Take colour away and compare the large regions
Open a copy of the astronomy club poster. It has a title, a Friday date and a way to reserve a place. A greyscale copy can help you see whether those pieces have a useful light–dark order. Try the check early, while the poster is easy to rearrange.
Three passes
First, make a greyscale copy or use your image editor’s desaturate preview. Keep the original. A screenshot works for the visual comparison if that is what you have. On a phone, view the two versions at the same size. You can do this exercise without changing the phone’s accessibility settings.
Second, blur the copy slightly. Compare the title, date and invitation at the size people will use. Try one change to size, spacing or brightness, then compare the two greyscale copies again. A large dark photograph may compete with small text, or two regions may look equally prominent.
Third, go back to the original. Ask someone who has never seen it what the event is, when it happens and how to reserve a place. Note what they found and what they missed. Try the same task on a phone for a web page, or at the intended viewing distance for a printed poster.
Measure contrast on the original colours
A greyscale preview compares light and dark regions. For web text, use a contrast checker with the original foreground and its actual background. If the text crosses a photo or gradient, check the background behind the letters where separation is weakest. Use a plain backing when that makes the text easier to keep readable.
In the WCAG calculation, sRGB channel values are linearised before they are weighted to calculate relative luminance. The weights are 0.2126 for red, 0.7152 for green and 0.0722 for blue. A plain average of the stored channel values or a fixed gap between two grey codes cannot give the contrast ratio. W3C relative luminance definition
For a quick illustration, pure red and pure green both have the same average of their three 8-bit RGB values. Their relative luminances in sRGB are 0.2126 and 0.7152 respectively. The same raw average can hide a large difference.
Try two 16px labels on a white background, one in #767676 and the other in #777777. The colours look close, but a contrast checker reports about 4.54:1 and 4.48:1 respectively. Only the first reaches the 4.5:1 minimum for ordinary web text. Compare the unrounded result with the threshold. In your own design, choose a margin above the minimum when you can. WCAG Contrast (Minimum)
Read the result
Even when a layout works in grey, read the labels, test any controls, and verify the original colours. For example, replace a red or green dot that alone means pass or fail with a visible text label such as Passed or Needs review. A greyscale comparison cannot tell you whether a screen reader can identify a control or whether a keyboard user can activate it. WCAG Use of Color
If you cannot use a visual comparison, read the title, date and invitation in document order, or use a screen reader on the page. Check whether the event and action are clear from the words. The visual comparison and the order check answer different questions.
Try it on your own design
Save the original and one changed version. Record the item you want a new reader to notice first, the one visual cue you changed, and what the reader actually found. For a web design, record one text colour pair and its contrast ratio. If the reader notices the right item but cannot read it, treat prominence and legibility as separate problems and test both again.
Check your understanding
Pure red and pure green have the same plain average of their 8-bit RGB channels. What should you use to compare their contrast as web colours?
Pick the one you would defend. Your answer stays private.
Your place and first choices stay on this browser. Classroom credit is recorded separately.
Stuck on something?Ask Meher, your AI tutor