What is Color Science?

[Admin]

‘Color Science’ as the words suggest is a scientific field that deals with content related to colors. It refers to ‘scientifically measuring colors’’. Not a lot of us are familiar with how color relates to science, so we question what ‘color science’ is. Perceiving colors is related to the sense perception of humans which makes the process very subjective. This makes it hard for people to relate it to science. They do not understand how this process could be made objective. According to them, how can color be measured in metrics in the same way as length or weight?

‘Color’ is something that humans feel. It is quite difficult to grasp how our visual system could align with that of others just because everyone goes through the same process, where the photosensitive cell present in our eyes absorbs light and our brain interprets that particular visual. But, let’s think of our daily lives. Everyone learns about colors since childhood. We exchange information using different color names when drawing pictures and buying and selling products. People use color names without much confusion. Not everyone may be the same, but we identically perceive colors.

[Characteristics of color perception by humans]

We need a source that emits light, an object, and most importantly a person (specifically, the human visual system) as shown in [Figure 1], to see with our eyes. The light from the source reaches the surface of the object and the object’s surface absorbs some of the light and reflects the rest depending on the unique reflective properties of the object. The light reflected by the object or the light from the source of illumination enters our eyes, this light is absorbed by the photosensitive cells in our retina and sent to the brain so we can ‘see’.

The physical stimulation that induces the cognitive stimulation of colors is light that is incident on the eye with a visible wavelength area in the range of 380nm~780nm. There are three kinds of color perception cells in the human retina. Even if the spectral characteristics of light are different, if the photoreceptor cells generate the same signal, we perceive it as the same color. Thanks to this property of ‘Metamerism’, we can see the same colors as we see in nature on the TV (additive mixture of red, green, and blue light) or in books (subtractive mixture of cyan, magenta, and yellow ink).

(Color Constancy)

When we describe the object that we see, we use color as one of its main characteristics. For example, color is one of the first things we describe for a newly bought outfit because we see it as an inherent characteristic of the outfit. It does not constantly change depending on the surroundings.

The color of an object varies depending on the reflection spectrum i.e. how much light is reflected by each wavelength. If there is an object that mainly reflects a long wavelength range area of about 600 nm and above out of a wavelength range area between 380nm to 780nm, and has the property of absorbing the remaining wavelength range, this object appears red, and if it reflects most of the light of the complete wavelength range of the visible ray, then it appears white. As such, the most basic physical property that determines the color of an object is the reflection spectrum. But the property of light reflected from an object varies greatly depending on the characteristics of the light coming from a particular source. In contrast to the bluish source of illumination which produces fewer long-wavelength rays, the yellowish source of illumination produces a lot of long-wavelength rays. Even for objects that have the same reflection spectrum, the spectrum of light reflected from these two sources of illumination is bound to take completely different forms.

Nevertheless, the fact that we can perceive the color of the object as always unchanged, is thanks to our visual system. As we can know from [diagram 1], the information that a person should get from color is of two types consisting of ‘information on the property of light from the source of illumination’ and the property of reflection of the ‘object’ but the information on the color that enters a person’s eye is a combination of these two. And so, the process of obtaining enough information from a color just does not end with measuring the light using the cells in one’s eyes but also requires the process of interpreting it in the brain. Visual information is processed in a place called the visual cortex, located in a part behind the brain. In the process of dealing with information, a person’s visual system offsets the intensity of light from the source of illumination or any color changes to some extent. So the lights we perceive in our surroundings are constant regardless of any change in the source of illumination.

This kind of character is called ‘color constancy’. One of the mechanisms that make color constancy possible is that we always perceive white as white and we perceive other colors against white making it a standard i.e. we always perceive white as the brightest color in our surroundings, we perceive red when more long wavelength rays are emitted and blue when a lot of short wavelength rays are emitted against the white standard. That is why a white paper always appears white regardless of whether it is seen in candlelight or broad daylight.

To understand constancy that depends on the change of intensity of the source of illumination, let us recall the experience of looking at a smartphone screen. If we look at the smartphone screen in a somewhat dark room indoors and then go somewhere bright outdoors, the screen suddenly looks dark. In contrast, if we look at the smartphone screen before we come to sleep and then turn off the lights, the screen suddenly looks bright. This is because, even though the brightness of the light coming from the display remains fixed, the sensitivity of our eyes changes automatically to make up for the change in brightness. If we go out to a bright place from a dark place, the screen looks relatively darker as the eyes are adjusting to the intensity of light in the bright place.

However, this does not mean that color constancy is perfect. This is because the colors we see may appear different depending on the situation.

