1. Can we know how an object (or space) is perceived if we have full knowledge of its physical characteristics?

The phrase ‘we have full knowledge of the object’s physical characteristics’ implies that we have some kind
of record of the object’s characteristics either in memory or in some kind of retrieval device. Then adding the
phrase ‘we can know how it is perceived’ could mean that we can know how the perception process works
when we come across this object at some point in the future, which is hardly likely, or it could mean that when
we come across that object at some point in the future then we know how it will appear to us.

Experience shows us that having this information does not always tell us how the object will appear, it will only
tell us how it should appear. There are many situations we come across where the object doesn’t appear as
expected, when the difference seems contradictory with our expectations then we call it an illusion. The good
news is sometimes the process of trying to shed light on these illusions leads us to a better understanding of
our perceptual system and / or designing spaces or objects that either exploit our perceptual shortcomings or
overcome them.

Take the Muler-Lyer illusion when we look at the two lines together with
the attached ‘arrows’, we perceive that one is longer than the other and
we don’t believe that they are equal until we measure them. In addition,
once we have having the knowledge that they are equal we still see
them as unequal in length as a result we are forced to admit that it is an
illusion and that our perceptual system is somehow at fault, we might
even try to understand how that might have come about.

Unfortunately, knowing and understanding the underlying mechanisms

that cause this illusion would not alter our perception, we would still
see one line longer than the other!
However it is comforting to know that at least having full knowledge of

The Muler-Lyer illusion: we still see one line longer than the other even
after we measure and confirm that they are equal in length. (p 249 in [2])

an object’s characteristics means that we can always tell when we are perceiving an illusion. That is we know
when we don’t have a perfect fit between perception and reality (‘full knowledge’ is assumed to mean reality).

There are many reasons why illusions occur, it trying to understand them we have come to realise that during
the perception process the brain sometimes uses perceptual cues or even fabricates information that seems to
be missing. There are also situations where illusions arise from within the object itself, as we will see below.

For instance since we know that there are perceptual cues that the
brain uses to ‘interpret’ a situation then we can demonstrate that
knowledge by manipulating the cues in such a way that creates the
illusion or we can do what the Greeks did (in the architectural design
of their temples) and manipulate the actual object to make it appear
different from what it actually is because we prefer it that way.

For example knowing that in our perceptual system parallel lines

extending out from an observer appear to converge, even though they
really don’t, and that when two objects are of equal size then the one
that is further will take up less of our field of view than the one that is
closer even though its not really smaller would enable a designer to
create a situation where they can predict that a small object will appear
larger than it really is when it is strategically placed on converging lines.
That is, the designer simulates depth when none exists and can predict
that an object will be seen as bigger than what it really is and indeed it
does.
The Ponzo Illusion. (p251 in [2])
This particular kind of illusion (that uses converging lines) is known
as the Ponzo illusion. Here both animals are the same size, hence
they have the same visual angle but the one on top appears longer because of the fabricated perception of
depth generated by the converging railroad tracks that make the top animal seem further away and so if this
was a real situation it should have a smaller visual angle, but since it doesn’t, the visual system is tricked.
So an illusion taught us something about our perceptual system &/or its deliberate creation allowed us to
demonstrate our awareness of our perceptual process.

The architects of ancient Greece were aware of many geometrical optical illusions and worked out details in
their architecture for counteracting them, although they had no knowledge or understanding of the underlying
physiological visual systems.

The Greeks knew, by observation, that the long lines of the architrave, appears to sag if it was actually straight,
and so they built them convex in order to make them appear straight. They did the same thing to stylobate.

For example in the Parthenon the Greeks did the following to compensate for optical illusions:
• the stylobate has an upward curvature of more than 100mm on the sides of the edifice and more than
60mm on the east and west fronts.
• Vertical features were made to incline inward in order to correct the appearance of leaning outwards at the
top.
• The axes of the columns are not vertical, but are inclined inwards nearly 70mm.

This is demonstrated in the following three diagrams:

The front of a temple as it should
appear. (p197-198 in [3])
The (exaggerated) appearance
if the temple was built without
compensations for optical
illusions
The built temple showing
the (exaggerated) physical
corrections in order that it
would be perceived as the temple
on the far left.

The Greeks also noticed that tall columns that are built straight tend to appear thinner in the middle, so they
built them slightly fatter in the middle in order to make them appear straight. This outward curvature in the
profile of columns is called entasis. In the Parthenon columns, which is 10.3m tall, they increased the column
thickness by 20mm to make it appear straight.

