Update contrast-minimum.html

[bruce-usab]

• Light editorial to opening intent.
• Add paragraph about color blindness being addressed.
• Add sentence about 4.5:1 permitting black-on-grey-on-white.
• Add link to NEI explainer on color blindness.

[bruce-usab]

This PR closes issue #2033.

[mbgower]

Looks good @bruce-usab. I committed a couple of trivial edits straight into your changes. Please take a look.

[bruce-usab]

Thanks @mbgower -- those are edits are very good.

[Myndex]

I have a longer response with deeper dives in the issue thread request. #2033

The Term "Color Blind"

Before I get into "technical stuff" I want to mention an agenda item I want to discuss for on the LVTF, namely the term "color blind" ... I am hoping we can start normalizing something that is more accurate, as 99.98% of CVD types are not color blind, they are color limited. But they still see colors.

EDIT >>>>: Deutan, protan, tritan are color vision deficient. Achromatopsia, blue cone monochromacy, are color blind, and commonly are also co-morbid with low vision and photophobia.
Deutan, protan have "low color vision", but also commonly have good or better than standard visual function overall.
<<<<

AS SUCH, I suggest that line in the PR say something like

"sometimes, though incorrectly, referred to as color blindness"

or perhaps (my preference but longer)

"formerly referred to as 'color blind', though in fact the vast majority of color-limited vision types do differentiate at least some colors"

A Couple Other Pull Issues

Additional technical discussion in the issues thread #2033 , but if possible can we:

FIRST LINE(18) (see next post):

For the first line in the PR, can we add "...so long as luminance contrast is high enough to support rapid reading." the suggested edit of the first line then should read:

For people without color deficiencies, hue and saturation have minimal or no effect on legibility so long as luminance contrast is high enough to support rapid reading (Knoblauch et al., 1991).

Please see post below for the revised line (18).

Second Line (19)

It reads "Color deficiencies can affect luminance contrast somewhat."

Can we change to:
"Color deficiencies can affect luminance contrast in some cases, as can other impairment types."

CVD Line (22)

I suggest removing any claims of assisting Monochromacy because that is unsupportable.

Other

In the issue thread I outline my other objections, most of which I have maintained since opening issue #695 in April 2019.

Thank you!

Andy

[Myndex]

Regarding the FIRST LINE (18)

While I suggested a minor edit in the previous post, I wanted to comment further as that line seemed a bit askew and out of context relative to some more recent research. I discuss in detail in this post in #2033

Suggested New Line 18:

For people with normal vision and reading very large letters, hue and saturation have minimal effect on legibility. (Knoblauch et al., 1991) However for people with low vision and for people with color limited vision, luminance contrast is the most important for readability (Legge, Psychophysics of Reading XI). This also holds true for normal vision when using small letters such as in body text. (Barten 2003, CSF)

Thank you,

Andy

[mraccess77]

I'd suggest we not say "normal vision" - what is normal? Perhaps typical or just stating without visual impairment.

[Myndex]

I'd suggest we not say "normal vision" - what is normal? Perhaps typical or just stating without visual impairment.

Hi @mraccess77

Hi Jonathan,

"Normal" vision is a specific definition, and a clinical definition:

• Snellen acuity of 20/20 or lower (20/16 is "perfect" acuity, 20/200 is SSA disabled)
• Peli-Robson contrast sensitivity of 1.95 or higher (2.25 is "best")
• Farnsworth Munsell Hue Color TES of 60 or less (TES 0 is perfect)
• 100% Visual field with MD not lower than -2db (0 is best, -20db is SSA disabled)

Depending on the study, "normal" can include "with refractive correction" if the correction can result in the above scores. An example: "normal with correction needed for presbyopia". Presbyopia is "normal" for over age 40 for instance.

"Normal" is not "most perfect"

Because normal is specifically defined, it also provides a useful baseline relating to user needs. Normal vision has natural limitations that define minimum requirements for readability. From this foundation of normal vision needs, we can then indicate the degree of additional accommodation needed for varying degrees of impairment(s).

