Okay. I think this is time for me to have a little self reflection. I confess. I often habitually use digital tools without really thinking about what is happening behind the scenes. For instance, I rarely put a thought into what the difference is between desaturation and grayscale, since the results of two look somewhat similar. But they are two different processes yielding varying results. Working in the media art and now studying data viz, I feel a need to build a better understanding of the theoretical background of various data processing techniques and be aware of the gap between human perception and computational processing. Subtleties of Color, the article and talk by Robert Simmon, gave me a good theoretical ground to reach this end.
Simply put, humans are not computers! According to Simmon, human perception is non-linear and uneven, governed by the physiological and environmental limit. We perceive light exponentially as opposed to the linear way of a computer. We are more sensitive to changes at low light levels, and the range of colors perceived by us is uneven. For example, 255 green in the RGB scale looks much brighter to human eyes than 255 blue. This is why numerical color mapping in codes do not always deliver the perceptual uniformity in values and why we are recommended to use certain color palettes to better represent differences in data.
Another interesting fact Simmon delivers is that human visual system is largely driven by lightness. As a good compromise, Simmon suggests combining a linear and continuous change in lightness with a shift in hue. This way, we add a lightness gradient to a change in hue and saturation, get perceptual linearity of a grayscale palette and achieve discrimination from the color component.
Simmon also addresses on designing visualization for color blind people in mind. A substantial portion of people have color deficient vision (8% of male population and 0.5% of female population), and it is only reasonable to use color palettes that can work for everyone. This inevitably leads me to think about a wider range of adaptive design approaches we should consider. In order to reach out to people with different types of disabilities and perceptual limit, we will have to adequately include other reasonable ways in our data “visualization”, such as tactile representation and sonification of data.
Side Note 1: I am happy to once again be confirmed that data visualization is innately multi-disciplinary, where we combine knowledge from many different areas in order to convey the underlying data adequately. It is fascinating that a lot of learning can be done by studying works of cartographers, artists, scientists, visualizers and specialists in the subject of interest, who have accumulated wisdom and honed crafts many years before the widespread use of computers or computational power.
Side Note 2: Robert Simmon's content was largely complementary to Healy’s comment on using already defined color spaces, which I found rather naive in last reading assignment (Link).