Another example of when inconsistency arises between how an object

is perceived and what we know of its characteristics is when the light
source is hidden and it comes from a direction other than from above.
This is because the brain uses information about patterns of shadows to
tell us about the shape of solid objects, so that if light comes from above
an ingoing dent would appear bright in its lower part and shaded at the
top, while the top part of an out going protrusion will be bright and the
lower part will be in shadow. It is obvious that if the direction of the
light changes then the pattern of light and shade will also change.

Knowing the direction of the light will lead to a perception that is

consistent with our knowledge of the physical properties of the object.
However, If there is no information about where the light is coming
Light is assumed to come from
from then the brain will assume that light is coming from above and if it
above, even when it is not
happens to be coming from below then the appearance of the object will
affecting our perception of form
still appear as though light is coming from above and thus be different (p331 in [1])
from our full knowledge of it, that is the dents will appear as bumps and the bumps as dents, when we know it
is in fact the other way around.

The last example demonstrates that the brain fills in gaps by fabricating information where information is
absent and as a result it sometimes makes mistakes. In this case when the direction of the light source is not
known the brain assumes that it is coming from above regardless of what we know of the object!

Knowing the full characteristics of an object and discovering an illusion can and does compel (some of) us to
search for physiological and / or perceptual explanations, as is the case with the Hermann grid, Mach bands,
simultaneous contrast illusions and many more.

The Herman grid illusion, gray Mach bands are light and Simultaneous contrast illusion,
images appear at intersections dark bands that appear at the two centre squares appear
but they are not really there. boundaries but they are not really different yet they reflect the same
(pp62-66 in [2]) there. amount of light into the retina

Goldstein offers an explanation for these illusions (pp 62-66 in [2]) in terms of lateral inhibition that occurs at
neural synapses. However he also says that the illusions disappear when we look at the affected areas directly
and that this must be explained by some other mechanism that he does not specify. In addition he also provided
an example, White’s illusion, that contradicts the lateral inhibition explanation.

The arrows indicate the amount of lateral inhibitions

White’s illusion. The rectangles at A and B appear
different, even though they are the same. (p67 in [2])
received by parts of rectangles A and B. B receives
more lateral inhibition than A because it is
surrounded by more white. This would predict that
B should appear darker than A (as is the case with
simultaneous contrast) but the opposite happens.
Therefore the lateral inhibition ‘theory’ can not
explain White’s illusion.

Therefore, in this case it seems that although we have full knowledge of the object and we can see the illusion
we can not yet explain the perceptual process that leads to this error in judgement.

The last example that we consider is the colour after effects illusion. This occurs when the retina is exposed to
one colour over a certain period of time it will then see its compliment where none exists. This illusion led to
the proposition that there are opponent processes at work in the brain and soon after we were able to discover
neurons that respond in opposite ways to blue and yellow and to red and green. It is argued (p217 in [2] that
these opponent processes provide a way of specifying wavelengths that may be clearer and more efficient than
the ratio of the S, M and L cone receptor responses’ alone. But even though we understand how we arrive at a
wrong perception we still get illusions of after effects that is, the system still has its short comings!
The brain seems to be pretty set in its ways. It uses certain perceptual cues in certain fixed ways that sometimes
lead to errors in perception, it makes assumptions and fills in information (that is not always correct) about
the environment when the information is missing and its mechanisms fatigue and require regeneration leading
to compensatory mechanisms that can lead to a faltering perception. The fact is even though we have full
knowledge of the object’s characteristics we have varying degrees of knowledge of the context that we will
find the object in and our brain has its limitations in capturing and interpreting its sensory inputs thus our
perception can be compromised to different extents depending on the situation.

Therefore, we are not able to know how we will perceive an object or a space with absolute certainty even if
we have a full knowledge of its physical characteristics. However, with all our internal shortcomings and with
all the external variations, we somehow get more things close enough to right to enable our species to survive,
maybe too right as we are over running the earth but maybe not right enough to maintain that which sustains
us, it seems. We even have time to sit back and wonder about how we do it all and how we might do it better,
which is all pretty incredible.
2. Central and Peripheral Vision with applications to architectural design:

Light energy that enters the eyes falls onto the retina where it activates receptors that convert this energy
into electrical energy that then travels to different parts of the brain resulting in visual perception. Central
and peripheral vision correspond to what happens physiologically and perceptually to the light energy that
falls inside and outside the fovea, respectively. Visual acuity and colour vision seem to be associated with
central vision while night vision, differences in illumination levels and motion perception are associated with
peripheral vision.

