[Webkit-unassigned] [Bug 216943] New: Chroma subsampling based on encode quality

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Thu Sep 24 14:36:28 PDT 2020


https://bugs.webkit.org/show_bug.cgi?id=216943

            Bug ID: 216943
           Summary: Chroma subsampling based on encode quality
           Product: WebKit
           Version: Safari 14
          Hardware: Macintosh
                OS: macOS 10.15
            Status: NEW
          Severity: Normal
          Priority: P2
         Component: Canvas
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: micah.millereshleman at gmail.com
                CC: dino at apple.com

Hi,

One unresolved browser inconsistency in the Canvas API is how image encoding quality affects chroma subsampling. 

In Firefox, chroma subsampling is turned off at quality >= 0.9, allowing crisp images to be exported from Canvas at relatively small file sizes (e.g. quality 92%). Chrome & Webkit only disable chroma subsampling at quality == 1.0 (100%), meaning that it's impossible to export a crisp canvas as a JPEG without also exporting a massive image.

I work at an online photo editor, and it means that we recommend Safari users save their images at 100% quality, and then use a different program to compress them (w/out chrome subsampling) before using them on a website. Other online photo editors bypass the browser altogether and load their own version of libjepg in WebAssembly. This seems a bit heavy-handed, and it'd be great if the browser default for Canvas produced images that were ready to use on the web.

Chrome recently opened a WHATWG issue for this, and it'd be great to hear from Safari. 
https://github.com/whatwg/html/issues/5395

Improving this small corner of the web platform would make Safari a better platform for our users. Thanks!

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20200924/e23f5aab/attachment-0001.htm>


More information about the webkit-unassigned mailing list