w3c / silver

Accessibility Guidelines "Silver"
https://w3c.github.io/silver/
Other
199 stars 44 forks source link

Usage of the word contrast throughout the guideline and how to's #332

Open sdw32 opened 3 years ago

sdw32 commented 3 years ago

I found 4 different usages of the word contrast throughout the guideline: 'contrast' used on its own, without prefix (e.g. from 7.5 https://w3c.github.io/silver/guidelines/#visual-contrast-of-text)

Provide sufficient contrast between foreground text and its background

'Luminance contrast', (e.g. from 7.5 https://w3c.github.io/silver/guidelines/#visual-contrast-of-text)

Provides adequate luminance contrast between background and text colors to make the text easy to read.

'Visual contrast' (e.g. from 7.5 https://w3c.github.io/silver/guidelines/#visual-contrast-of-text)

Visual contrast of text

'Perceived contrast', (e.g. from 'Get started' within visual contrast of text how to (https://www.w3.org/WAI/GL/WCAG3/2020/how-tos/visual-contrast-of-text/)

EXAMPLE: you can increase perceived contrast by:

When contrast is used on its own in this guideline, I did not understand which kind of contrast was being referred to. Also, I didn't understand the difference between 'Perceived contrast' and 'Visual contrast'.

'Visual contrast' appears to be being used as meaning something approximately equivalent to perceptibility, for which a dictionary definition is (https://www.dictionary.com/browse/perceptible).

capable of being perceived; recognizable; appreciable:

I had a look at the usage of the phrase 'Contrast of text' on the web, outside of WCAG 3.0 and related documents. Almost universally, this phrase was being used to mean 'Luminance contrast'. I couldn't find anywhere else on the web where the phrase 'Visual contrast of text' was used in an equivalent manner to its usage here in WCAG 3.0.

Therefore, I would conclude that introducing this new phrase 'Visual contrast of text' in this guideline will likely cause significant confusion, with many people not understanding how this is different to 'Luminance contrast'.

Instead, I would suggest the guideline should use the words 'Luminance contrast' when this is being referred to. I would suggest the guideline should avoid using the word contrast on its own. Of course, 'Saturation contrast' and 'Hue contrast' may also be relevant, and these phrases should be used where appropriate.

Critically, I would suggest that the ultimate outcome that the guideline should aim to achieve is to make sure that:

The phrases 'Visual contrast' and 'Perceived contrast' are not outcome measures in their own right, and they are not properties that web developers can directly specify, and I don't currently understand if or how they can be objectively measured, so I am not sure of the benefit that is gained by introducing these terms. Instead, the guideline can just refer the extent to which the outcome measures (legibility and readability) are influenced by the things that web developers are in control of: primarily line-height, font-name, font-size, font-weight, and by choosing foreground and background colours, the luminance contrast, and if relevant, the hue contrast and saturation contrast.

If it is necessary to separate out the extent to which text can be perceived, as one several factors that influence readability, I would suggest using the word perceptibility. As far as I can gather, this would be approximately equivalent to current usage of phrases 'Visual contrast' and 'Perceived contrast', although I do not entirely understand the intended meaning of these terms, so I might be mistaken!

The heading for the section that talks about these issues might perhaps be better named 'Visual accessibility of text', which can then introduce readability and legibility (depending on the use case). Visual perceptibility of text might also be another viable option to explore.

Myndex commented 3 years ago

@sdw32 Thank you for commenting.

Unfortunately, there were many changes that were conducted by multiple individuals that have resulted in some language inconsistencies, in addition, some parts were written a year apart, by partially different teams. MOREOVER, there is significant effort being made to make everything "plain language" and removing technical terms and jargon.

Luminance contrast is one specific kind of contrast, you see it hitting high in search results, because it is WCAG 2.x. WCAG 3.0 Visual Contrast is a COMPLETE break from WCAG 2.x

Plain Language

For plain language, "lightness contrast" or "lightness/darkness" contrast is probably going to replace most uses of "luminance contrast" in WCAG 3.0, because unlike WCAG 2.x, WCAG 3.0 and APCA are perceptually based, and we should drop the term luminance contrast completely in the forward facing plain language materials.

