Light, Vision and Imagery
By: David Wood - November 28, 2001
Now that we have an understanding of what light is, we can begin to study how we see it. The sense of vision, taken for granted by most of us, is very useful and has many advantages:
It allows us to detect an approaching predator before it bites our arm off.
It allows us to recognize and locate food without tasting everything we find.
It allows us to communicate using body language, facial expression and writing.
It allows us to navigate our surroundings without injury.
It allows us to recognize a compatible mate, avoiding the potential embarrassment of attempting to mate with a person of the wrong gender.
When you stop to think about how useful vision really is, it is not surprising that there was an evolutionary need for most creatures on this planet to develop it. The part of our bodies that we developed for the sense of vision we call "eyes".
The Human Eye
The human eye is a sensor that can detect and focus on EM radiation between the wavelengths of about 400 and 700 nanometers. This may seem kind of limited considering that modern camcorders can record in the infrared band too, but before you let your new digital camcorder go to your head consider the following…
The eye is a complex piece of precision optical biology. The complexity of the human eye is often used in arguments for the existence of a divine creator because it is so extraordinary that such an excellent piece of optical machinery could have evolved purely by chance. Whether our eyes are an example of biological evolution at its best, or proof of the existence of god is an argument best left outside of this article, but either way, we are gifted to have them.
Here is a cross section of the human eye, the front of the eye being on the left side of the image. Light passes through the cornea where the image of the outside world is refracted. Contrary to what many believe most of the refraction in the eye takes place in cornea, not in the lens.
The lens is covered by the iris and the hole in this iris we call the pupil. The iris can change shape, increasing or decreasing the size of the pupil and therefore increasing or decreasing the amount of light entering the eye. When you look at a bright light your pupil is smaller than when you look into a dark room. This is a subconscious process.
The light that passes through the lens projects an image onto the retina, the lens is used to fine tune this refraction and focus the image. Man-made optical instruments tend to focus this way by moving a set of lenses back and forth. Think of the telescoping lens on a camera, obviously our eyes don't use this method and it would look very odd if they did. Our eyes instead focus by changing the shape of the lens itself. The lens is not solid, it can bend and stretch when the ciliary muscles pull on it.
The retina is covered with millions of light sensitive nerves called photoreceptors which are categorized into rods and cones. They send electrical impulses down the optic nerve bundle to the brain when they are stimulated by light. Rods are "tuned" to the whole visible light spectrum and are more sensitive to light than cones. Cones however are tuned to a specific wavelength of light and they are what allow us to see in color. Cones come in three flavors each tuned to a different color as can be seen on this graph:
If you were thinking of "red", "green" and "blue", like on a TV or computer monitor, you weren't far off. The human eye is most sensitive to the colors "reddish-yellow", "greenish-yellow" and "bluish-violet" (575, 535 & 445 nanometers).
At the center of the retina is a little dimple called the fovea. Here there are lots of cones and no rods (since the cones take up all the space). When we focus on something in the center of our vision we are concentrating on the image projected onto the fovea and sensed in high-resolution color. Because cones aren't as sensitive as rods, the fovea is actually a disadvantage under low light conditions. Because of this, soldiers training for night operations are taught that when observing an object with the naked eye they are to look slightly to the side of it instead of directly at it. The way that rods work better under low light conditions also explains why colors appear to be more vivid during the day than they do at night.
The human visual system does some image processing before the conscious mind sees it. The fact is that what you see isn't 100% reality, which has all sorts of philosophical implications that we wont go into.
First of all, we see the world upside-down. The refraction that takes place within the eye causes the image projected on the retina to be wrong way up, we simple perceive this way as the right way. Another example of image processing in the eye is the blind spot. Where the optic nerve joins the back of the retina (refer to the cross section of the human eye show previously) there are no rods or cones at all. This would mean that in each eye we would see a black spot permanently in our field of vision if not for the fact that visual system removes it. This removal method is quite clever; it fills in the spot based upon the colors and shapes of the image that surround it. You can "see" this blind spot if you know how to look for it…
View the above image at arms length, look at the cross and cover your right eye. Slowly move your head closer to the page, keeping your vision focused on the cross, but pay attention to the dot in the corner of your eye. As you move closer the dot will appear to vanish. This is because the dot is now over the blind spot on the retina of your left eye and your visual system is filling in the gap with the white color from the surrounding page.
We should also note that cones and rods do not transmit instantaneously. In fact in low light conditions they can take a few thousandths of a second longer to transmit signals than they do under higher light conditions. Also, when the light is no longer stimulating the rod or cone it will still continue to send signals for a fraction of a second afterwards. This is called persistence of vision. Because of this affect we can perceive a television screen, that is flickering at 50-60 times a seconds, as a steady continuous moving image. Perhaps the best illustration of persistence of vision is when writing your name in the air with a "sparkler" firework.
Depth perception allows us judge distances, which is an advantage to predatory mammals (such as man). It is the reason why we have two eyes instead of one.
Stereo vision is our primary means of depth perception. It is also the reason why human beings have their two eyes in the front of their heads facing forwards. An alternative configuration might be like that of some lizards with eyes on the sides of their heads giving them almost 360 degree vision. Although we don't have wide-screen vision like a lizard, our forward facing eyes are ideal for stereo depth perception.
