- From: Greg B via GitHub <noreply@w3.org>
- Date: Fri, 18 Jul 2025 15:07:42 +0000
- To: public-css-archive@w3.org
Thank you, Chris. Just to be complete, here would be an adaptation of the HTML I originally shared to view with the background more accurately matched to the image. ``` <div> <img src="./stripes-12bit-PQ-P3.avif" alt="linear32 (P3 to best match sample photo)" style="width:100vw; height:100vh; background: color(display-p3 1 0 0);" /> <img src="./stripes-12bit-PQ-P3.avif" alt="linear32 (sRGB, not matched)" style="width:100vw; height:100vh; background: color(srgb 1 0 0);" /> <!-- no CSS color spec for Rec2020 testing --> </div> ``` This does not change the conclusions: there is significantly browser variability for alpha and the results are not as expected. It would be ideal for standards to clarify use of alpha as HDR further expands the potential variability in transparent images given a much wider range of potential EOTFs. Based on results I've seen, I suspect visible banding of solids/gradients for 10-12 bit images may also occur when the encoding / decoding of the alpha channel are not consistent (unsurprisingly, as this is akin to editing a finished image and risks quantization). -- GitHub Notification of comment by gregbenz Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/12492#issuecomment-3089799375 using your GitHub account -- Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Friday, 18 July 2025 15:07:43 UTC