New issue
Advanced search Search tips

Issue 798754 link

Starred by 1 user

Issue metadata

Status: Available
Owner: ----
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux , Mac
Pri: 2
Type: Bug



Sign in to add a comment

feSpotlight behavior differs across browsers

Reported by t...@tobireif.com, Jan 3 2018

Issue description

UserAgent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_13_2) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/63.0.3239.108 Safari/537.36

Steps to reproduce the problem:
Load https://codepen.io/smfr/pen/aEWgXj

What is the expected behavior?
feComponentTransfer should work correctly.

What went wrong?
feComponentTransfer doesn't seem to work right.

Did this work before? N/A 

Does this work in other browsers? N/A

Chrome version: 63.0.3239.108  Channel: stable
OS Version: OS X 10.13.2
Flash Version:
 

Comment 1 by t...@tobireif.com, Jan 3 2018

The linked test SVG gets rendered (very) differently by current Chrome, Firefox, and Safari.
Labels: Needs-Triage-M63
Labels: -Needs-Triage-M63 Hotlist-Interop
The feComponentTransfer seems to be behaving the same in Firefox and Chrome. It's the lighting that differs.

I've updated the bug name appropriately. Not sure whose behavior is right.
Status: Available (was: Unconfirmed)
Summary: feSpotlight behavior differs across browsers (was: feComponentTransfer doesn't seem to work right)

Comment 5 by f...@opera.com, Jan 4 2018

Status: WontFix (was: Available)
I see two major differences:

1) In Gecko the spotlight has no falloff. [1] seems to imply there should be some form of falloff (although it talks about "antialiasing", which is mildly confusing I think.)

2) The feDiffuseLighting filter in Gecko seems to always operate in "sRGB". It appears that is the BZ in [2]. (One can observe the difference in Blink to some degree by adding color-interpolation-filters="sRGB" to the primitive in question.)

(1) appears to be more of a spec issue than an implementation issue, and (2) seems to be handled correctly in Blink AFAICT. Filed [3] for (1).

I don't see anything that we ought to be doing here at this point in time, so WontFixing.

[1] https://drafts.fxtf.org/filter-effects/#element-attrdef-fespotlight-limitingconeangle
[2] https://bugzilla.mozilla.org/show_bug.cgi?id=993934
[3] https://github.com/w3c/fxtf-drafts/issues/244
Please test https://codepen.io/smfr/pen/aEWgXj on macOS. Chrome there shows a weird purple halo around the green spotlight, which I believe is an artifact from the feComponentTransfer (its output should be binary green).

Comment 7 by f...@opera.com, Jan 4 2018

Status: Unconfirmed (was: WontFix)
@schenney, could you test that?

Comment 8 by t...@tobireif.com, Jan 4 2018

Here's a screenshot of the SVG at https://codepen.io/smfr/pen/aEWgXj in Chrome on Mac OS:

screenshot_chrome_mac.png
34.9 KB View Download

Comment 9 by t...@tobireif.com, Jan 4 2018

The purple haze is not there in Firefox, for example.
(But in Firefox there seems to be another issue:
https://bugzilla.mozilla.org/show_bug.cgi?id=1427976 )
Status: Available (was: Unconfirmed)
Yep, I repro the haze on mac. Very odd.

Comment 11 by f...@opera.com, Jan 4 2018

Components: Internals>Skia Internals>GPU>Rasterization
Labels: OS-Linux
I can actually see it on one of my Linux machines too, but there it's a reddish-brown haze.

Given that it seems to coincide with the portion of the spotlight that has alpha != 1, so something with the unpremultiplication of alpha in the shader perhaps...
Project Member

Comment 12 by sheriffbot@chromium.org, Jan 7

Labels: Hotlist-Recharge-Cold
Status: Untriaged (was: Available)
This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
Status: Available (was: Untriaged)

Sign in to add a comment