Color
From Visualization Sp06
Lecture on Feb 16, 2006
Readings
- Color and information, In Envisioning Information, E. Tufte
- A rule-based system for assisting color map selection, Bergman, Treinish, Rogowitz, (html)
- Color guidelines, Brewer, (html)
Demonstrations
Jason - Feb 17, 2006 03:36:17 pm
After playing with the ColorBrewer a bit more after Pat's lecture, I wonder if there is not a mistake in the icon displays at the bottom left. If you click on the "photocopy friendly" icon, it explains that diverging schemes cannot be photocopied successfully but that differences in lightness should be preserved with sequential schemes. This makes sense, but none of the 5-color sequential schemes (even the one with 5 different grayscale values!) are listed as "photocopy friendly." Is this a mistake?
Maneesh - Feb 17, 2006 06:01:23 pm
Jason - That is weird. Can anyone figure out what is going on with ColorBrewer. Does seem like it might be a bug.
Noaa - Feb 20, 2006 02:17:27 pm
I'm pretty intrigued by the whole color thing... especially with colorblind people. What do they see? In my head red/green colorblind people see gray instead of red or green, but I'm told that's wrong. What is it then? And what is the effect of having red and green as such important meaning-bound colors in society like traffic lights, stop signs, and so on and so forth. In lecture there was something mentioned about an app or website where it shows what colorblind people see... any chance we can get a link up to that? If it hasn't been put up yet, that is, since it's entirely possible I missed it :p
Edit: Nevermind I Googled it and wow... being colorblind is a trip. But I guess if you're born that way you don't know any different so it wouldn't seem weird. Still strange to see how something looks through other people's eyes, though. If anyone else is interested, here is the link I saw: http://www.tsi.enst.fr/~brettel/colourblindness.html
Jheer - Feb 20, 2006 03:46:41 pm
For those of us with Macs, there is another great app simulating a variety of color-blindness phenomena: http://www.michelf.com/projects/sim-daltonism/
Load it up and check out your own graphics work or web page...
Lesliei - Feb 20, 2006 03:00:39 pm
Another good website on color blindness is Vischeck. This website shows you what color blind people see (check out their Vischeck tool) and will also show you how to correct your images for particular types of colorblindness (the Daltonize tool). Incidentally, this website is also recommended by Cynthia Brewer, author of ColorBrewer, who says the work behind it is "sound".
Raymond - Feb 21, 2006 12:12:04 am
This is by far my favorite lecture so far in this class, due mostly to the fact that color is such an interesting topic. Noaa- I checked out the website you showed me and it was pretty trippy to see how colorblind perceive the world. I wish we could've talked about the cognitive/cultural connotations people have about color, though.
Bryan - Feb 21, 2006 09:49:49 am
I'll speak up as someone with (mild) red/green color blindness. Usually the shades I have difficulty with are very desaturated or dark shades. The misconception usually comes from a first impression: I'll see a shirt, think it's dark green, then only years later discover that it's gray (it's led to some interesting color combinations). One I only just discovered a couple of weeks ago was that the little "walk" character at crosswalks is white, not light green like I had always imagined. It's as though my mind filled in what color it expected (green go, red stop). I should mention that I have no trouble with the red and green of stoplights, although they always have a consistent position for people who have near complete overlap of medium/long cones.
IvanTam - Feb 21, 2006 11:23:54 am
I think an important thing to keep in about when working with color is the physiology of the human eye. An interesting and crucial point to to made here is, as a 2003 Slashdot posting put it, "our eyes suck at blue."
This article, I think does a good job of quickly illustrating how JPEG's perceptual model takes advantage of the human eye's relaative dearth of blue-receptor cones in our retina's. Since our eyes are less sensitive to blue than red or gree, comparatively, JPEG is more "lossy" in the blue channel.
The take home lesson, I think, is to avoid using colors in such a way that the view must distinguish amongst many shades of blue.
Yi-Tao - Feb 21, 2006 03:49:43 pm
IvanTam - I'm glad you brought up the issue of jpegs. I realized that they were lossy a few years ago while I was editing an image. Although the article points out that the loss in quality isn't noticeable in one compression, keep in mind that the loss compounds over many compressions. Thus, after editing and saving the image a few times, I noticed the substantial loss in quality between the original and final image. After that incident, I keep my intermediate work in a loss-less format.
Nchentan - Feb 22, 2006 12:34:10 am
Before this lecture, I always thought that the raindow color used in most scientific visualization was "the appropriate" color map to use for visualizing quantitative data. The lecture and the papers were very informative for me, in that they point out the weakness of this colormap and suggest several useful alternatives!
Ashley - Feb 22, 2006 07:50:46 pm
In response to IvanTam, I don't think that that article proves anything about our eyes being less sensitive to blue. Our eyes have S, M, and L receptors, which are receptors that have a peak sensitivity at short, medium, and long wavelengths respectively. While it's true that we have fewer S-cones, recent color theory (opponent process) supports the new theory that a) S cells aren't the only source of blue information and b) the L and M cells also contribute to the perception of blue. What I think that article showed was that the picture had a lot of blue in it, so decreasing the resolution of the blue channel resulted in a less jarring image than decreasing the resolution of the other channels. Also, by pointing out that the blue-channel image was much blurrier, they showed the fact that the minimum threshold needed to detect a difference in hue at the blue end of the spectrum is higher than at the center of the visible range. (It's also higher at the red end, but not as high.)
Brien - Feb 23, 2006 05:24:33 pm
I agree that color is very interesting. This lecture gave me an appreciation for the depth of the subject. I guess I never understood the 128 pack of colored pencils. The singularity/phase plots were especially interesting to me. Some time ago I was playing with putting functions on the screen. I'd take three surfaces, one each (r,g,b), and plot them across the screen, maybe scaled and mod 255. Anyway, you could see where the different color surfaces were all in nearly the same equivalence class by the gray areas, which roughly meant they were all behaving similarly in that area. Although it wasn't very useful in general, it was kind of cool in some cases. Like the phase plots show, it's very interesting how coloring things can highlight characteristics that would not at all be obvious otherwise.
Cynbot - Feb 24, 2006 01:56:51 pm
I hope that Pat can convince video game makers to be sensitive to the colorblind. So many games, esp Halo 2 and Doom 3 use color schemes that make it impossible to segment enemies.