LUMINANCE is a linear measure of LIGHT, but it is not linear to perception. Luminance contrast, especially suprathreshold, does not accurately predict contrast perception.

LIGHTNESS is a perception. PERCEPTUAL LIGHTNESS such as Lstar is a linear model of perception — linear to perception, and as a result non-linear to light. BRIGHTNESS is a different perceptual quality, as is DARKNESS.

APCA uses perceptual lightness modeling as part of how it functions to predict perceptual contrast for readability.

Visual Accessibility

The actual term I have wanted to use (since 2019) is "Visual Accessibility" under which the many aspects of visual presentation are organized. "Visual Contrast" is an interim compromise for a variety of reasons.

Again, much of this is discussed at https://www.w3.org/WAI/GL/task-forces/silver/wiki/Visual_Contrast_of_Text_Subgroup

VISUAL CONTRAST IN BRIEF

Newer Terms

readability contrast

The predicted perceptual contrast for readability

I am using the term "readability contrast" to encompass the primary collection of contrast metrics required for best readability. This is more plain language and more understandable than "perceptual contrast" I think, and I am using it more instead.

"... I don't currently understand if or how they can be objectively measured..."

Here is the link (previously provided) to the very brief whitepaper on how it functions. This discusses an earlier version, but the general concepts are the same.

The underlying science is also described the bibliographies provided on the VC link, which have a wealth of information, as well as the books I've suggested on color appearance models, and general color science, which might be a good place to start.

"...Saturation contrast' and 'Hue contrast' may also be relevant,.... "

As far as "readability" perceived lightness contrast (from luminance processing) is what is important. Hue, chroma contrasts are processed separately, and far weaker, and far lower resolution than luminance-based perceived contrasts. Chroma/Hue are not directly useful to readability, except that some hue/chroma combinations can hinder readability by causing glare or chromatic aberration related effects.

"....Single words of text (e.g labels) are presented in such a manner that they are clearly legible (legibility)..."

We will not be using the term "legibility" like this, as legibility implies threshold, and threshold is not an appropriate place for any design guidelines to say the least. There is fluent readability, and spot reading, which have different critical values of certain metrics, but both are suprathreshold.

Thank you,

Andy

Andrew Somers W3 AGWG Invited Expert Myndex Color Science Researcher https://www.myndex.com/APCA/simple

ChrisLoiselle commented 3 years ago

@jspellman This is ready for survey. Potential action item for W3C would be to add all of these terms related to contrast to the W3C glossary of terms.

ChrisLoiselle commented 3 years ago

DRAFT RESPONSE: We believe the questions have been answered, per Andy's comments https://github.com/w3c/silver/issues/332#issuecomment-787050774

A resulting action item for W3C would be to add all of these terms related to contrast to the W3C glossary of terms.

I hope this answers your question, if it does not, please feel free to follow-up.

Thank you, Chris

sdw32 commented 3 years ago

@myndex thanks for your comments.

I think Perceived Lightness Contrast is an excellent way of describing this concept. I prefer this to readability contrast, because it's ambiguous as to whether increasing the size would (or wouldn't) increase the readability contrast. I would think the most straightforward way of describing these terms as to say that the readability of text is an outcome that depends on many factors, but especially the perceived lightness contrast of the foreground/background colour combination, and the size of the text.

I am assuming here that perceived lightness contrast is intended to mean a concept that is independent of size (in APCA I believe this is Lc which is purely as a function of the foreground/background colour combination). However if I have misunderstood this intention then please let me know!

And yes, I agree that legibility implies threshold and 'spot reading' is a better term. So it's about spot reading for labels etc. and fluent reading for body paragraphs of text.

I will look forward to further discussions on this topic!