Image Processing Reference
Visual Display Artifacts
If you try to drive the display too hard by setting a raster size and scan rate that is near the
performance limits of the display, the picture starts smearing in the horizontal axis and
also ringing when hard edges are drawn.
Smearing happens because the scan circuits simply cannot change the intensity of
the beam fast enough to cope with the resolution.
Ringing is seen as ghost copies of the image offset a quarter- or half-inch (depending
on the display size) to the right. These are due to the electronic effects of throwing a high-
frequency signal down a piece of connecting wire. At the very high pixel clock frequencies
we use now, a bad joint in a cable causes a reflection of the signal that bounces back and forth
to create the ghosting effect. The distance of each ghost image is related to the time taken for
each reflection to go back down the wire until it hits another obstacle. If you want to know
more about this topic, you'll need to study a physics or electronics book and seek out a
description of transmission line theorems. Be warned, though: It is a very dense subject.
Making visual quality checks and reducing pixel clock rates to more realistic levels
help to solve this problem. Using higher-quality cable and placing ferrite rings on the
cables to increase the inductance will also alleviate the effects.
There are other artifacts that affect the geometry of the display but these are largely
to do with the regulation of the scanning circuits in a CRT monitor. Using flat-screen dis-
plays based on one of the following technologies eliminates most of the linearity and
Liquid Crystal Display (LCD)
Light Emitting Diode (LED)
Organic Light Emitting Diode (OLED)
Digital Light Projection (DLP)
Dithering to Improve Image Quality
Since the 1970s we have been using dithering techniques to improve image quality in sit-
uations where the color palette is limited, and the technique was known before that in the
In the context of computer imaging, four techniques have been used widely. There
are others that are less popular variants of these four basic techniques.
This information is important to anyone who renders images that are computer gen-
erated and then displayed on a TV video system. Some weird artifacts appear if you try
and video capture newsprint or other dithered images on a rostrum camera. There are also
implications here for preprocessing images and dealing with image data in the decoded
output that yield harsh contouring artifacts. Adding a little noise to the image in order to
perturb the contoured edges disguises the edges. That unfortunately makes the compres-
sion algorithms slightly less efficient. A little noise applied carefully helps to make the