https://bugs.documentfoundation.org/show_bug.cgi?id=156685

--- Comment #9 from Myndex <m...@andysomers.com> ---
Hi Regina, Heiko, Rafael,

Sorry for my delay in response, and sorry this is causing trouble from the
other bug being fixed.

To highlight the importance of corrections for the underlying issues:

For color RGB(114,159,207), white is higher contrast than black. Depending on
the size of the font this can be quite important to readability and visual
accessibility.

FWIW I have been practically unable to use LibreOffice because of how text
colors are unreadable when my OS is in dark mode, so getting some of these
issues corrected is important for the LibreOffice UX overall.


The present bug

Instead of lowering the threshold, could the default blue be lightened? That
seems like a proactive step in the right direction.



Notes on contrast & methods:

The WCAG 2 contrast math is incorrect, particularly for dark colors. An
APCA-based calculator gives useful results over the visual range.
https://www.myndex.com/APCA/

If the auto color is ONLY white or back, then changing the math is not needed
(probably). The existing but archaic method is fine for a single flip point.
The 156 came from testing and aligning with perceptually uniform methods, and
exactitude is not needed. RGB(114,159,207) with black text rates a contrast of
Lc 50, which is lower than ideal, okay for larger / bolder fonts (32px normal,
21px bold) if readability is important.

RGB(114,159,207) is close to contrast center, so it's perhaps black text is not
a show-stopper, but white text has about 15%-20% more readable contrast.


Perception and Flips

To keep text black with a perceptually similar blue, rgb(125,179,232) is a
reasonable minimum, rgb(135,197,255) is better for black text, and
rgb(158,206,255) is preferred for smaller black text.


Dark mode (white text) perceptually uniform inversions:

rgb(112,155,201) reasonable minimum
rgb(96,136,178)  better for white text
rgb(90,125,156) preferred for smaller white text


Thank you for reading

Andy

-- 
You are receiving this mail because:
You are the assignee for the bug.

Reply via email to