|
2.12 After-images
2.13 Stabilised vision
References
3. OBSERVER VARIABILITY
If the contrast of a stimulus against its surroundings is progressively reduced
there will naturally come a time when it can no longer be seen. However, this
‘threshold’ point will vary from observation to observation. If many observations
are made, and if the number of times that the threshold is above given values is
plotted as a percentage of the total number of observations, a cumulative
probability curve will result (see for instance Blackwell ). Such a curve is
conventionally known as a ‘frequency of seeing’ curve. One may equally expect
a form of frequency of seeing curve if presentation time is varied for a fixed
stimulus, if size or scene luminance is progressively varied for fixed contrast and
presentation time, or if a stimulus is caused to move at various rates during
presentation. However, the most common form of frequency of seeing curve to
be found in literature is, without a doubt, that as a function of stimulus
contrast. This is reasonable since contrast is, by definition, a measure of
differential energy, and hence should be related to concepts of signal to noise
ratio.
In studying threshold performance, then, it is possible to study the
probability of acquisition as a function of one or more basic properties of the
viewed scene, or to determine the trend of the 50% probability of acquisition as
a function of scene parameters. The majority of threshold studies to be discussed
|
