apca-introduction

commit

95880645e9e0baaf647ee4cde44cde4cc342d62a

parent

3441ef2954cee304745ba0dcb3dc429d45d3690e

Author

Tobias Bengfort <tobias.bengfort@posteo.de>

Date

2022-07-31 22:18

expand conclusion

Diffstat

1 files changed, 21 insertions, 9 deletions

diff --git a/analysis.md b/analysis.md

@@ -456,15 +456,27 @@ Algorithm (APCA), a new algorithm to predict visual contrast. I compared it to
  456   456 an existing algorithm that has been part of WCAG 2.x, the current standard for
  457   457 accessibility testing for the web.
  458   458 
  459    -1 Though still in early development, APCA already makes two major contributions:
  460    -1 
  461    -1 -   a different color contrast formula that assume much more ambient light
  462    -1 -   a more sophisticated link between spatial frequency and minimum color
  463    -1     contrast that allows for more nuanced thresholds
  464    -1 
  465    -1 It is hard to evaluate APCA from a purely theoretical standpoint. Instead,
  466    -1 thorough empirical validation is required. This has not yet started and will be
  467    -1 a considerable effort. See <https://github.com/w3c/silver/issues/574>.
   -1   459 Though still in early development, APCA is very different from the older
   -1   460 algorithm in many key aspects:
   -1   461 
   -1   462 -   It uses a different luminance calculation that deviates from the standards
   -1   463     but is supposed to be closer to real world usage.
   -1   464 -   It uses a different way of calculating a contrast from luminances.
   -1   465 -   It uses different scaling. Crucially, this scaling is based on a difference
   -1   466     rather than a ratio.
   -1   467 -   It uses a more sophisticated link between spatial frequency and minimum
   -1   468     color contrast that might allow for more nuanced thresholds.
   -1   469 
   -1   470 The new contrast formula agrees with WCAG 2.x for 83.5% of color pairs. That
   -1   471 number rises to 91% for a modified WCAG 2.x formula with an ambient light value
   -1   472 of 0.4. This could indicate that APCA assumes more ambient light. It would also
   -1   473 explain why APCA reports lower contrast for darker colors.
   -1   474 
   -1   475 I like many of the ideas. However, I am concerned by the lack of publicly
   -1   476 available evidence. It is hard to evaluate APCA from a purely theoretical
   -1   477 standpoint. Thorough empirical validation is required. This has not yet started
   -1   478 and will be a considerable effort. See
   -1   479 <https://github.com/w3c/silver/issues/574>.
  468   480 
  469   481 [Web Content Accessibility Guidelines]: https://www.w3.org/TR/WCAG21/
  470   482 [sRGB color space]: https://en.wikipedia.org/wiki/SRGB