Now I understand about the extended color space. But with regard to color matching, isn't it affected by variation in perception? The same colors may look different to different people.
- Yamada
- That's right. Let me tell you about a second advantage of Kyuanos: color coordinates are assigned quantitative values much closer to human perception.
We had to choose a model for the color space that devices can produce; their three-dimensional color gamut. ICC profiles use a method called Lab, which is based on how humans perceive color.
It's said that people sense three aspects of color: lightness -how bright or dark colors appear; color tone-whether colors are closer to red or yellow, for example; and, saturation-whether colors are vivid or dull. These three attributes are represented in a color space as the coordinates L (lightness), ab (hue/saturation). Lab values can be easily converted from color measuring result.
But human perception complicates it. There's a gap between the exact numeric representations of colors and the impression we actually get from them. The question of how to render color perception into spatial models was the topic of several international conferences starting in the 1990s. In 2002, a standard organization involved in colorimetry, called the International Commission on Illumination (CIE) established CIECAM02. We use this color model, which emulates actual human color perception with high accuracy.
Is it much different from the Lab color model used for ICC?
In addition, we reviewed the color-matching method, considering what impression colors give us and how we can enable digital processing throughout the process from input to output. (Yamada)
- Yamada
- In tests using color patches, we found that some colors represented by Lab values match colors as we actually perceive them, but there's nonlinearity in blues, greens, and yellows, for example. There's less nonlinearity with colors in the CIECAM02 model. Because this model closely matches how we perceive color, we can attain highly accurate color performance.
So with Kyuanos, colors are expressed in a way much closer to human perception of color.
- Yamada
- We started developing Kyuanos by quantifying colors. Some factors affect how color appears to us. Obviously, we don't see colors at all without light. So we wondered, how could we quantify lighting? Colors also look different depending on reflectivity, so we wondered how to measure and quantify reflectivity. In addition, we reviewed the color-matching method, considering what impression colors give us and how we can enable digital processing throughout the process from input to output.
To quantify colors, did you measure the individual colors that would be in users' original data, one by one?
- Yamada
- Yes. But although we used measuring instruments, it's more complex than it seems because there are variations even in instruments from the same manufacturer. So, we then began to think how we might manage the measuring instruments themselves, and by extension, all other related processes. We got pretty carried away with it actually. (laughs)
We started off by creating various lighting environments in order to research standard scenarios for the technology.
You must have also paid special attention when you evaluated displays, right?
- Yamada
- Yes. That's because color appears different when you look at something that is light-emitting, such as a display, or something that reflects light. Even colors that give us the same value with a measuring instrument might not look the same to us. For cases like this, we considered psychophysical processes that come into play.
About lighting, it's overwhelming to imagine how many types of lighting there are around the world.
- Yamada
- Yes, I agree. Here, our own development efforts were aided by measurements we commissioned from research organizations in Japan and overseas. We measured the lighting actually used in various environments. We investigated the lighting in offices, art galleries, homes, and other settings. This data collection also involved taking measurements at major electric appliance retailers. That was quite an ordeal! (laughs)
Whenever anyone (including myself) went on a business trip in Japan or abroad, they were asked to take portable measuring instruments with them to measure the ambient lighting conditions.
How did you choose the places for measurement?
- Yamada
- In Japan, North America, and Europe, we chose settings where printed images are generally viewed, such as homes and offices. Apparently, fluorescent lights in Japan seem bluish white from a foreigner's perspective. But to us, the same kinds of fluorescent lights in the United States and Europe seem to have a reddish tinge. Obviously, such lights appear perfectly normal to the people who usually use them in their respective environments.
