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

On Sep 19, 2013, at 6:25 AM, Rik Cabanier <cabanier@gmail.com> wrote:

> 
> 
> 
> On Wed, Sep 18, 2013 at 9:06 PM, Dirk Schulze <dschulze@adobe.com> wrote:
> 
> 
> On Sep 19, 2013, at 5:28 AM, "Rik Cabanier" <cabanier@gmail.com> wrote:
> 
>> 
>> 
>> 
>> On Wed, Sep 18, 2013 at 8:15 PM, Brad Kemper <brad.kemper@gmail.com> wrote:
>> 
>> On Sep 18, 2013, at 4:33 PM, Rik Cabanier wrote:
>> 
>>> Blurring of the backdrop is not quite the same as a CSS image.
>>> You typically only want to blur where you're actually painting. For instance, if you draw large text or an SVG graphic, you only want to blur where the pixels of the text or graphic are drawn.
>>> 
>>> I guess you could pull in the backdrop as a rectangular region of a certain size, but seems too primitive for CSS.
>>> 
>> 
>> The backdrop is whatever is behind the element, right? That can change, due to scrolling, resizing, dragging, animating, etc., and you'd want the blurring to continue to blur whatever is behind it.
>> 
>> Yes, since this is all declarative, the effect should look correct if the content or its backdrop change. (Just like it should with blending) 
>> 
> 
> Exactly that was my intention with 'backdrop'. It always references what is behind the element. (And that might change by transforming elements or scrolling.) The effect is of course limited to the background size and position of the element. This gets you exactly the possibility to blur or blend the areas that are exactly behind your element. I don't think this is to primitive though.
> 
> So, is the area of the backdrop that is filtered rectangular, or is it the shape of the content you draw?
> SVG's backgroundimage primitive will pull in a rectangular area so if you want it to just affect the content, you'd have to calculate the coverage of the element and use that as a mask/clip

I suppose we would use the "background painting area", which indeed is rectangular.

Greetings,
Dirk

> 

Received on Thursday, 19 September 2013 05:55:46 UTC