The visual receptors on the retina are either rods or cones. Since all receptors on the fovea are cones, it is safe to
assume that cones dominate the initial stage of central vision. The great majority (97%) of the receptors outside
the fovea are rods and so it is likely that they have a major impact on peripheral vision. However, we also note
that 99% of all cones on the retina are also spread outside the fovea, but as their numbers quickly drop off with
increasing distance from the fovea we assume they operate at the boundary between peripheral and central
vision, so the distinction between central and peripheral vision is a continuous transition rather than a sudden
discrete jump between them.

When cones in the fovea and hence in central vision are destroyed as with a condition called macular
degeneration, a blind spot is created in the direct line of vision. However, when rod receptors are destroyed, as
in the case of retinitis pigmentosa, peripheral vision is gradually destroyed and total blindness results when the
cone receptors in the fovea are also destroyed. This seems to support the assumption that cones and rods drive
central and peripheral vision, respectively.

When receptors convert light energy into electrical energy, the electrical signals travel to ganglion cells via
bipolar cells. As it turns out, there are far more receptors than ganglion cells, in fact the ratio of receptors
to ganglion cells is 126:1. Therefore each ganglion cell is connected to one or more receptors. According
to Goldstein (p58 in [2]) many of the foveal cones have one cone per ganglion cell while in the peripheral
retina there are many rods per ganglion cell. This structure between receptors and ganglion cells is called
convergence. We will use this notion of convergence in explaining differences between central and peripheral
vision in the context of brightness perception and visual acuity.

The ganglion cells can be further classified as parvo cells or magno cells. Parvo cells have the smaller cell body
and shorter denser branching of their dendrites. They are also far more numerous than magno cells. The magno
cells have opposite characteristics to parvo cells. They have a large cell body and longer, sparser branching
of their dendrites. There are virtually no magno cells in the fovea and their number increases towards the
peripheral retina. This observation together with the above regarding convergence and the retinal distribution
of receptors suggest that foveal cones converge with parvo cells, thus operating during central vision while rods
converge with magno cells hence operating during peripheral vision.

Ganglion cells then gather together to form the optic nerves and exit from the eye at the blind spot. The two
optic nerves (one from each eye) meet at the optic chiasm where they cross to opposite sides of the head. From
the optic chiasm two optic tracts are formed some lead to the thalamus in the mid brain where they terminate
in the lateral geniculate nucleus (LGN) the others branch off and go to the brain stem. Corresponding to the
parvo and magno ganglion cells in the retina are corresponding parvo and magno cells in the LGN. While the
brain stem contains projection of magno cells.

The two different pathways to the visual cortex (one via the LGN and the other via the tectum, in the brain
stem) seem to serve different perceptual functions that are also somewhat related to central and peripheral
vision. The former pathway, via the LGN, is involved in the fine grained perception of patterns and colours and is
related to central vision while the later, via the tectum, seem to co-ordinate the localisation of objects in space,
the guidance of eye movement and gross pattern perceptions and is related to peripheral vision; with some
overlap between the two processes. The LGN is also involved in peripheral vision during scotopic conditions.

We now consider brightness, visual acuity, colour and motion in more detail with respect to central and
peripheral vision.

Brightness is the perceptual impression of the amount of light that is being emitted from a source or reflected
from a surface. Dark adaptation experiments indicate that ‘two separate physiological mechanisms are involved
in the perception of brightness: the cone system (in central vision) for brighter illumination and the rod system
(in peripheral vision) for dimmer illumination’ (p109 in [1]). This perceptual difference can be somewhat
explained physiologically in terms of the difference in convergence (synaptic connections) between the
receptors and the ganglion cells in the retina.

The many to one connection between rods and ganglion cells are more likely to fire under dim light conditions
(because they can receive signals from any one or more rods) than ganglion cells connected to cones (because
the only way it will receive a signal is when that particular cone fires). Thus peripheral vision is more sensitive
to relatively dark illuminations than central vision.

There is also some physiological evidence suggesting that ‘when bright light is present, and the cones are active,
they inhibit or turn off the action of the rods.’ Suggesting that central vision is better than peripheral vision in
bright light.

