Re: CSS3 transition strangeness when using background-color:transparent;

On Jan 29, 2010, at 9:43 am, L. David Baron wrote:

> On Friday 2010-01-29 09:25 -0500, Boris Zbarsky wrote:
>> 624       // FIXME (spec): The CSS transitions spec doesn't say whether
>> 625       // colors are premultiplied, but things work better when they are,
>> 626       // so use premultiplication.  Spec issue is still open per
>> 627       // http://lists.w3.org/Archives/Public/www-style/2009Jul/0050.html
> 
> I feel somewhat strongly at this point that animation of colors
> should be in premultiplied space; this was mostly after
> experimenting with the results of both on:
> http://dbaron.org/css/test/2009/transitions/transitions-alpha
> Doing this does remove some author control, as I said in
> http://lists.w3.org/Archives/Public/www-style/2009Jul/0048.html ,
> but I think the tradeoff of getting expected results is worth it.

I agree that animating colors in premultiplied space gives better results
most of the time.

> (Another related point is that animations of colors should probably
> honor the 'color-interpolation' property in
> http://www.w3.org/TR/SVG11/painting.html#ColorInterpolationProperties .)

Color interpolation should probably match gradients too (which implies
that CSS gradients should also honor 'color-interpolation').

Simon