Alliance for Lighting InformationAlliance for Lighting InformationHuman visual systems work differently under different levels of stimulation - and of course differently from each other as well. But certain trends are generally recognized.
The retinal photoreceptors are commonly divided into two types - rods and cones. Cones come in three "flavors" - red, green and blue so to speak - and fill the central part of the retina called the fovea. With these photoreceptors working together, the visual system - including the brain - creates perceptions of the world in color, by combining the responses of the three different types of cones. This works because of the multiple layers of groupings of photoreceptors, as individual cones directly contribute (positively or negatively) to several nerve cells or receptive fields, which may in turn contribute to other groupings.
All rods are the same and fill the periphery of the eye, and are especially sensitive to motion, parallax, orientation and other visual phenomena. Since there is only one type of rod photoreceptors, the signal from rods is always black and white and gray - luminance contrast but no color contrast.
The location of the rods and cones differ - the rods are on the periphery with a few cones mixed in, and the cones are alone in the center, and there is some mixing outside the fovea. If you are "looking at" something (up to around 2 degrees of visual angle) it is imaged by the visual system onto the fovea and is seen solely with cones. There are no rods in the fovea, but there are many densely packed around it and they become first more dense and then less dense as the distance off-axis increases.
Since the cones have three different types, the resolution with cones is always a bit less than with rods also. This occurs despite the fact that the foveal region is densely packed with cones. As the off-axis distance increases, the grouping of photoreceptors increases, so the sensitivity to motion or orientation increases.
Adaptation - the response to changing levels of stimulation, which for photoreceptors means light - also differs for rods and cones. Since the cones have three separate "channels" the overall sensitivity is lower, and the rods are much more sensitive than cones at lower levels of light. Because of the nature of human visual response, these levels are typically measured in terms of luminance (candela per unit area), but on a logarithmic scale. This means that differences of 3:1 or 1:1/3 are the smallest discernible steps, despite the seemingly large changes in the associated luminance values.
Photopic is the term used to describe the adaptation range where cones are completely functioning. By definition this corresponds to complete color vision, which is feasible under the higher adaptation levels corresponding to around 3.4 cd/m2 and up. Note that the amazing human visual system can "handle" luminances up to 100 to 1000 times this level, although we are bothered by such very high luminances even if they are uniform. (If luminances are very non-uniform, this is defined as high contrast, and may become glare. This may effect adaptation levels, but is in general a different topic, since the contrast which is "glare" means that the high luminance must be significantly different from the adaptation luminance.)
Scotopic is the term used to describe the adaptation range where cones are not functioning and rods are the only working photoreceptors. By definition this corresponds to black, white and gray vision, with luminance contrast but no color contrast, because there is no color without cone perception. Furthermore foveal vision is not working at all, and for details to be discerned, they must be located off-axis. This is why to see the stars clearly at night, it is better to "look away" from them slightly.
Mesopic is the intermediate range, between photopic at the higher luminance levels and scotopic at the lower levels. By definition, across the mesopic range both cones and rods are working, but the contribution of the cones varies, decreasing as the levels approach the mesopic-scotopic boundary.
Where is this boundary? The Illuminating Engineering Society of North America (or "IESNA, The Lighting Authority") has several publications which define the boundary. The definitions differ depending on where you look.
In older texts, such as the IESNA Lighting Handbook Reference Volume, 1984, the mesopic range is defined as "between about 3.4 and 0.034 candelas per square meter" and this is the same definition used in the IESNA Lighting Fundamentals Course 1999 edition. However, the IESNA Lighting Handbook, 8th edition (2000) reports several different values for the boundary between mesopic and scotopic. In the discussion on Luminous Flux and the Lumen in Chapter 1, the text discusses "mesopic ... between 0.001 and 3 cd/m2" which is 1/30th the previous value for the boundary. This value is reported again in the definitions of the terms in Chapter 3. In between these two statements, the Chapter 2 Introduction refers to "intermediate luminance levels, between approximately 0.01 and 3 cd/m2 - the mesopic region", which brings up the question of which order of magnitude should we be considering as the boundary. Even when the values are compared in logarithmic scale - as they should be - the difference between 0.001 and 0.01 is substantial - and what ever happened to 0.034 c/m2?
