I have red-green colourblindness, and I’m trying to make a colour chart suitable for colourblind people.

To my sight, all the following colours have nearly equal greyscale brightness, though the red is a bit darker and perhaps some others are a bit lighter.
rgb values are marked.
Anyone else with red-green confirm or reject this?

For those without red-green, rank the colours in order of greyscale brightness.

Next step after this is to rank the colours by which are most similar and which are most different.

PS. Dark orange is far right. It looks similar to dark yellow to me.

• For those without red-green, rank the colours in order of greyscale brightness.

Define greyscale brightness…

“I have red-green colourblindness, and I’m trying to make a colour chart suitable for colourblind people.”

Your objective is unclear to me.

I presume you have seen the colour space diagrams?

http://www.color-blindness.com/2009/01/19/colorblind-colors-of-confusion/

They give you the lines of colour confusion for the three generally defined types of colour blindness. Click on the spot for protan (poor red, a red light may be invisible at night it is so dull), deutan (poor green, a green light may be so desaturated that it looks white) and tritan (rare, and different genetics)

mollwollfumble said:

I have red-green colourblindness, and I’m trying to make a colour chart suitable for colourblind people.

Luckily this has already been done…

http://colorbrewer2.org

> Define greyscale brightness…

I’ve got a formal and an informal definition. The formal definition is that if a colour is placed immediately adjacent to a grey colour, and if immediately next to the border between the two the colour side appears lighter than the grey side then the colour is lighter than the grey. And visa versa. I’m finding that good, but not good enough. It seems to fail for pure red and pure blue. For pure blue that version gives a blue which in bulk appears too light. So my informal guide is whether the colour in bulk appears brighter or darker.

I can do the formal method myself without help. But that tends to give a wrong answer for primary bulk colours, which is more subjective, hence my need for help. With the formal method I get an equivalence of rgb 50,50,50 with blue 0,0,255. But looking at the bulk colour that blue appears too bright, it’s more like an equivalence of grey 50,50,50 with dark blue 0,0,200 or even darker.

> Your objective is unclear to me.

Normal colour charts, paint charts for instance, with colour names like eggshell, olive, sea blue, confuse the hell out of colourblind people. I’m looking for a set of colours (using a selection of traditional names) that will be instantly separable, comparable and recognisable.

> http://www.color-blindness.com/2009/01/19/colorblind-colors-of-confusion/ They give you the lines of colour confusion for the three generally defined types of colour blindness. Click on the spot for protan (poor red, a red light may be invisible at night it is so dull), deutan (poor green, a green light may be so desaturated that it looks white) and tritan (rare, and different genetics)

Thanks for that buffy! I’ve been looking for a chart like that on the web for ages, with a complete lack of success. I’m deutan, but the lines of confusion on the chart don’t take into account greyscale. I do not confuse green, red and yellow for instance. I confuse vivid green with pale orange and dirty yellow. Whereas the lines of confusion on the chart appear to convert a 2-D colour space into a 1-D colour space, that isn’t true. The lines of confusion convert a 3-D colour space into a 2-D colour space. And a 2-D colour space is just the perfect dimensionality for drawing a colour chart.

> Luckily this has already been done… http://colorbrewer2.org

That’s a start. But what it really does is to eliminate the screamingly bad, not optimise the result to make it the best. Also, as above, what it ends up with is a 1-D colour space when colourblind people really do have a 2-D colour space.

Here are four colour strips, where I’ve tried to match the greyscale brightness for each one. Comparisons please.

You also have to remember that there are degrees of colour blindness. Some are worse than others, it’s a graded scale. The way I demonstrate to parents is to tip a pack of 36 derwent pencils on the floor and ask the child to give me all the green ones. They will give you green, brown and sometimes some yellow. But some give you more than others. There are protanopes and protanomals, and deutanopes and deutanomals.

buffy said:

You also have to remember that there are degrees of colour blindness. Some are worse than others, it’s a graded scale. The way I demonstrate to parents is to tip a pack of 36 derwent pencils on the floor and ask the child to give me all the green ones. They will give you green, brown and sometimes some yellow. But some give you more than others. There are protanopes and protanomals, and deutanopes and deutanomals.

You will also know that the term “colour blindness” is just a useful convention. Deutans, protans and tritans all possess three different colour sensors in their cone cells. It’s not a deficiency, it’s a remapping. For a deutan for example, the colour space is Y,G,B rather than R,G,B. Which means that with a deutan-optimised colour chart it is the supposedly normal trichromats that have the lines of confusion – in this case a deutan is far better at distinguishing different narrowband wavelengths of green than a trichromat. This has been used successfully in the army to fix camouflage, because lines of confusion for trichromats are not confused by deutans. More familiarly it is seen in the ishihara test, where a deutan will see colour differences that are confused by trichromats. It also has been noted in the interpretation of some classical artworks.

It’s just annoying that rgb and cmyk colour maps are so prevalent that one tends to forget that up to 8% of men are seeing in a different 3-D colour space.

