Projets

Colour blindness self–testing

What colours am I unable to see exactly? An attempt to self diagnose.

One of the classes that my school offered dealt with colour, its perception, codification and restitution. One of the few courses that had the ability to gather all of the artistic fields that are taught in the school.


However, the issue is that I am one of many that deals with colour blindness. It is a very common decrease in the ability to perceive color, and can affect different ranges of colours. Mine is in the red to green zone, which is the most common. It is rarely called a disability, because of its effect on the everyday life, but in the case of this class, it became apparent that it was a disability, as we were trying to distinguish colour samples that were very close spectrally, trying to determine a vocabulary to describe all visible axes, focusing primarily on intensity, tint, and brightness. Though the class went deeper into these parameters, and inspect them in as many conditions as possible, I had to take a step back to try and determine how much of a handicap this was going to be for me. Parallel to the class momentum, I focused on this issue in tandem with a teacher of the class.

I was told previously that my inabilities were in the red to green zone. Initially the professional I went to see told me about deuteranopia, and how it related to the actual cones in my eyes that were affected, or abnormal. I tried to dig deeper into this but most research on this field concentrates on the origins of the problem and its hereditary transfer, but very rarely on perception.

img1

I was eager to find a way to being able to describe the situation more precisely than just “red to green”, which was too wide a description, furthermore, I am very well able to see red and shades of - as well as - greens and variants. 

I had to make a few decisions to narrow down my research, taking into consideration the fact that I was looking to find something which I am unable to see. How could I even be sure I was going to get to anything conclusive? Along with the teacher that followed the research, we suspected that if my vision was flawed, I would not perceive the subtleties between certain colours, meaning I could observe color charts and I might see flat zones, areas of no progression, sections where nothing was moving.

I started with a very simple sketch that displayed the LAB color system with a variant of brightness (or as we called in the French class, clarté). Looking through contrasted steps of the system's pages would I be able to point out zones of the displayed colors that looked the same to me? Here are some extracts of this step:

img2
img3

I drew lines onto the graph to point out flat zones, trying to get an idea of scale or range or type of color.

img4
img5

This step was not very interesting in retrospect, but did confirm how little a needle I was looking for.

img6
img7

I decided that I had to work in a restricted spectral zone. A codified color space that would narrow down the infinite range of possible colors, LAB was too large, so I switched to RGB. The system conveniently enabled human readable values to colors that I was going to be looking through.
The second decision was to stabilise the variants: when looking at the previous ranges on the LAB swatches, it was very hard to focus, as the environment was so full of colors, and we knew that colors side by side affect each other’s perception. 

I returned to what we had observed at first in the class, lighting. Colour of light and coloured light. I was going to need to be able to produce lights in different colors and set up an environment to control them.

The next step was to run ranges of colours in a fast slideshow on a computer screen. My plan was to have displayed colours incrementally but regularly switch on screen, continuing to look for a gap in my color vision. In this case the gap was going to translate into a flat zone, a set of frames that were going to be virtually the same to my eyes. I accomplished this thanks to Processing, a simple sketch to run a slideshow of RGB values that I could vary over time. Being my first experience in processing I tackled it the only way I knew how, which was by hand. I'm sure I could have built a function to progressively reassign values to the colour viewer, but duplicating the code using a spreadsheet was easier.

img8

I then decided that working on a computer screen was not ideal because of its actual rich colour spectrum, as well as its not so bright functions. Plus, moving the sketch from one screen to another produced dramatically different results. But I knew the process was working, I knew how frequently the colours were supposed to change, and after some long testing, I found (time) zones when I could not see differences between the displayed colours.

I then decided that switching the screen to something more restricted would be a logical step. I got the idea from these websites, showing how easy it was to set up a colour recognising system with arduino.


http://mattmccoyee.com/other-projects/color-sensor/ http://reibot.org/2011/05/28/color-detection-with-arduino-for-under-5/ http://www.instructables.com/id/Using-an-RGB-LED-to-Detect-Colours/


Although I did not go all the way with the colour sensor etc, thanks to those tutorials, I had enough courage to set off in arduino, if only to reproduce what I had got to on screen with processing - after all they are similar languages. Additionally the spectral zones I had found to be problematic in my previous processing experiments could then be more reliably tested with this new setup.
After some tweaking, the arduino setup was quite simple. In fact, one of the very first tutorials people suggested to learn arduino involved making a light blink, it was all there just waiting. This was what the setup looked like in the end:

img9

I placed a ping-pong ball on top of the led serving as a colour diffuser, and watched.

After a certain duration, the project came to a halt. My research had produced conclusive results, and I had seen (or not seen in fact) zones within which I could not make out differences between two colours that I knew to be different. I have found some greens that are exactly similar to my eyes, in a range between (115,110,110) and (135,110,110) and some reds that are very similar to me (not as clearly as the greens) but even with this setup, the time to go through all the combinations and all the possibilities would take days.

It is also worth mentioning that given the components I used to produce the colours, I could not guarantee the LED was actually reproducing the steps I had programmed into the sketch. The zones that seemed similar to me, might well have been areas in which the (cheap) LED was just weaker at reproducing such subtleties. I showed the specific progression to others who told me they did perceive differences, but very minute ones.

I found that my colour blindness was not very strong, and affected only very small parts of my responses. Pinpointing them would have been interesting, but given the end-of-year period, I had other projects to finish, so I never got around to it. I am glad I took it this far though, because the initial doubts I had in my perception of colours was an obstacle I had to tackle to take the class seriously, I was constantly thinking that I might be looking at a colour that others see in a different way, and I had to find a way to diagnose.

I could see this project taking on a new dimension having seen what other people were able to do in arduino, setting up a system that would warn me if any of the problematic colours were detected, but I have not yet taken the time to get to that.

I had spent a lot of time on this considering the 1h/week time the class was supposed to take, and I had to put other classes into perspective. More importantly, I convinced the teacher that followed the project to grade the project for its process and not its outcome. The steps in between the three main parts displayed here were the most important, about thinking of this specific colour subject as research and challenging the things that were presumed as given.

Return to project listing