Color Oracle vs. Sim Daltonism

If you’ve been enjoying Sim Daltonism, you may be interested in another Mac OS X application to simulate color blindness. I’ve just found out about Color Oracle. Color Oracle let you see the whole screen as a color-blind person would see it which in some way makes it better than Sim Daltonism. The drawback of this approach is that you cannot interact with what you see; you have to leave the filtered mode before doing anything else. Beside that, both use a different filtering algorithm, which, interestingly, does not give the same results.

I’ve done some tests with Color Oracle and Sim Daltonism to see how different the results where between the two. The test is simple: apply each color blindness filter to Mac OS X’s color selector’s color wheel. The wheel looks like this unfiltered:

Now, if I filter the same wheel with the deuteranopia filter, here is what I get (Color Oracle on the left, Sim Daltonism on the right):

While the result is somewhat similar, there are four striking differences:

• Sim Daltonism alters the grays while Color Oracle does not (look at the striped background);
• Sim Daltonism push blues, greens and reds towards aqua, orange and brown while Color Oracle keeps the blues more pure and the greens and reds more yellow;
• Sim Daltonism replace yellows by whites while Color Oracle does not;
• Sim Daltonism puts some sort of dark spot deeper blues.

The same differences are appearing with protanopia but to a lesser extent (Color Oracle on the left, Sim Daltonism on the right):

With tritanopia however, the results appears to be practically the same:

There is still some differences however: whites are a little whiter and grays (look at the striped background) are a little blueish on Sim Daltonism’s side.

So which one is the best? I’d suspect Color Oracle beats Sim Daltonism on the quality of its filter. I’m not an expert in color blindness nor color theory, but I say that because I don’t see any valid reason for any of the differences noted above. There is no reason for a dark spot in the blues, for the yellow to become white, nor the grays to become something else.

I’ll see what I can do to improve things a bit. That would probably imply using the same algorithm as Color Oracle which is described in detail in this paper (PDF).