- From: Sebastian Zartner via GitHub <noreply@w3.org>
- Date: Mon, 09 Feb 2026 11:04:08 +0000
- To: public-css-archive@w3.org
> You can reach any value, it's just about how we define the behavior. For example, we could define that the slope for some filters for values < 100% is negative.
Can you expand on this? E.g. imagine a filter of `brightness(200%)` and you want to apply a mask that goes from 0 to 1, i.e. a brightness value between 0% and 200%. How would that look like?
> I think it's more important to be consistent about the mask-image, so that 0 is always no effect, rather than 0, and 1 is full effect. Otherwise, your no-effect-value on the mask can vary depending on the specified value of the filter, and that I think makes the mask-image less reusable between different filters/values.
A 0-value behavior across filters meaning no effect does make some sense. Though I don't see how you can reach any value with that approach.
We could also define additional parameters defining the range of the multiplier the mask should cover. So something like
```ebnf
mask() = mask( <mask-reference># <integer>{2}? )
```
where the first `<integer>` defines the lower bound and the second the upper bound of multiplier. So a fully transparent pixel in the mask corresponds to the lower bound and a fully opaque pixel the upper bound.
Though the 0-means-no-effect idea and my proposal above only apply to filters with a single parameter. `drop-shadow()` and especially custom filter effects aren't reasonably covered by that.
And allowing to define the parameter that is influenced by the mask is definitely out of scope for an MVP. Though in case of `drop-shadow()` I think it's a reasonable default to let the value only apply to the standard deviation value, meaning a mask value of 0 does _not_ mean no effect.
Sebastian
--
GitHub Notification of comment by SebastianZ
Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/13288#issuecomment-3871053594 using your GitHub account
--
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Monday, 9 February 2026 11:04:09 UTC