Status Update
Comments
eh...@gmail.com <eh...@gmail.com> #2
Information redacted by Android Beta Feedback.
eh...@gmail.com <eh...@gmail.com> #3
Thank you for reporting this issue. We have shared this with our product and engineering team and will update this issue with more information as it becomes available.
ka...@google.com <ka...@google.com> #4
ka...@google.com <ka...@google.com> #5
Thank you for reporting this issue. We have a fix rolling out in an upcoming release.
eh...@gmail.com <eh...@gmail.com> #6
pe...@google.com <pe...@google.com> #7
ka...@google.com <ka...@google.com> #8
Observed the gray box is renedering
Note:-Isuue not seen in Win 10 and Linux debian, Mac 14.3
Reproducible on
==============
123.0.6303.0 - Canary
123.0.6300.3- Dev
122.0.6261.49 - Beta
121.0.6167.160 -Stable
Forward Bisect Information:
---------------------------
Good Build:116.0.5816.0
Bad Build:116.0.5817.0
Bisect Script:- python3 bisect_builds.py -o -a mac -g 116.0.5816.0 -b 116.0.5817.0
CHANGELOG URL:
Suspect:-
Change-Id: I2a04d15f89612d3898b139e5c2852236213aa585
Reviewed-on:
Christopher Cameron@Please help us in re-assigning if this is not related to your change.
Thanks..!!
cc...@chromium.org <cc...@chromium.org> #9
This goes away when running with the flag --disable-features=CoreAnimationRenderer,CoreAnimationHDRUnderlays
I would guess this to be a bug in how we handle solid color CALayers.
tc...@chromium.org <tc...@chromium.org> #10
Do you have any updates on this issue?
cc...@chromium.org <cc...@chromium.org> #11
Haven't dug in on this.
pd...@chromium.org <pd...@chromium.org> #12
Moving to "Chromium > Internals > GPU > Image" which is where similar bugs are filed, and because #9 mentions this is related to CALayers.
su...@chromium.org <su...@chromium.org> #13
[GPU bug triage] ccameron@ gentle ping for an update here :)
zm...@chromium.org <zm...@chromium.org> #14
eh...@gmail.com <eh...@gmail.com> #15
cc...@chromium.org <cc...@chromium.org> #16
I'll try to get another swing at this.
cc...@chromium.org <cc...@chromium.org> #17
The bug appears to be in HDRCopierLayer (which ideally wouldn't exist).
cc...@chromium.org <cc...@chromium.org> #18
Indeed, it performs color space conversion without un-premultiplying first.
cc...@chromium.org <cc...@chromium.org>
ap...@google.com <ap...@google.com> #19
Branch: main
commit 9f6aa3f74f98b64a4355f2859e19d572d8baa57f
Author: Christopher Cameron <ccameron@chromium.org>
Date: Sat Aug 31 11:40:21 2024
metal::HDRCopierLayer: Un-premultiply before color conversion
Color conversion only makes sense on opaque color values. Un-premultiply
before performing color space conversion, then re-multiply afterward.
The test in this bug used the premultiplied sRGB pixel value of
(0.5, 0.5, 0.5, 0.5). Converted to linear space, this should still be
(0.5, 0.5, 0.5, 0.5). Because the color space conversion was done on
the premultiplied pixel values, it evaluated to
(0.214, 0.214, 0.214, 0.5).
Bug: 325133349
Change-Id: I1ff3f8b000a803bacb92988e6e83d136e21eb101
Reviewed-on:
Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Commit-Queue: ccameron chromium <ccameron@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1349560}
M components/metal_util/
va...@google.com <va...@google.com> #20
Verified the fix on latest Canary M130 #130.0.6693.0 using Mac M1 14.0 as per #
Hence, the fix is working as expected and adding verified labels: TE-Verified-130.0.6693.0 & TE-Verified-M130
Attaching screencasts for reference:
Thanks..!!
Description
Steps to reproduce the problem
Problem Description
When you render an HDR image inside of a white container, and that container has opacity less than 1, it will display the wrong color. See attached test file. The box renders as gray on Chrome (incorrect) and white on Safari (correct). Note also if you move the left position of the image so that it overlaps the left box, the color will change to white. See attached test file and video.
Summary
HDR gainmap image causes wrong background color to render when using opacity less than 1
Custom Questions
Which component does this fall under?
Blink>Image
Does this work in other browsers?
Yes - This is just a Chrome problem
Additional Data
Category: API
Chrome Channel: Stable
Regression: N/A