- From: David Dailey <ddailey@zoominternet.net>
- Date: Thu, 23 Jan 2014 13:37:26 -0500
- To: "'Tab Atkins Jr.'" <jackalmage@gmail.com>
- Cc: "'Dirk Schulze'" <dschulze@adobe.com>, "'www-svg'" <www-svg@w3.org>, "'public-fx'" <public-fx@w3.org>
on: Thursday, January 23, 2014 10:39 AM Tab Atkins Jr. writes On Thu, Jan 23, 2014 at 6:55 AM, David Dailey <ddailey@zoominternet.net> wrote: >> <linearGradient> >> <stop offset="0" stop-color="url(image1)"/> <stop offset="1" >> stop-color=" url(image2)"/> </linearGradient> >Gradients, while 2-dimensional, have a 1-d data model - they're colors >(0d) painted on a line, which extend from the line perpendicularly to fill the plane. >(Or, for radial gradients, extend from it and loop around in an ellipse.) >As such, painting images in a gradient doesn't make much sense, since images are 2d to >start with. >I suppose you could think of it as painting the line (and the extended >perpendicular) with the corresponding pixels from the image blended with the given amount >of progress, but that's reaching a bit. Well, reaching is a relative thing I suppose. That's what I had in mind. Transitions between an image which blend gradually (along a line) into a color might be more common than transitioning between N images, though a reflected gradient that has its stop-offsets animated and which paints transitions between two images would be a very nice effect. Both effects would normally be done with a mask, just as <someshape fill="image"> would normally be done with a clippath. So long as one is extending the model, why not make it transformative? Cheers D
Received on Thursday, 23 January 2014 18:38:04 UTC