The output resolution of a visual stimulus generator is the number of distinct levels that can be used to reproduce the stimulus. On a colour CRT display this corresponds to the resolution with which the voltages to drive the red, green and blue guns can be created. (Laser printers are given greyscale capability using a method of super- pixels which is discussed elsewhere.) A typical personal computer graphics card will use one 8bit Digital-to-Analog Converters (DACs) for each of the three outputs resulting in a resolution of 1 part in 256 for each colour. Imagine using this resolution to represent a spatial waveform with a contrast of 0.5% - a not untypical requirement, and you will see that there aren’t very many luminance levels left to define it (one or two). The situation is made worse by the fact that CRTs have a non-linear transfer function between voltage and luminance (approximately x2) so that voltage steps near the top end of the voltage range produce a far greater change in luminance than those near the bottom. In a colour system, the situation is worse still as most of the luminance comes from the green gun with very little being contributed by the blue yet they have the same voltage resolution for both. A good visual stimulus generator therefore should have high resolution for the blue output, very high resolution for the red output and ultra resolution for the green output. Failing this though, high resolution for all three guns will do. Look for at least 12 to 15 bits per gun. Do not confuse these with true colour graphics systems that still only have 8 bit DACs on their outputs and remember that it is perfectly possible to generate high quality stimuli with only an 8 bit framebuffer provided that it has high resolution DACs at the output.

^ Back to top