Design Review #1030

#1030: [HTML] 2D canvas floating point support

Opened Dec 18, 2024

I'm requesting a TAG review of 2D canvas floating point!

The output bitmap of a CanvasRenderingContext2D (or offscreen) defaults to being 8-bit on all browses (although it is not specified this way). This change makes that behavior explicitly specified, and add the option, via CanvasRenderingContext2DSettings, to specify that it can be 16-bit floating point.

Explainer¹: canvas_float.md
Specification: PR in development
WPT Tests: In development
User research: N/A
Security and Privacy self-review²:
GitHub repo:
Primary contacts:
- @ccameron-chromium, Chromium, co-author of specification
- ColorWeb-CG
Organization/project driving the specification: ColorWeb-CG
Multi-stakeholder support³:
- Chromium comments: None (author)
- Mozilla comments: https://github.com/mozilla/standards-positions/issues/708
- WebKit comments: https://github.com/WebKit/standards-positions/issues/91
Status/issue trackers for implementations⁴:
- Chromium: crbug.com/40277213

Further details:

I have reviewed the TAG's Web Platform Design Principles
Relevant time constraints or deadlines:
The group where the work on this specification is currently being done: ColorWeb-CG
Major unresolved issues with or opposition to this specification:
This work is being funded by:

You should also know that...

A superset of this functionality was reviewed by TAG as part of the proposal for Adding support for High Dynamic Range (HDR) imagery to HTML Canvas. That proposal is being split into several independent components to be added individually.

Discussions

Comment by @ccameron-chromium Dec 18, 2024 (See Github)

The comments for the larger TAG review that are relevant for this the ones relating to toDataURL and toBlob, which have been clarified (both in terms of immediate behavior and future extensions in HDR work) in the smaller explainer.

Discussed Mar 1, 2025 (See Github)

Matthew: I left a comment... looked at it from a perspective of consistency with other things... whether the explainer seems to be missing anything...

... definitley some things that sound reasonable. But one thing : a separate issue thread that I've linked to... They've explained why this format of encoding fits the hardware and matches WebGPU... They have talked about the way they have named stuff. That's a summary of my comment. If we do have someone who is expert, then fine...

Jeffrey: I agree it looks fine and I think Anne's concerns have been solved on the other threads.

we agree to set to proposed closed and close at the plenary.

Comment by @zcorpan Mar 27, 2025 (See Github)

The WebKit s-p comments link to https://github.com/whatwg/html/pull/10951, is that the relevant spec PR?

Comment by @ccameron-chromium Mar 27, 2025 (See Github)

Yes. There is also a related ImageData PR at: https://github.com/whatwg/html/pull/11143

Comment by @matatk Apr 3, 2025 (See Github)

Hi @ccameron-chromium, thanks for your review request. We have had a look (thanks for the clear and helpful explainer), and are happy to see the work proceed, and this gap being addressed.

Thanks to @zcorpan for your help with this review.