[patrickhlauke]

"Normal" vision is a specific definition, and a clinical definition

doesn't the same also apply to "Color Blind"?

[alastc]

It looks like most US places default to color blindness, but I quite like how the NHS in the UK approach it:
https://www.nhs.uk/conditions/colour-vision-deficiency/

Defaulting to "Colour vision deficiency (colour blindness)".

[Myndex]

Hi @patrickhlauke

"Normal" vision is a specific definition, and a clinical definition
doesn't the same also apply to "Color Blind"?

No, the clinical term for "color blind" is Color Vision Deficiency (CVD), and in every clinical or academic example I've seen, where they may use the colloquial term colorblind early in the article, they also immediately point out that people are not actually blind of color except the very rare cases of achromatopsia.

I would like to see the public adopt the term "Color Limited."

[Myndex]

I'd suggest we not say "normal vision" - what is normal? Perhaps typical or just stating without visual impairment.

Hi Jonathan @mraccess77

I just wanted to follow up on this — it looks like the term "standard vision" is used often in research, and I think it is more neutral than "normal vision" as such I intend to start using "standard vision" in my writing.

Thank you for pointing this out,

Andy

[mbgower]

@alastc I suspect this is another situation where an outstanding PR was missed because the issue was never marked for survey, so I just added it.

[bruce-usab]

It looks to me like these edits got out of sync with other proposed changes. I am all for closing this PR and working from the most resent version, since that might be the easiest way to resolve the conflicts.

No, the clinical term for "color blind" is Color Vision Deficiency (CVD)...

I agree with Understanding using more precise and correct terminology. Please note that in the U.S. (at least) CVD is fairly widely understood (among laypeople) to mean cardiovascular disease.

...where they may use to colloquial term colorblind early in the articly, they also immediately point out that people are not actually blind of color .except the very rare cases of schromatopsia.

I am of the opinion that 1.4.3 Understanding should fit that pattern.

I would like to see the public adopt the term "Color Limited."

That term is not very catchy! Yes, people's use of words changes over time. I am big proponent of people-first language.

IMO it's a matter of pointing out that "Color Blind" is ableist.

I have been in (live) conversations where "ableist" was used to effectively argue a point. Regardless, it is not a term that I (personally) feel comfortable using. I am skeptical that it would help this Understanding page.

[patrickhlauke]

reads a bit odd/informal with the "somewhat" at the end. Maybe "Color deficiencies can have a slight impact on the perception of luminance contrast"

[bruce-usab]

I agree about the "somewhat" at the end. How about using parallel of previous sentence? So instead of:

Color deficiencies can affect luminance contrast somewhat.

We would have:

For many people with color vision deficiencies, luminance contrast has a significant impact on legibility.

[gundulaniemann]

I feel the impact of luminance difference is on all people and not specific to color vision deficiencies. To the contrary, people with color vision impairments often have a very big luminance sensitivity.

How do you like the following:
The intent of this Success Criterion is to provide enough contrast between text and its background so that it can be read by people with moderately low vision.
This user group does not typically use contrast-enhancing assistive technology.
For many people, luminance contrast has a significant impact on legibility.
For people without color deficiencies, hue and saturation have minimal or no effect on legibility as assessed by reading performance (Knoblauch et al., 1991).
Yet color deficiencies can also affect luminance contrast.

[Myndex]

Hi @gundulaniemann

I feel the impact of luminance difference is on all people and not specific to color vision deficiencies.

As far as reading and readability is concerned, it's all about luminance contrast, for all sighted users.

The visual word form area (VWFA) is the pre-lexical processing that recognizes whole words in the letter pears, and it functions on the luminance channel exclusively without consideration of color as a hue or colorfulness. MRI studies using the Stroop test confirm this (as does the Stroop test itself for that matter).

