Bug 121173
| Summary: | REGRESSION: Color rendering varies between images and HTML elements | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Pokey <webmasters> |
| Component: | New Bugs | Assignee: | Tim Horton <thorton> |
| Status: | RESOLVED WORKSFORME | ||
| Severity: | Major | CC: | ap, dbates, simon.fraser, thorton |
| Priority: | P1 | ||
| Version: | 528+ (Nightly build) | ||
| Hardware: | Mac | ||
| OS: | All | ||
| URL: | http://login.cridb.com/crid/render.html | ||
Pokey
REGRESSION: Colours render differently when the same value is assigned to an image (jpeg,png,gif) and an HTML element (div bgcolor). This makes it difficult to create seamless transitions between images and HTML elements. The visibility of the problem varies depending on the monitor hardware (on a low-end LG monitor the difference screams as opposed to a high-end NEC monitor where the difference is subtle but still visible.)
On the webpage the problem shows on the background of the arrows in between the columns.
DigitalColor Meter shows a difference in the readout: (HTML element) L*a*b* 80.39-2.51-4.52 (GIF image) L*a*b* 79.22-2.51-3.51
This problem does not appear in Chrome, Firefox, or Internet Explorer (all up-to-date versions as of 9/11/2013).
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Pokey
This bug did not occur in Safari 5.
Alexey Proskuryakov
Could you please attach a screenshot with this issue visible?
Pokey
I get a 500 error when I try to upload a screenshot.
Here is a test page that shows the problem with a checkerboard pattern: http://login.cridb.com/crid/render.html
The colour difference can be very subtle, but again the use of the DigitalColor Meter will confirm the colour shift.
Thanks for your interest.
Alexey Proskuryakov
Thank you, it is certainly easier to see with the checkerboard.
I can reproduce this behavior with Safari 6.0.5 on 10.8.5. I don't have Safari 5 to test with.
Pokey
I might add that the colour of the graphic is accurate, it is the colour of the html element that is incorrect.
Pokey
It has been a while since I checked but this bug appears to be fixed in the latest WebKit build (Version 7.0.5 (9537.77.4, r176554)).
Simon Fraser (smfr)
Sounds good.