New issue
Advanced search Search tips
Note: Color blocks (like █ or █) mean that a user may not be available. Tooltip shows the reason.

Issue 877484 link

Starred by 1 user
Status: Assigned
Owner:
reed@google.com
Cc:
█ allanmac@chromium.org
█ csmartdalton@chromium.org
mtkl...@google.com
█ caryclark@google.com
reed@google.com
Components:
EstimatedDays: ----
NextAction: ----
OS: Mac
Pri: 2
Type: Bug-Regression



Sign in to add a comment

Bad antialiasing for object composed of arcs

Reported by martin.s...@gmail.com, Aug 24 
UserAgent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_13_6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/70.0.3532.0 Safari/537.36

Steps to reproduce the problem:
1. Open the link https://jsfiddle.net/fu7srdc4/19/ (or have a look at the attached screenshot 'Canary70')

What is the expected behavior?
The edges of the objects (drops) should like those of the line. It worked in Chrome 65 (see attachment 'Chrome65', broken with Chrome 66).

What went wrong?
The edges of the objects (drops) look stepped.

Did this work before? Yes Version 65.0.3325.181 (Official Build) (64-bit)

Does this work in other browsers? Yes

Chrome version: 70.0.3532.0  Channel: canary
OS Version: OS X 10.13.6
Flash Version:
Canary70.png
269 KB View Download
Chrome65.png
267 KB View Download

Comment 1 by senorblanco@chromium.org, Aug 24 

Can repro on Linux (so it's probably utilizing CPU raster, not GPU).

Will do a bisect.

Comment 2 by senorblanco@chromium.org, Aug 24

Cc: reed@google.com
Owner: liyuqian@google.com
Status: Assigned (was: Unconfirmed)
Bisect says:

---

You are probably looking for a change made after 539952 (known good), but no later than 539955 (first known bad).
CHANGELOG URL:
  https://chromium.googlesource.com/chromium/src/+log/a2154fd14494fd1229c70e24a22ca5a325850133..612feb71496d3492dadd3d60cb61414d987797be

---

The only rendering-related change seems to be "Enable Delta Anti-Aliasing and Rebaseline" https://chromium.googlesource.com/chromium/src/+/e3053cdf8cfac55d0a4589a7f1fcc60796d60c41

Yuqian, could you take a look or reassign? Thanks!

Comment 3 by liyuqian@google.com, Aug 24

Cc: allanmac@chromium.org csmartdalton@chromium.org
This is a known issue with Skia's newer AA algorithms: skbug.com/6886

The solution is to not put a lot of closely overlapped edges in a path: https://jsfiddle.net/fu7srdc4/28/

Comment 4 by martin.s...@gmail.com, Aug 25 

That's unfortunately not a solution in my case, as I actually use arcs to render pen strokes with interpolated width based on velocity and pressure. Overlapping arcs are more than sufficient from performance perspective, with very good results, and few effort.

How are the chances that the Skia fix will make it?

Comment 5 by reed@google.com, Aug 25

Cc: mtkl...@google.com caryclark@google.com
Will take a look on Monday

Comment 6 by martin.s...@gmail.com, Sep 9 

Did you have a chance to look?

Comment 7 by liyuqian@google.com, Sep 9

Owner: reed@google.com
►

Sign in to add a comment