Chapter 4
Human Perception Overview

What This Chapter is About

Human perception involves numerous disciplines spanning art, graphics design, optics, physics, philosophy, psychology, physiology, psychophysiology, and biochemistry, among others. The complexity of human perception and our poor understanding of it prevents giving a complete treatment here.

Light Intensity Effects

When light intensity levels change, as with setting suns, our eyes become light adapted. Cones adapt more rapidly, adapting fully in about seven minutes. Rods can take up to ten times longer. At a given adaptation level, the sensors can differentiate from 12 to 24 different intensities. The neural processing circuits in the retina rapidly adjust the sensitivity and output of the cones and rods based on local luminance and contrast. Thus, the resulting working range of the eye is increased to between 100 and 200 intensity levels.

As light decreases, high photopic (cone) vision gives way to scotopic (rod) vision. At the crossover point, it is fair to say we have two different intermingled retina systems since we actually have both sensor types operating simultaneously. At the lowest light levels, only the rods operate and full night vision occurs below 0.01 to 0.005 lumens. While single quanta stimulate individual rods, we cannot see them though we can see as few as five or eight under optimum conditions.

Acuity is the ability to distinguish object details with a given contrast and luminance. Because cones provide acuity, are concentrated in fovea, and require relatively intense light to function, it follows that acuity decreases with decreased light intensity. As illumination levels decrease, all colors fade as the Purkinje shift occurs. Our color vision becomes more sensitive to blue wavelengths that seem brighter and less sensitive to reds. The spectral-response function actually shifts as Figure 4-1 illustrates.

Figure 4-1  The Purkinje shift

Things that were once blue become brighter than things that were once red. But there is little or no color, because this is due to a shift to rods that have only one spectral response function and hence cannot “see” color. Finally, increased luminance increases acuity also but decreases absolute luminosity difference detection.

Decreasing light levels cause the eye to increase its light integration time before firing optic nerve signals. This delay, called the retinal action time, can increase delays as much as one-tenth of a second. While light’s finite speed means we always sense historical data, this extra dark adaptation delay can be disastrous for night-time drivers. For example, if a car travels 60 miles per hour, that equals 88 feet per second. The one-tenth of a second delay equals 8.8 feet that may be dearly needed in a stopping emergency. The same applies to digital video game response times. In other words, effective video games likely provide high radiance images.

Adaptation is useful, effortlessly allowing us to visually discriminate under dramatically varying light conditions. For example, viewing photographs may involve surveying hundreds of input intensity levels. Since a given eye adaptation differentiates between 12 to 24 levels for the given level, rapid and unconscious processing allows eyes to shift to different adaptation levels and accommodate significantly more radiant energy levels, thereby increasing the eye’s working range. But, on the other hand, adaptation conceals radiant energy level perceptions essential for reliable camera operation.

Finally, note that our ability to distinguish between radiance levels depends on the total light present. For example, one flashlight’s intensity may seem noticeably stronger than another’s at night, but the intensity is likely inconsequential during the day.

Visual Input

Earlier discussion indicated spectral light has frequency. Our visual system assigns sensations to light frequencies and their combinations. We assign a given sensation a name that allows individuals to converse though they may actually experience different (but consistently different) sensations.

Visual Processing

The retina consists of brain cells that tether via the optic nerve to six billion visual processing cell counterparts located in the brain. Rods and cones gather input and pass it to neurons in the retina which integrate the input over time and across numerous retinal cell interconnections. The retina must eliminate possible random signal errors. It does this by increasing integration time and confirming signals from neighboring receptors.

The neurons in the retina combine, amplify, integrate, and differentiate the signals from the rods and cones. Chrominance (color) information is separated from luminance (black and white) information, although the same cells and fibers may encode both kinds of information with different codes. Retinal neurons also process the luminance information by exaggerating luminosity differences using a process referred to as sharpening which makes object edges more apparent. It then transmits the processed data to the brain’s recognition mechanism via the optic nerve, delivering significantly more luminance than chrominance information.

Perception Summary

Perception is the miraculous output from many slow, highly interconnected cells. It usually provides a predictable internal interpretation of an authentic external world we construct using information our brains previously accumulated and cataloged. While contemporary charge-coupled device sensors may rival retinal sensor densities, lack of accumulated knowledge will hinder their recognition capabilities for the foreseeable future.

Our eye-brain perception system is likely a probability engine that:

• Receives feeble, fragmentary input
• Eliminates some signals
• Operates as a differential input device that measures detected differences in color and luminosities off the achromatic axis
• Sifts through numerous alternatives searching for a best interpretation
• Presents the most plausible best-fit contender based on stored knowledge