The first image below represents a background of asphalt, snow, and evergreen trees, with cyclists wearing different high-visibility colors in front.  The second image is the same thing, processed by Vischeck to show what someone with the most common form of colorblindness would see.  My preliminary conclusion is that yellow and royal blue work well, whereas red and some shades of bright green don't work well at all.  Royal blue has the additional advantage that it would also stand out against fall leaf colors; yellow would not.  However, yellow is generally considered a more highly visible color, and one color-blind cyclist wrote to tell me that bright yellows work better for him than bright blues.  Perhaps a contrasting combination of yellow and blue is best, but it would seem that one could be much more confident of yellow than of blue.

            


Note that my conclusions only really hold for achieving visibility through color contrast.  Lights and reflectors that work on the principle of standing out by being bright may work well even if they are the "wrong" colors.  However, some colors are hard for some people to see; in particular long-wavelength red light is hard for some to see.  Red LEDs can be problematic, because they emit a narrower range of wavelengths than a typical incandescent lamp with a lens.  Anecdotal evidence suggests that 630 nm red LEDs are OK, whereas 640 -660 nm LEDs can be problematic. Unfortunately, when you buy an LED bike light, you don't get to find out what wavelength the LEDs use.  And from the Agilent LED catalog, it looks like the brightest red LEDs are 637 nm, which may not be as good a choice as a 626 or 615 nm led with a slightly lower rated brightness, but with better visibility to the color-blind.

Here's a very telling comparison using a photo from a web site that sells high-visibility stuff for kids, http://www.playitsafe.biz/ .  It seems to show that most of it just doesn't work for people who are color blind.  The high-visibility green and orange hats become a light color that blends well with the trees.  The hot pink beocmes grey.  The only things that work are the yellow kerchief on the right and the white shirts.  The cute smiles come through too, but that's hard to maintain on a long commute.  

     

One of the questions that these images raise is whether the popular greenish yellow jackets and jerseys that are used by cyclists for visibility are more like the yellow bandana (bottom right) or the green hat (top left).  To check this, I found everything yellow in my cycling wardrobe and set it up outside on a cloudy day, as shown below:



The left and right jackets are both made by Pearl Izumi.  The green one on the right seemed to stand out the most in person, and I think also in the picture above.  The jersey next to it seemed to me to be a similar hue, but not nearly as bright.  I assume that this is because the jacket has more fluorescent dye, which converts blue and UV light to yellow/green, whereas the jersey simply reflects the yellow/green light that hits it.

Here are three simulations from Vischeck for the three different types of color blindness:

Deuteranope (most common):


Protanope (also common):


Tritanope (very rare):


The main conclusion is that all the shades of yellow or yellow-green I tested worked well for any type of color blindness.  The advantage of the fluorescent yellow-green jacket seems to be diminished a bit for color-blind people, but it is still on par with other types of yellow cycling gear.  Anecdotal reports from a (very) few color-blind cyclists I've heard tend to confirm that fluorescent yellow-green jackets are good.

I'd encourage you to try pictures of your own at the Vischeck web site.

Comments?  charlie.sullivan@dartmouth.edu