The ganglion parvo cells have low contrast sensitivity while the magno cells have high contrast sensitivity.
Following the pathway of electrical signals to the LGN in the mid brain we find that the LGN’s magno cells are
more sensitive than parvo cells to the magnitude of the change in luminance at edges. Assuming a connection
exists between parvo cells and central vision and between magno cells and peripheral vision, these differences
also indicate that peripheral vision is superior at detecting differences in illumination levels than central vision.

An application of this distinction between central and peripheral vision in relation to brightness perception is
in restaurant lighting design. Suppose we are designing the lighting for a large open restaurant for dinner time
and suppose that our objective is to give the dinners the feeling of privacy and intimacy. If the restaurant is
dimly lit throughout then having better peripheral vision than central vision would mean that the diners will be
more aware of their neighbours and surroundings than each other and their dinner. They may not even be able
to read the menu, since that would involve using their compromised central vision.

On the other hand, if the restaurant is brightly lit, then peripheral vision will be compromised and their central
vision will be working perfectly. But then every time their central vision shifts slightly from each other or the
food they will clearly see their surroundings and their neighbours and that would take away their feelings of
intimacy and privacy,

So the ideal case scenario is to have relatively low general illumination but sufficient light directed at each
table that is enough to activate the cone receptors of the diners at that table. Furthermore, since illumination
decreases with the square of the distance from the source, we can also make sure that the tables are sufficiently
far apart so as to not receive much light from neighbouring sources, but enough light to be able to move
between them. This way, as the central vision shifts from the diners companion or food to the surrounds they
will see very little of their neighbours and when it shifts back to look at each other or the food they will see each
other well. Although peripheral vision works relatively well in low illumination, we will soon see that is has
very poor visual acuity and so the diners although aware of their surroundings through their peripheral vision
they won’t see much detail and hence the objectiveof privacy and intimacy will be achieved.

Visual acuity or the ability to see detail, is far greater in central vision than it is in peripheral vision. As with
brightness vision, this too can be explained in terms of differences in convergence between receptors and
ganglion cells in the retina.

In this case, two different light patterns that fall on the fovea, i.e. in central vision, will activate correspondingly
two different combinations of ganglion cells due to the one-to-one connections between foveal cones and
ganglion cells and so they will be perceived as different. On the other hand the many-to-one connections
between the rods and ganglion cells in peripheral vision mean that there are potentially different stimuli that
could result in the same combination of ganglions firing. Therefore, the combination of firing ganglion cells for
each stimuli may not be sufficiently different for detection. Hence the lower visual acuity in peripheral vision
than central vision.

This higher acuity in central vision than in peripheral vision also translates into a higher ability to see texture
and pattern in central vision than in peripheral vision.
The difference in visual acuity between central and peripheral vision clearly applies to the above restaurant
example. Another example is to suppose we have a large open factory where attention to detail is necessary
for both safety and performance. One approach would be to brightly illuminate the whole space, but a better
approach from both an environmental and economic perspective would be to use relatively bright task
lighting combined with general medium ambient lighting. Thus focusing the light where it is mostly needed
and lowering the overall consumption of electricity. Shielding the task light would also reduce glare when it is
viewed from a distance.

Colour vision corresponds to our ability to perceive different wavelengths of light as different colours. Overall
the rods and cones have pigments that differ in their absorption of wavelengths of light. In addition to that there
are three different cone pigments, denoted S, M and L each having a different absorption spectrum but only
one rod pigment with its own absorption spectrum. The absorption of the rod’s visual pigment closely matches
the rod’s spectral sensitivity curve (in scotopic conditions where the rods are most sensitive) while the S, M an
L cone pigments add together to result in a spectral sensitivity curve similar to that of the cones (in photopic
conditions).

Each wavelength of light will produce a ratio corresponding to the amount of light absorbed by each of the
S, M and L pigments. It is this ratio that allows us to detect (and match) different wavelengths of light that
correspond to different colours. In photopic conditions the cones are active and colour is perceived in central
vision, while in peripheral vision, the rods are too saturated to contribute to colour perception. In scotopic
conditions, the cones are not working and only the rods are active. Therefore in the absence of a ratio between
pigments, different colours can not be perceived, only the blue-green region of the spectrum can be detected,
where the rods are most sensitive.