The discussion under Exterior Lighting refers to "scotopic conditions, below approximately 0.01 lx" which presumes some reflectance to convert illuminace to luminance. Making use of the assumption of a "perfectly diffuse reflector" or PDR leads to the reflectance being equal to the luminance times pi divided by the illuminance. With this equation, the boundary value of 0.01 lx must be falling on a surface with 31% reflectance to produce 0.001 cd/m2 - which is the lowest boundary value previously mentioned (by comparison, the widely applied reflectance value developed by Kodak is 18% for exterior surfaces in general.) Obviously the 0.01 lx value does not allow for a boundary above 0.003 cd/m2 - 1/10th the value in the older texts.
This contradiction in IESNA definitions in the Lighting Handbook, 8th edition, may be related the previous publication RP-33-99, Lighting for Exterior Environments. This document also cites the value of 0.001 cd/m2 for the boundary between mesopic and scotopic. Inspection of the references indicates that this change in the definition of the mesopic - scotopic boundary probably comes from a 1995 CIE publication by T. McGowan and M. Rea, or a 1996 LD+A essay by M. Rea.
Apparently somewhere between 1984 and say 1996, the boundary between mesopic and scotopic was decreased by a factor of 30.
The reason the boundary between scotopic and mesopic matters is that this is the point where cones no longer contribute at all to visual perception. This means that there is a significance in how the light level - set by the criteria - interact with the visual system.
For example, the roadway lighting Recommended Practice, ANSI/IESNA RP-8-00 (and its previous versions) call for roadway lighting levels to have a maintained average luminance of at least 0.3 cd/m2 for the very lowest levels, which correspond to local roadways with low pedestrian (previously residential) classification. This criterion can be exceeded and still meet the practice, and for all but the moment before maintenance occurs, average luminance will be higher. This means the asphalt roadway surface should be - at the very darkest - about 1/10th the level of the top boundary of the mesopic range. This is two "brightness steps" of 3-to-1 in logarithmic terms - and 300 times the value of the lower boundary in its recent definition.
The roadway surface is characterised by the r-tables in the RP-8-00 and the typical surface is considered to be R3, which has an overall associated reflectance of 7%. Comparing this with other exterior surfaces' relectances shows that the roadway is typically darker than surrounding surfaces. Although detailled information on the reflectance of exterior surfaces is limited, the Kodak value of 18% has been considered accurate. Under the same illuminance (which is a bit of an assumption), this reflectance would correspond to an off-axis luminance level of 2.5 times that of the roadway average, or about 1/4th the top of the mesopic range and over 700 times the level of the lower boundary. Note that this is for the very darkest allowable level on the darkest roadway - busier roads have criteria up to 1.2 cd/m2 (minimum value for the maintained average luminance) - and does not include the contribution made by headlights at all. Obviously, properly designed roadway lighting systems provide mesopic levels, but rather near the edge of the photopic range, when headlights are included. This is recognizable by the relatively good color vision which results from the combined illumination of properly designed roadway lighting systems and headlights.
In comparison, the criteria for walkways and parking lots are significantly lower lighting levels, down to 2 lx minimum for parking lots (IESNA RP-20-98) and 2 lx average for residential walkways (IESNA DG-5-94). Using the same assumption of PDR and the equation discussed above, 2 lx with a reflectance of 18% corresponds to over 0.1 cd/m2 - about 1/30th of the top boundary and 100 times the lower one.
The following diagrams show the mesopic range from 0.001 to over 3.4 cd/m2 on a logarithmic scale, and the luminances corresponding to three different surfaces. The surfaces are roadway asphalt (7% reflectance), the surroundings (estimated at 18% reflectance per the "Kodak outdoor reflectance" and a maximum reflectance surface (estimated at 50% for exterior settings). With these three reflectance values (and some assumptions), the same illuminance will produce a luminance 7 times as high from the max. as from the road.
The first diagram shows the three luminances associated with full moonlight, with the darkest location on the darkest class of road, and the brightest spot on the brightest class of road, according to the IESNA RP-8-00 national standard.
The second diagram shows the three luminances associated with full moonlight, with the darkest location on the darkest class of parking lot, and the brightest spot on the brightest class of parking lot, according to the IESNA RP-20- 96 recommended practice.
The third diagram shows the three luminances associated with full moonlight, with the darkest location on the darkest class of a walkway, and the brightest spot on the brightest class of walkway, according to the IESNA DG-5-94 design guide.
Therefore properly designed systems (except for residential walkways ) will provide lighting levels which are - at their very lowest - in the upper-middle of the (new) mesopic range, and never approach the lower boundary between mesopic and scotopic vision.
This site is provided as a public service by resodance publishing co. and no guarantee is implied or provided.
This page is served & maintained by &
last changed on 24 Sep 04 by