All vision types require ample luminance contrast for best readability. By "ample" I mean what Ian Bailey/Jan Lovi-Kitchin call the "contrast reserve", for best fluent readability, they suggest 20 times the threshold (JND).

To the contrary, people with color vision impairments often have a very big luminance sensitivity.

Yes, the deutan/protan/tritan forms of color vision deficiency (CVD) have as good or better contrast sensitivity as standard vision. The exception is protan and red/orange against black.

I am going to respond to this more in depth in the later post.

[patrickhlauke]

maybe add a "do" to make it a bit clearer ("For people without color deficiencies ... People who do have a color vision deficit..."

"will also have adequate contrast between the text and the background." maybe at the end, add - to make it a more complete sentence - "if the requirements of this Success Criterion are followed", as currently it's left a bit hanging...

[bruce-usab]

Agree that this needs improvement.

[mraccess77]

I think what the text may be trying to say is that contrast while important to all - low contrast may still allow for reading of text because of the difference in hue can augment it - for example red text on green background with typical color vision. However, in this example - for people with limited color perception the aid of hue is not available in some cases and thus sufficient contrast is more important in these specific situations.

[Myndex]

Short Recap

For readability, all sighted users need ample luminance contrast.

For all sighted users, saturated blue should be the darkest of a color pair, as blue has nearly no luminance. This goes for saturated reds and oranges, but reds and oranges are critical because they are seen even darker by the protan. Reds, oranges, and purples should generally not be paired with a darker color like black.

In other words, per the most recent peer reviewed science, luminance contrast is not negatively affected among color vision deficient,²⁻³ with the exception of reds paired with darker colors like black.⁴

A Gentle Objection

These lines lack accuracy:

Therefore the contrast is calculated in such a way that color (hue) is not a key factor.

Despite summing the colors to a calculated achromatic luminance, hue is still a key factor, because the WCAG 2.x math is not perceptually uniform and does not consider the luminance deficit experienced by certain CVD types, namely protan who see reds and oranges darker. See the "Minutiae" twisty below.

People who have a color vision deficit (often referred to as "color blindness") will also have adequate contrast between the text and the background.

Not necessarily. The WCAG 2.x contrast math can not deliver on this promise, and therefore this is not a statement that should be made IMO. WCAG 2.x contrast does nothing special nor particularly helpful for color vision deficiency. See the "Minutiae" twisty below.

---

Please Consider This Alternate Wording:

The intent of this Success Criterion is to provide enough luminance contrast between text and its background so that it can be read by people with moderately impaired contrast perception, without the use of contrast-enhancing assistive technology.

Hue and saturation have minimal to no effect on readability, as assessed by reading performance.¹ Further, Color Vision Deficiencies (CVD) per se, do not negatively affect luminance contrast perception.²⁻³

However, because some CVD types see colors containing a substantial amount of red as darker (less luminance), colors such as saturated reds, oranges, and purples, should usually be the darkest of a color pair.

Refs:

1. Effect of chromatic and luminance contrast on reading Knoblauch, Arditi, Szlyk, 1991. DOI: 10.1364/JOSAA.8.000428
2. Luminance contrast sensitivity in color-deficient observers
3. Contrast sensitivity of patients with CVD

---

---

The Minutiae

CLICK for foundation and details

The deutan/protan/tritan forms of color vision deficiency (CVD) have as good or better contrast sensitivity as standard vision²⁻³. The exception is protan and red/orange against black⁴.

In fact dichromat CVD such as deuteranopia and protanopia have slightly better visual function overall than standard vision, a current theory being due to a lower neural noise.

As far as readability is concerned, and luminance contrast is concerned, the one issue is with protanopia. As protanopia is missing L cones, they are insensitive to red and colors composed mostly of red, such as orange. They see reds and oranges darker, and therefore, when reds or oranges are against black, the protan has significantly reduced luminance contrast perception.

Generally speaking, reds, and some oranges, should be the darkest of a color pair and this does also apply to standard vision, but it is critical for certain CVD types. In other words saturated reds and many oranges & purples should be paired with white.