Ganglion parvo cells are colour sensitive while magno cells are colour blind. and further down the track we find
that the LGN’s parvo cells respond to differences in wavelengths of light, while the mango cells respond in the
same way to all wavelengths of light. Thus supporting the idea that central vision is better at colour perception
that peripheral vision provided of course that parvo and magno cells tell us about central and peripheral vision
respectively. However, the magno system is not completely insensitive to all aspects of colour, perhaps this has
something to do with connections between it and the cones in the peripheral retina.

Continuing on to the cortex there is evidence that support the notion that ‘hue and brightness information are
processed in separate places in the brain ... (however) the wavelengths of light that are present are not the only
factors that determine our perception of hue. A number of factors, such as stimulus intensity and duration as
well as the characteristics of surrounding stimuli, can also alter the perceived colour’. (p169 in [1]).

Suffice it to say colour perception and related physiological processes are far from simple, but broadly speaking
we can say that colour vision is far better in central vision than it is in peripheral vision during photopic
conditions and perceiving differences in illumination are better using peripheral vision than central vision in
scotopic conditions.

Our ability to differentiate between colour in central vision than in peripheral vision can be combined with
our increased visual acuity in the design of spaces. For instance suppose an architect wishes to highlight
certain sculptural features in a building’s structure then choosing a contrasting colour to the background could
achieve that objective when one is viewing the sculpture using their central vision. Or in situations colour
coding adjacent pipes that transport different materials (such as water, air, electrical wiring, gas) in a building
could facilitate identifying them quickly for maintenance. An example where this is taken to extreme is on the
facade of the Pompidou Centre in Paris where the enormous coloured utility pipes are also a form of artistic
expression. Also, as we are more sensitive to blue/green light in scotopic conditions through peripheral vision it
would make sense to colour areas that need to be highlighted in the dark in blue, such as having exit lights in a
blue green colour, that can also be found quickly in an emergency situation.

Motion perception also begins at the retina. Signals from the receptors reach the optic tract and find their way
towards the brain stem. The visual centre in the brain stem is known as the tectum. The part of the tectum
that receives most of the incoming fibres from the optic tracts is an area on the back surface of the brain stem,
known as the superior colliculi. The vast majority of cells whose axons make up this pathway appear to be
projections of magno ganglion cells. The front portion of the cells in the upper layers of the superior colliculi
represent the central visual field and the back region represent the visual periphery. This back region is ‘not
very sensitive to details of shape (such as orientation) or to colour, but they are quite sensitive to motion and
location.’ (p89 in [1]), which is consistent with their magno cellular input.

The superior colliculi also receives inputs in the form of projections from the primary visual area of the cortex
and from an area called the medial temporal region thought to be a centre for visual motion processing. ‘Both of
these back projections are primarily of the magnocellular type, thus contributing to the view that the superior
colliculi are important in the analysis of movement and location. From the superior colliculi the pathway
continues to the thalamus and eventually to the secondary visual areas of the cortex.

Being able to detect motion through peripheral vision than through central vision can be exploited where we
need to see moving objects through our peripheral vision. This is particularly important when driving, as we
need to see moving vehicles around us. This is especially the case when they put their turning indicator light on
in order to turn their vehicle in front of us. Having a moving flashing light instead of a sustained light markedly
improves our ability to see the turning vehicle. It may also help in industrial situations. A worker whose central
vision is focussed in front of them can see movements in components of a machine beside them better if it is
moving. So movement could indicate when something is ready to be attended to.

The above discussion demonstrates that there are differences between central and peripheral vision and
that their are underlying physiological differences that relate to perceptual differences, yet the physiological
differences are quite complex, they overlap and not fully understood. But knowing these differences can help in
the lighting design of three-dimensional spaces that are potentially safe and aesthetically pleasing.
3. Appearance of objects in daylight and at night and how this might affect design:

The visual characteristics of an object can be measured, using various instruments, and recorded in terms of
dimensions, distance from the observer, reflectance and luminance. When these characteristics are reported
by an observer without the aid of measuring instruments but by using either words or even presenting the
appearance of light in paintings as artists try to do, then they are presenting the visual characteristics of objects
subjectively. In this case we are talking about the appearance of the object. For instance, colour corresponds to
the subjective experience of wavelength, brightness is the subjective experience of luminance or illuminance
and lightness is the subjective experience of reflectance.

