Re: [css-images][css-compositing] Blurring an element’s backdrop

On Thu, Nov 28, 2013 at 11:54 PM, Alan Gresley <alan@css-class.com> wrote:

> On 29/11/2013 5:35 PM, Lea Verou wrote:
>
>> On Nov 29, 2013, at 12:45, Rik Cabanier <cabanier@gmail.com> wrote:
>>
>>> This might work. My main concern is that it would overload
>>> mix-blend-mode too much since we're also hoping to repurpose it to
>>> blend different areas of an element. I'm unsure how we would
>>> reconcile that. Do you have a suggestion?
>>>
>>
>> If the syntax for that is what was discussed in the FXTF a while ago,
>> I don’t see what the conflict would be. Care to elaborate? :)
>>
>> ~Lea
>>
>
> From what I have gathered from this thread, it not the syntax but rather
> how do you blur a backdrop if the filter causes isolation (staking context)
> and how does this blur less as you go outwards from a given element.
>

The filter/blend does cause stacking, but the backdrop is not affected by
that. (stacking is even needed to have a backdrop because otherwise you
draw according to appendix E [1])


>
> Take this demo where it animates both opacity and a blur together. Only
> works in Chrome.
>
> http://css-class.com/test/css/3/filter/blur-opacity-cross-fade.htm


This demo is not working for me in Chrome 31 on mac (but it could be
Mavericks as Chrome is very flaky since I installed it)


> What would or should happen over a non uniform backdrop (currently just a
> background colour of black) when we have opacity(0%) and blur(1em)?
>

I assume the blur is for the backdrop and the opacity for the elements?
If opacity goes to 0, it should be as if that element is invisible. So, the
area that is covered by the element is untouched/not blurred.


>
> Chrome itself is not following the spec since when opacity has a value of
> '1', there should not be a stacking context but if it did not disobeyed the
> spec, it would animate strange. Anyway, the bug is not present in this demo
> since the images are absolutely positioned.


1: http://www.w3.org/TR/CSS21/zindex.html

Received on Saturday, 30 November 2013 03:34:46 UTC