Colors that look different depending on the surroundings

If you look at [Figure 2], there are two grey squares one over black and another over white. These two grey colors are created using the same digital signal but the grey over the black looks brighter than the grey over white. This is known as ‘brightness contrast’. Similarly, in [Figure 3], the pink color made up of the same digital signal is placed over a deep red background on one side and a grey background on the other. Physically, they are made up of the same stimuli but our eyes perceive the color over the grey background as a deeper pink.

As we can see in [Figure 2] and [Figure 3], the colors look different based on the background it is against even though the object is the same, this phenomenon also occurs in the process of a human’s visual system trying to gather color information. We need to obtain meaningful information through ‘vision’. If we take the examples in [Figure 2] and [Figure 3], we can say the space where the color starts changing is more significant than the part where the color remains the same. In the event of new information like this, our eyes consider a difference in information to be more important than a change in physical size. i.e. our eyes perceive the change in the respective backgrounds of black and white with the grey square on them more distinctively. Thus, viewing it on a darker background than a white background makes it look brighter.

It’s not only the background color that affects the color we see. Let’s go back to our daily life. Everything looks dark and drab when it is cloudy. On the other hand, on a sunny day, the color of all objects such as the green of leaves, flowers planted along the roadside, clothes worn by people, etc. look bright and clear than on a cloudy day. Apart from the color difference caused by ambient brightness, the color of objects in a room looks a bit different depending on whether the source of illumination emits a bluish light or a yellowish one.

I think by now the readers are probably confused. We mentioned previously that the visual system of a person offsets the intensity of light from the source of illumination or any changes to color to some extent and this is why the colors of objects look constant regardless of any change in light. And right after it says that color perception could differ depending on the surroundings. These two phenomena which instantly sound contradictory, occur simultaneously. This is because color constancy is not perfect and just takes place to some degree. This imperfect color constancy enables us to distinguish between an early morning and a bright daytime, between a bluish and a yellowish light even while giving the perception of a ‘certain degree of’ fixed color in objects.

[How to measure color? – CIE (Colorimetry)]

Color holds great importance in modern society, especially for business purposes. Product design and color are of the essence when buying cloth, furniture, electronics items, etc. The designer may play a big role in making high-selling products but let’s think about the problems after making a design. Maintaining consistency in color is crucial for a company’s image. If the color of a product differs slightly every time we buy it, or in case of the clothes, if the color of a jacket and pants are somewhat different, the quality of it comes into question on top of the product looking shoddy. In other words, the quality of high-quality goods should be tightly controlled. Thus, to manage the quality of products, we need a method to measure and digitize color just as we do for height or mass.

We can say that modern color science took shape in 1931 when the business world demanded this sort of “color quantification method”. For the last hundreds of years, much research was conducted to understand the mechanism of identifying colors, but in 1931, an international standard was proposed for the first time to express color through numbers in the Internal Commission on Illumination (CIE). All color-related internal standards and color markings in the business world proposed later are based on this method.

[Figure 4] conceptually shows the color quantification method (Colorimetry) proposed by International Commission on Illumination in 1931. The source of illumination is shown on a power distribution spectrum and the object on a reflection spectrum. Human eyes are expressed via three types of sensitivity functions (color matching function) labeled as the standard observer. Color can be denoted by the tristimulus values of XYZ by combining the three elements of source of illumination, object, and human eyes. All colors occurring in nature can be shown within a horseshoe shape by CIE xy color coordinates where XYZ values are projected on a two-dimensional plane in the last graph shown in [Figure 4].

As we saw in the descriptions above, humans interpret colors using information from our surroundings too and so even the same physical stimulus could be perceived differently depending on the environment. Colors seen with the eyes in the scope of ambient colors and diverse observation environments cannot be described with just XYZ. So, a lot of experiments have been conducted since the 1980s on how colors perceived in different observation environments change. Various color application models have been proposed based on this. The color appearance model provides information on color, brightness, and saturation as seen by the eye by obtaining information on tristimulus values and the ambience. The CIE proposed a standard model called CIECAM97s in 1997, modified the same and, proposed CIECAM02 in 2002, and the CIECAM16 model recently.

[Color coordination amongst imaging devices]

We wake up and fall asleep watching numerous videos taken on the camera on displays like the TV or smartphone every day. As such, advanced color engineering technology is used in displays and cameras, products that are indispensable to modern man.

Both cameras and displays use the three color signals of red (R), green (G), and blue (B) by default. The camera absorbs external light to produce RGB signals, and the display combines three types of RGB lights in an appropriate ratio to reproduce various colors. Now, the RGB created by the camera and the RGB output by the display take on physically different meanings. Consequently, color scientists like me, research and develop technologies to implement the same color amongst imaging devices with different physical properties and apply them to products.