First we consider the difference in appearance of objects in outdoor daylight (when objects are illuminated
with sunlight and /or skylight) with how they appear in outdoor night light (when objects are illuminated with
star light &/or moonlight). But as humans, we also spend a great deal of our time indoors and we also invented
electric lighting to try and ‘extend’ daylight hours. However, daylight appears different indoors from outdoors
and we haven’t been able to mimic daylight using electric lighting, mainly because of the dynamic nature of
daylight and its wavelength distribution differs from that of electric lighting. Therefore, in considering lighting
design we also need to consider these additional conditions or limitations of enclosed spaces and electric
lighting with respect to how objects might appear under those conditions.

Light can vary along three dimensions: intensity, wavelength and duration. The main difference between
daylight and night light is intensity, where the intensity of daylight is much higher.

In daylight objects that are relatively near appear in focus and in colour. Light at the long wavelength end of
the spectrum (which we report as reds) appears brighter than the light at the short wavelength end (which
we report as blues to greens). In the night light the colour sensation is lost, that is colours appear de-saturated
and the the apparent brightness is greater at the blue/green end than at the red end. In fact red appears almost
black in night light. Detail is also lost in night light and so is texture. This is because:
• the cone receptors that we use in the day time have a much higher threshold to light intensity than the rods
and so we rely on rod vision at night and cone vision in the day time.
• rods have only one visual pigment, rhodopsin and so we can’t see colour in night light
• there is a shift in sensitivity (called the Purkinje shift) towards the blue end of the spectrum when we
change from cone vision to rod vision. This shift means there is a point at the red end of the spectrum
beyond which the rods are blind.
• daylight cone vision is central vision where we can see detail, these cones are not sensitive in night light so
we can’t see detail at night.
Thus night light objects appear blurry and colourless.

Interestingly, objects that have familiar colour are seen to have a hint of that colour in night light. For instance
a banana would appear a little yellow, while an unfamiliar object with the same colour as the banana would
appear totally de-saturated. This phenomena is referred to as memory colour, it is an example of a situation
where prior knowledge affects our perception. (p220 in [2]).

Shadows result from the blockage of light from a directional source. Shadows helps us to perceive objects in
3D. In daylight shadows help us to see texture and form. When light falls on a textured surface at an angle the
texture is more easily apparent. In bright sunlight shadows often take on a bluish colour, this is due to ‘physics
and visual coding’. It is because ‘whenever an object is lit by two sources from different directions and of
different colours, the cast shadows will have colour because they are illuminated by the source that is not being
blocked’. As Da Vinci explained the appearance of a white object that is partly in the sun ‘the part not facing
the sun stays in the shadow and partakes (only) of the colour of the air (which is blue skylight).’ (p47 in [4].
This property of coloured shadows has been heavily exploited in impressionist paintings. Night light on the
other hand is not intense enough to result in easily perceivable shadows, so objects appear as flat grey to black
silhouettes, thus we can not see 3D forms and texture in night light.

With increasing distance objects in daylight appear less sharp and often have a slight blue tint. This is referred
to as atmospheric perspective. The reduced acuity is because with increasing distance there is basically more
stuff (air and particles) in the way between the eyes and the viewed object. Thus occluding the detail and
reducing the amount of light reaching the retina. Distant objects can appear bluish because of the way the
particles scatter the light waves. In night light distant objects appear as colourless flat silhouettes, simply
because we are using rods in night light, so no colour and no detail is present. Also in daylight we can see
texture for some distance, however, ‘elements that are equally spaced in a scene appear to be more closely
packed as distance increases’. This is because more distant objects take up less of our field of view. The notion
of a texture gradient at night doesn’t even apply. In order for us to optimally recognise shapes, faces, depth and
movement and these differences we need luminance (brightness) differences, but not so great as to produce
glare. At night the brightness differences are very small and so vision is hugely compromised without the aid of
artificial light at night.

Provided there is enough light for cone vision to operate, the colour appearance of objects remains constant
under wide variations in illumination conditions that ranges from direct sunlight, to spaces in the shade to
indoor spaces illuminated with electric light. This is the case even though sunlight is 1000s of times brighter
than indoor light and electric light has a different wavelength distribution to daylight. This phenomena is
referred to as colour constancy and there are a number of theories that attempt to explain it. They include
chromatic adaptation, ratio principles and the presence of cues, (pp220 - 224 in [2]).

