly? Is the image blurred in the corners? Are all corners on the same quality level (optical decentering)?

4. There are cameras that actively reduce noise in the images. In many cases this does not affect the resolution in our black and white structure and decreases the noise level. But still the images often look strange somehow. The reason is the loss of low contrast texture in the image due to the noise reduction algorithms. Look at the 4 scenes (8 images) on the right side of the image that we have implemented at two different contrast levels. Can you still see the stones, the grass, the bricks, and the fabric or are the details all gone? Especially in this area (at ISO speed levels > 800) you will see the difference between a small consumer camera and a digital SLR.

5. You should also look at the colored patches (with the “abc” in it). Are the colors correct? Is the text sharp and clearly visible or do the colors look “washed out” due to the noise reduction?

This entry was posted Friday, 19. September 2008 at 08:13 and is filed under BIOS Pixel initiative Photokina 2008. You can follow any responses to this entry through the RSS 2.0 Feed. You can leave a response, or Trackback from your own site.