Color estimation for both camera and display is possible by functions of the same type, i.e. the difference between function types in technology to be applied to camera and display is not very large in terms of technology used to improve image quality. However, this only means that the apparent form of the technology is similar, not that the same technology can be applied to both. The researcher/developer has to properly understand the physical characteristics of the input device in the case of a camera, and the physical characteristics of the output device in the case of a display, reflect the difference between the two when developing a technology.

[Figure 5] concisely shows how colors captured by the camera are brought up on the monitor and printed by a printer. The camera records the captured color values in RGB signals. The RGB value recorded for the same scene filmed by different cameras may vary greatly depending on the type of camera. Red, green, and blue filters are always placed in front of the video sensor in the internal structure of cameras. Light passing through the lens passes through the filter, gets absorbed by the sensor and turns into electric signals. Therefore, even if the same light enters the camera, completely different red, green, and blue signals are created depending on the filter properties and exposure time.

Now, let’s think about one RGB video made by a camera and sent to two different monitors. LCD monitors are commonly used these days. If we look at the rough structure of the LCD monitor and there is a backlight or light at the very back and a liquid crystal display in front of it. Liquid crystal has the property of changing penetration ratio in response to applied electric signals, the brightness in LCD monitors can vary using these. Color signals are made when light passing through the liquid crystal passes through red, green, and blue color filters attached to the front. As such, the color of a monitor is affected by the characteristics of the backlight, liquid crystal, and color filter. The colors red, green, and blue that can be emitted by a monitor change depending on the materials used in manufacturing it and on the way it is operated. And so, even for the same signals, monitors with different product characteristics inevitably emit different colors.

[Figure 5] Colors look different resulting from the unique properties of imaging devices

Finally, let’s think about how we send the images we see on a monitor to the printer. It is common for printers to create colors using the four inks of CMYK, namely cyan, magenta, yellow and black. The ink used by each company is different and thus their color properties also vary. So even if the same CMYK signals are sent, the color on the actual print is contrasting based on the printer type. Not only that but printers are greatly affected by the properties of the paper used and that of the external light.

As such, imaging devices can display all colors created by each of them with RGB or CMYK signals and exchange color information by these color-space coordinate values. Color coordinate values such as RGB or CMYK are all ‘Device Dependent Color spaces’. So, to display the colors created by imaging devices based on colors as seen by people and not as understood by imaging devices, color information needs to be exchanged using ‘Device-independent Color Spaces’ such as CIE XYZ.

[Fig.6] conceptually shows how imaging devices ensure consistency between colors. First, if we want the color in the camera and monitor to be consistent, it is necessary to know what the colors referred to by the camera RGB signal mean. To do this, we can mathematically model the relationship between RGB signals and the colors we see by analyzing the color properties of the camera. The same applies to monitors. The combined technology from these two technologies can help predict the actual color from the RGB value captured by the camera after which the monitor can calculate which RGB value to output to reproduce the same.

Over the past few decades, technologies have been researched and several standards made to reproduce the same colors amongst imaging devices, but it is still not easy because of several reasons. In many cases, it is physically impossible to reproduce the same colors between devices. The reason is that the range of colors that can be reproduced as per the imaging device and the viewing environment of it varies. It can be created on the monitor, but some colorscannot be printed and vice versa.

Another thing to keep in mind when ensuring that colors are consistent between different imaging devices is that implementing better image quality is more important than color consistency. We think that liking something is subjective and thus our preferences do not need to coincide with that of others. But, when someone asks us for our favorite color, it is different from when we are asked if we like pink or blue in the sense that the prior is abstract. But what if we are asked about the color of a pretty face and that of an appetizing apple? What’s unbelievable (or maybe not) is that there’s not a lot of difference between how people perceive the color of their skin and that of an apple (red). We have slowly formed an impression of what color daily objects around us should be, like the color of skin, that of the sky, and that of an apple. These are called ‘memory colors’. But as it happens, the actual color of the object may not be as we remember it. In many cases, we tend to remember colors darker than they are. We have also come to know that our preference for a familiar object in a picture goes up if the color of it is closer to how we remember it as a memory color. This exact relation between the color in our memory and the color we prefer is applied to the image quality improvement technologies of imaging devices such as TV or camera.

A smartphone is a product where color science is most intensively applied. This is a product where the camera and the display communicate organically to provide optimal picture quality. As with all fields of science and engineering, color science continues to develop. Also, new and improved image quality technologies based on the knowledge of this new color science continue to develop. We surely look forward to the future of image quality in smartphones.

Reference

[1] This article is written by Professor Youngshin Kwak of UNIST (Ulsan National University of Science and Technology) . The ideas and content expressed here are not the official description or standpoint taken by Samsung. They are the ideas and opinions of Professor Youngshin Kwak.