For instance, in a space illuminated by tungsten light, according to the adaptation principal, the eyes become
adapted (or fatigued) to the long wavelength (because a tungsten light is a ‘long-wavelength-rich’ light) and so
becomes less sensitive to it. Apparently, this relatively ‘excessive’ red wavelength that is reflected by the object
has less effect on its appearance following adaptation. On the other hand the appearance of achromatic colours
such as white, grey and black also remains constant under different illumination conditions. This is explained
in terms of percentage of light that is reflected from an object, hence it is explained in terms of proportions and
ratios not in terms of adaptation. It is argued that ‘our perception of an object’s lightness is related not to the
amount of light that is reflected from the object, which can change depending on the illumination, but to the
percentage of light reflected from the object, which remains the same no matter what the illumination’. (p220
in [2]), This is referred to as lightness constancy. However, this works best for flat evenly illuminated objects.
In a 3D space uneven illumination results from shadows and from the orientation of the object relative to the
light source, where some sides are more intensely illuminated than others. For lightness constancy to work in
uneven illumination the visual system uses more cues than just ratios. Here the appearance of graded contours
in shadows such as the appearance of a penumbra together with cognitive processes, such as the ‘shadow’s
meaningful shape’ are used as perceptual cues. That is, for lightness constancy to occur, the visual system must
have adequate information about the conditions of the illumination.

Now we turn our attention to applying this information to lighting design. We saw that although daylight is
constantly changing and we can see the changes taking place, the appearance of the objects with respect to
colour, lightness and depth remains relatively constant. However, the appearance of objects in night light is
monochromatic and blurry with no depth information. The fact that we want to move around at night safely,
and to perform various tasks that require illumination after dark it then makes sense to design spaces that
maximise daylight penetration and to provide adequate electric lighting in the night time to provide a level of
light where at least cone vision is active, in both outdoor and indoor spaces.

But this is not achieved by simply adding more and more electric lamps until adequate lighting levels are
reached because ‘very often we find that (adding more light) does not help the quantity of light is not nearly
as important as its quality.’ (p4 in [5]). Ultimately, the lighting design of a space depends on many factors that
include functionality, aesthetics, user comfort, cost effectiveness and energy efficiency, to name but a few.
Functionality needs alone are too many to address in the space of a few pages. For instance the lighting needs
in a private sitting room at home, an art gallery, a restaurant, a hallway, an office block, a garden after dark etc.
are all different. Even during the daytime, where potentially no electric lighting is required, in a situation that
doesn’t seem to require any specialised lighting except to maximise daylight into a space, can not be solved
simply by the introduction of large windows, because large glazed areas can result in undesirable heat gain
and glare from direct beam penetration. So, in day lighting design the issue is how to introduce the beauty and
functionality of daylight while reducing its potentially undesirable effects.

In what follows we consider some daylight related issues that could arise inside buildings and introduce some
appropriate daylight strategies. The human eye can adjust to high levels of luminance without producing
discomfort. However, veiling reflections reduce contrast and so obscure detail, while excessive brightness
contrast in the field of view causes glare and thus discomfort to the occupants of a space. So the challenge is
how to achieve a balance between maximizing daylight penetration into a space while providing a means of
controlling brightness contrast.
Extending the perimeter form of a building may improve the building’s performance by increasing the total day
lighting area. However, the thermal impact of electric lights and the cost of increased linear length of windows
will need to be taken into account.

The geometry of a building’s walls, ceiling, floors, windows and how they relate to each other has a significant
impact on the quality of daylight within a building. A number of design elements can be employed to increase
daylight penetration into a space and to even out brightness distribution patterns.

Overhangs can block direct penetration of beams of

sunlight and reduce the amount of the sky seen from
a room while catching and directing reflected light
from the ground or other surfaces to the interior of a
room, resulting in higher overall illumination levels
and a more even distribution of light in the space.

Light shelves can be both interior and exterior to

the space. They are used mainly to reduce window
brightness by blocking direct beam sunlight but can
also increase room brightness by the reflection of
light deeper into the space.

Horizontal louvers can block direct beam light An overhang to control direct beam penetration and
during the summer months when sun angles are direct light deeper into a space. (p8 in [5])
high while allowing some sun penetration during the
cooler seasons.

Vertical louvers are useful for east and west

orientations to block direct beam light and to reflect
light into the interiors.