Prevalences and genetics, if you need them. Not many tritanope/anomals about. I know one. Female.

http://www.nature.com/eye/journal/v24/n5/fig_tab/eye2009251t1.html

buffy said:

Prevalences and genetics, if you need them. Not many tritanope/anomals about. I know one. Female.

http://www.nature.com/eye/journal/v24/n5/fig_tab/eye2009251t1.html

Thanks buffy, I hadn’t run across them since a scientific american article about them in the 1970s.

mollwollfumble said:

> Define greyscale brightness…

I’ve got a formal and an informal definition. The formal definition is that if a colour is placed immediately adjacent to a grey colour, and if immediately next to the border between the two the colour side appears lighter than the grey side then the colour is lighter than the grey. My informal guide is whether the colour in bulk appears brighter or darker.

Here are four colour strips, where I’ve tried to match the greyscale brightness for each one. Comparisons please.

No answers? OK, then I’ll simplify the question.

In this colour swatch, which grey colour is the closest match to the yellow?

I find myself unable to match brightness of that yellow and the greys.

But if I had to, I’d probably go for the middle one.

dv said:

“I have red-green colourblindness, and I’m trying to make a colour chart suitable for colourblind people.”

Your objective is unclear to me.

The final objective will look approximately like this, with all the numbers deleted and the colours filled in.

The Rev Dodgson said:

I find myself unable to match brightness of that yellow and the greys.

But if I had to, I’d probably go for the middle one.

Yeah, probably the middle one.

Wouldn’t a computer match these better than a human?

The Rev Dodgson said:

I find myself unable to match brightness of that yellow and the greys.

But if I had to, I’d probably go for the middle one.

Thank you!

The middle one is not a uniform colour, it varies from 240,240 240 at the top to 250,250,250 at the bottom.
So it’s very similar to the left hand one.

mollwollfumble said:

dv said:

“I have red-green colourblindness, and I’m trying to make a colour chart suitable for colourblind people.”

Your objective is unclear to me.

The final objective will look approximately like this, with all the numbers deleted and the colours filled in.

The colour chart style below is more cerebral and less empirical – it should be better. The greyscale lightness is now on a slant, which nicely takes account of blue being lighter than black and yellow being darker than white.

By the way, I’m now using 20,20,20 for black because the difference between 20,20,20 and 0,0,0 is very difficult to see on a computer screen. And I’m using 150 as the halfway point for brightness rather than 127.

mollwollfumble said:

mollwollfumble said:

dv said:

“I have red-green colourblindness, and I’m trying to make a colour chart suitable for colourblind people.”

Your objective is unclear to me.

The final objective will look approximately like this, with all the numbers deleted and the colours filled in.

The colour chart style below is more cerebral and less empirical – it should be better. The greyscale lightness is now on a slant, which nicely takes account of blue being lighter than black and yellow being darker than white.

By the way, I’m now using 20,20,20 for black because the difference between 20,20,20 and 0,0,0 is very difficult to see on a computer screen. And I’m using 150 as the halfway point for brightness rather than 127.

Got it. On the following chart, the measured distance between colours is very similar to the perceptual distance apart of rgb colours for a red-green colourblind person of type deutan. The red area to the left attaches to the centre of the area on the right so that the final result is actually Y-shaped.

Be aware that this chart is optimised for computer screens – all the colours come out differently for printers. Also be aware that the full set of colours of the spectrum contain more colours than can be seen by a deutan on an rgb screen.

rgb distance can differ significantly from perceptual distance. For instance, the rgb distance between 255,0,0 and 255,60,0 is 60 which is three times the perceptual distance between 255,60,0 and 255,80,0; but the perceptual distances are very similar.

mollwollfumble said:

mollwollfumble said:

mollwollfumble said:

The final objective will look approximately like this, with all the numbers deleted and the colours filled in.

The colour chart style below is more cerebral and less empirical – it should be better. The greyscale lightness is now on a slant, which nicely takes account of blue being lighter than black and yellow being darker than white.

By the way, I’m now using 20,20,20 for black because the difference between 20,20,20 and 0,0,0 is very difficult to see on a computer screen. And I’m using 150 as the halfway point for brightness rather than 127.

Got it. On the following chart, the measured distance between colours is very similar to the perceptual distance apart of rgb colours for a red-green colourblind person of type deutan. The red area to the left attaches to the centre of the area on the right so that the final result is actually Y-shaped.

Be aware that this chart is optimised for computer screens – all the colours come out differently for printers. Also be aware that the full set of colours of the spectrum contain more colours than can be seen by a deutan on an rgb screen.

rgb distance can differ significantly from perceptual distance. For instance, the rgb distance between 255,0,0 and 255,60,0 is 60 which is three times the perceptual distance between 255,60,0 and 255,80,0; but the perceptual distances are very similar.

I’d like to see the middle one completed.

The lowest chart has lots of contrasting darker and lighter and colour-changing vertical lines in it. So it doesn’t work for me. (I’m moderately red-green colourblind.)