When we view a scene with our two eyes, each eye sees the scene slightly offset from the other (because our eyes are offset on our face by a couple of inches). This difference is called binocular disparity and the brain uses this to perceive depth. This form of depth perception is called stereopsis.
The principle behind stereopsis is actually simple trigonometry. If you are focused on an object close to, then both your eyes are pointing inward and if you are focused on an object further away, then both eyes are looking almost parallel to each other. If you are looking to the horizon then you eyes are in a totally parallel position. If you are trying to focus on something right on your nose, then your eyes are pointing severely inward and that is when the person is said to be "cross-eyed". Your brain can sense how far inward your eyes are pointing to know how far away the object of focus is. Your brain basically triangulates the position.
This method of triangulating distance is used to calculate distance on a much large scale too. Have you ever wondered how astronomers know how far from the Earth the moon is? Well, one way is to view it from two different locations on the Earth simultaneously. Since we know the distance between the two locations, and we know the angles at which they view the moon, we can use trigonometry to calculate its distance. Our eyes can't measure the distance to the moon, it appears as far away to us as the stars behind it. This is because our eyes are too close together and the difference in viewing angle would be too small to sense. But for perceiving depth at smaller distances, like the objects in the room, stereopsis is ideal.
Not everybody has sight in two eyes, yet there are some other ways in which we can get some idea of depth. These methods do not require stereopsis, and they can easily be recreated on 2D media like pictures or film.
Depth of Field
Our eyes, like a basic camera can only focus on one distance at a time. If an object is closer to or further away than the object on which we are focused, it appears blurred. This blurring is called depth of field and is used by photographers, film directors, and computer games to convey the illusion of depth and to direct the viewer's focus on a particular part of the image. To see this effect close one eye, hold your finger about a foot away from you and focus on the room behind it, your finger will appear blurred. Now focus on your finger and the room behind it will appear blurred. Below is an example of depth of field, courtesy of Robert Ian Axford.
If you look at the picture above, you can tell that the handle of the guitar is closer to you than the guitar player or the plant behind him. Obviously the image doesn't just leap out at you like a hologram, but when we are viewing the real world we do so dynamically and not viewing everything as a static 2D dimensional image like the one above. Our eyes will focus back and forth as we concentrate on the different objects around us and in doing so we get an idea of how far away or close to we to our surroundings.
Perspective and Parallax
Take a look at the following image and figure out which tree is closest:
The tree on the left is of course the closest one... or is it? The image is entirely computer generated, so this image is entirely two-dimensional. Why then, do we perceive the tree on the left as being closer? There are several characteristics of the image that make us perceive it this way.
First of all, parts of the tree on the left obscure some of the branches of the tree on the right (this is called occlusion). Secondly, the tree on the left takes up more space in the image. Our brains assume that both trees are roughly the same height and therefore if one looks larger than the other then it must be closer.
If we see something that stretches out into the distance, such as a road, then its width will appear less and less the further down the road we look until it reaches a tiny point on the horizon called the vanishing point (marked VP on this picture). If we were to drive down this road then the lines on the road and the street lamps at the side of it would appear to get larger and larger as they approached us. This effect is called linear perspective and is quite easy to draw from hand on a two dimensional picture, or recreate on a computer screen. You may have drawn pictures like these in high school.
This effect can also be seen in a lot of computer games, below is a screenshot from a popular computer game called CounterStrike. In games like these the perspective is based upon the center of the screen where the cross hair is. Notice how the wall on the left appears to get smaller and smaller into the distance. Perspective can also be seen on the entertainment center, the blue sofa, and the gun.
Apart from things looking smaller the further away they are, they also tend to move slower across our field of vision. This effect is called motion parallax. Parallax can be seen when you are riding in a train, looking out of the window and observing the countryside scrolling by. Things close to the train like trees along side of the track will fly past very quickly, on the other hand, things in the far distance will move by more slowly. Motion parallax is also quite easy to recreate and is often seen in cartoons and computer games.
Correctly shading an image, casting shadows and drawing highlights in the right places can help to make it look as though it has more of a three dimensional shape. Computer graphics often use this form of shading to make buttons look 3D, text areas to look indented, windows to look higher than others etc. This illusionary form of 3D only works when the viewer believes that the light is being cast on the object from the correct direction. In computer graphics we usually pretend that the light source is shining from behind us, from above and to the left. Look at the picture of blue text. If you make believe that there is light shining on it from the lower-right direction, then you can convince yourself that the "raised" text is actually lowered and the "lowered" text is raised.
The air around us is not totally clear, it is filled with dust and other particles and over great distances some objects can appear dimmer than others if they are further away. When viewing distant objects like mountains on the horizon, our senses of stereo depth perception and focus are of no use since our eyes are focused to infinity. In a natural scene like this our sense of perspective is also limited because there are no flat objects or surfaces to give a perspective effect. At these distances the only thing that tells use that one mountain is further away than another is that there is more haze in front of it. A variation of the haze effect is fog. On a foggy morning objects will become more visible the closer they get to us, and less visible the further away they are.