WCAG 2.x contrast does not calculate to accommodate this, and instead frequently rejects white yet allows black, the opposite of what is needed for color vision deficiencies. (this is partly because the WCAG2 math is not divided around contrast center for the top and the bottom 4.5:1). (Performance claims, if they exist, should be removed).

Aside from protanopia, the actual color-blind types, i.e. the rare blue cone monochromat or rod-only achromatopsia, also do not "see red". But these types require AT as they are co-morbid with low vision and severe photophobia⁵.

Tritans are not affected as far as luminance contrast is concerned, as S cones (blue) are not a part of luminance in the human vision system.

Deuteranopia, having all L cones and no M cones have essentially standard red sensitivity (but only luminance and a 'saturation' for distinguishability from green), and as the L cone response overlaps M, there is little change and studies have confirmed that deutans have equal or better luminance contrast sensitivity compared to standard vision²⁻³.

REFS
2] Luminance contrast sensitivity in color-deficient observers

3] Contrast sensitivity of patients with CVD

4] What's Red and Black and Also Not Read

5] CVD Simulator with experimental achromatopsia sim

[Myndex]

Hi Jon @mraccess77

...low contrast may still allow for reading of text because of the difference in hue can augment it...

At the higher spatial frequencies of most text, color contrast becomes insignificant to readability. There are some studies that have shown that color contrast can be readable... but they were also using gigantic letters (over 560px !!).

Here's an example, the first three are maximum color (hue) contrast, with little to no luminance contrast. The last one is luminance contrast and related saturation contrast, but no hue contrast.

It's important that the majority of text, especially text that is intended to be read, rely on achromatic luminance contrast.

EDIT TO ADD:

... difference in hue can augment it ...

Also, Legge et alia and others have determined that color contrast as in hue sat, does not add to luminance contrast, they are separate processes, and follow separate paths. I.e. if you have x amount of luminance contrast and x amount of color contrast, you do not get 2x contrast.

Also, MRI studies indicate that color contrast (hue) is not a factor in the visual word form area of pre-lexical processing.

[bruce-usab]

@Myndex I very much like your updated intent:

The intent of this Success Criterion is to provide enough luminance contrast between text and its background so that it can be read by people with moderately impaired contrast perception, without the use of contrast-enhancing assistive technology.

Your next two paragraphs are educational, but I do not think they on topic for Understanding 1.4.3.

Hue and saturation have minimal to no effect on readability, as assessed by reading performance.¹ Further, Color Vision Deficiencies (CVD) per se, do not negatively affect luminance contrast perception.²⁻³

However, because some CVD types see colors containing a substantial amount of red as darker (less luminance), colors such as saturated reds, oranges, and purples, should usually be the darkest of a color pair.

[Myndex]

Hi @bruce-usab

Your next two paragraphs are educational, but I do not think they on topic for Understanding 1.4.3.

I agree...

Rather than delete these three sentences, I was rewording/correcting some of what was above, i.e. could be deleted, here's why I reworded:

Color deficiencies can affect luminance contrast somewhat.

No, except for protan with reds on black. Otherwise, CVD have equal or better visual function, including contrast sensitivity, as standard vision.

Therefore the contrast is calculated in such a way that color (hue) is not a key factor.

MMMmmm... This is an unusual statement—reducing the color to luminance does not disregard hue, hues are weighted per expected luminance contribution... But more importantly we don't want to make people think that hue is not important, because it still is for some things, just not readability.

People who have a color vision deficit (often referred to as "color blindness") will also have adequate contrast between the text and the background.

No, not with WCAG2 math they won't. This is a performance claim which shouldn't be made.

References supporting these supra (in the above posts)

[bruce-usab]

I am taking a fresh stab at this, with a third PR.

[alastc]

Sorry, not superseded by #3240, re-opening.

[alastc]

Ok, clarified with Bruce, this is superseded by #3284