There are also daylight tracking and reflecting

systems that are dynamically controlled to follow the
sun’s path or use strategically placed mirrors that
direct beam light where it is needed. These devices
are used to extend daylight hours and the number of Employing light shelf for deeper light penetration. (p7
months that natural light can effectively replace or in [5])
complement electric lighting.

The kind of glazing used affects both the light and heat penetration into a space. Effective aperture is a measure
of the light admitting potential of a glazing system. It is calculated as a product of the visible transmittance
of the glass used by the window to wall ration (where the window to wall ratio is the ratio of the net window
glazing area to the gross exterior wall area). According to Ander, (p7 in [5]) when effective aperture is around
0.18, daylighting saturation will be achieved and that ‘additional glazing area or light will be counterproductive
because it will increase the cooling loads more than it will reduce the lighting loads.’ Hence the effective
aperture information can be used in evaluating cost effectiveness and the daylighting potential of a schematic
building configuration.

The reflectance values of room surfaces impact the performance of a daylit space. The ceiling is the most
important surface in reflecting daylight coming into a space onto the work plane, followed by the side walls and
finally the floor. In fact dark floor colours will have the least impact on the daylit space.

So, to maximise the ceiling effect, sloping a ceiling away from the fenestration area increases its brightness
and evens out the brightness distribution within a space, while increasing the height of a window from the
floor will dictate the depth of daylight penetration where the higher the window the deeper the penetration.
Generally, the depth of daylight penetration is about 2.5 times the distance between the top of a window and the
window sill. So, spaces that are deeper than this will require supplementary electric lighting.
 The depth of light penetration depends on the height
of the window relative to the ceiling (or floor). (p13 in
[5])
Sloping the ceiling can increase the overall brightness
of a room. (p8 in [5])
Finally, internal day lighting controls can be achieved by the use of venetian blinds, draperies and roller shades
of various degrees of opaqueness. The advantage of including them in the day lighting design strategy is they
can be easily and completely retracted when sunlight and / or skylight are desirable.

Apertures that allow daylight into a building space could be positioned in the walls and / or ceilings. Vertical
apertures in walls are windows or clerestories. We have already discussed windows, clerestories are special
kinds of windows. They are vertical or near vertical openings whose sill height is above eye level but below
ceiling height. Cleresories tend to result in a broader and deeper distribution of daylight than windows and a
reduced likelihood of excessive brightness in the visual field.

Apertures could also be places in ceilings as skylights. These are obviously restricted to the upper level of
a building. Careful control is required because occupant discomfort could result from direct beam sunlight.
However, certain spaces such as circulation spaces can be enhanced with direct beam sunlight. Generally, small,
closely spaced skylights provide more uniform lighting conditions and greater energy saving than larger more
widely spaced skylights, but are likely to be more costly to install.

Roof monitors are raised building elements of a roof with vertical or sloped apertures on one or more sides.
They can potentially allow the top floor of a building to benefit from daylight with less heat gain than with
skylights.

A skylight design allowing daylight penetration to a Roof monitor example. (p18 in [5])
lower level. (p17 in [5])
Atriums increase daylight penetration as well as provide an element of circulation and spatial order within
a building. Since they tend to be within the internal volume of the building they effectively create a second
perimeter zone, allowing the entry of daylight from multiple directions within a space (from the outside
perimeter of the building and from the inside perimeter of the atrium). The intervening surfaces of the atrium
walls can be used to reflect light to adjacent spaces on lower floors.

A complete daylighting system also involves supplementary electric lighting. Controls of electric lighting could
be manual where the users simply turn the lights on when required or an automatic control system could be
employed. These could be switching controls, stepped controls and dimming controls. The switching controls
operate by turning lamps off when a photocontrol senses that a certain daylight threshold level is reached. This
can result in sudden variation in illumination levels which can be disconcerting to occupants. ‘These types of
controls may be effective in building types such as large warehouses where light quality and variations may
not be as significant as in spaces such as offices and libraries.’ (p43 in [5]). Stepped controls offer intermediate
levels of electric lighting and tend to integrate better with daylight availability. They can allow a variety of
lamps to operate within luminaires. For instance a three lamp fixture could allow progressive switching off of
the lamps from 3 to 2 to 1 to 0. Finally dimming controls continuously adjust the electric lights by modulating
the power input to complement or enhance the level of illumination provided by daylight. I guess these can
potentially lead to the highest satisfaction level by the occupants and the greatest degree of energy savings
because the amount of light will be continuously adjusted to meet the required design level within a space.