Poor bold text rasterization on openbenchmarking.org
Categories
(Core :: Graphics: Text, defect)
Tracking
()
People
(Reporter: 1justinpeter, Unassigned)
Details
Attachments
(3 files)
comparison.png
(First bug here, apologies if I mess this up)
Noticed on https://openbenchmarking.org/s/Intel%20Core%20i7%20950. Comparison with Chrome and font data attached. Reproduces after a browser restart, with a fresh profile, and when manually enabling webrender.
|
Reporter | |
Comment 1•5 years ago
|
||
|
|
Reporter | |
Comment 2•5 years ago
|
||
Also reproduces on nightly.
|
||
Comment 3•5 years ago
|
||
Justin Peter, thank for the report. Can you attach about:support to this bug?
|
|
Reporter | |
Comment 4•5 years ago
|
||
|
|
Reporter | |
Updated•5 years ago
|
|
||
Comment 5•5 years ago
|
||
The web site specifies only font-weight: 400; for web font in CSS(<link href="https://fonts.googleapis.com/css2?family=Heebo&display=swap" rel="stylesheet">).
I think that the website should also specify the bold Web font.
<link href="https://fonts.googleapis.com/css2?family=Heebo:wght@700&display=swap" rel="stylesheet">
|
||
Comment 6•5 years ago
|
||
(In reply to Alice0775 White from comment #5)
The web site specifies only font-weight: 400; for web font in CSS(<link href="https://fonts.googleapis.com/css2?family=Heebo&display=swap" rel="stylesheet">).
I think that the website should also specify the bold Web font.
<link href="https://fonts.googleapis.com/css2?family=Heebo:wght@700&display=swap" rel="stylesheet">
Right. Because the site only specifies a normal-weight (400) face for the web font, but then uses it in elements with bold font-weight, it'll get synthetic bolding applied. This is generally inferior to a true bold face; and sometimes, the DirectWrite "bold simulation" effect really messes up, as here. We've seen this with a few other fonts as well. It appears to be a DirectWrite bug (perhaps triggered by some specific detail of how the glyphs are constructed).
(Chrome includes an embedded copy of FreeType and probably uses it here to render the glyphs instead of using DirectWrite, and that's how it avoids the DW bug.)
See bug 1587094 for more details.
Description
•