Closed GoogleCodeExporter closed 8 years ago
Thanks Rakesh! Can you also include an HTML file that embeds this test SVG file
using
SVG Web?
Original comment by bradneub...@gmail.com
on 28 Oct 2009 at 4:57
I had used the demo.html that comes with the svgweb source code
Original comment by raks...@gmail.com
on 29 Oct 2009 at 6:05
I've tried rendering Rakesh's example locally on XP. Using r964 and
javascript-samples/svg_object.html as a template with Flash (10.0.32.18) and a
variety of browsers - Explorer (8.0.6001.18702), Opera (10.00), Safari (4.0.3),
Chrome (3.0.195.27), Firefox (3.5.2).
None of the browsers displayed as expected (as seen natively on Opera or ASV in
IE).
Repeated attempts on the same browser result in sometimes different rendering - one
arm on the bear, sometimes no paddle. Always, the boat is at center. The
bear's
eyes and mouth never move. Never any movement of the waves. Fairly consistent
CPU
usage between browsers - basically hog the CPU from loading to long after the
animation appears to finish or until the tab or browser is closed.
Rakesh's example initially caught my eye because of the many instances of
animateMotion. I have had a lot of trouble getting the simplest animateMotion
example to render with SVGWeb (still can't). I was hoping that this may be the
problem here too. So I tried removing all the animateMotion elements from
Rakesh's
example and testing it as above. It renders the same as the original but with
more
frames per second - smoother. Still, no waves, boat at center... The CPU usage
is
still high and doesn't stop being high until after closing the tab or browser.
I think animateMotion is a problem with this file but not the whole problem.
Hope
this helps.
Ken
Original comment by k...@svgmaker.com
on 3 Nov 2009 at 5:48
Thanks Ken; did you add ?svg.render.forceflash=true to the URL when you tested
on all
the browsers? If you got differing behavior that might be because the native
renderers did the actual rendering, each of which have differing levels of SMIL
support.
Original comment by bradneub...@gmail.com
on 3 Nov 2009 at 6:40
Hi Brad; All my tests used svg_object.html from the samples/javascript-samples
directory which forces flash via the meta tag.
I removed the following animations and managed to eliminate the "missing"
arms/paddle
behavior.
<animate attributeName="display" begin="Default.begin;Default.repeatEvent"
calcMode="discrete" dur="4" end="Default.end" fill="freeze" keyTimes="0;1.00"
values="none;inherit"/>
In this file it doesn't achieve anything since display is "inline" from the
beginning.
After that, the only differences between browsers that I can easily make out is
the
frame rate of animations. Opera, Safari and Chrome appear the same and smooth.
Explorer is jerky. Firefox appears static (I have intermittent non-SVGWeb
performance problems with Firefox/Flash - others may get different results).
I'll
try the tests on a different (Vista) machine later today.
Original comment by k...@svgmaker.com
on 3 Nov 2009 at 11:28
My assertion my previous comment that animating display was not required is
wrong.
It is required. By removing it, it just made the incorrect animation look ok.
I just checked the original file again with Chrome. A few successive loads of
the
file yielded a different result wrt the the arms/paddle - one arm+paddle, no
arms+paddle, one arm no paddle... The display attribute is animated separately
on
each arm and the paddle.
Original comment by k...@svgmaker.com
on 4 Nov 2009 at 3:56
Original comment by grick23@gmail.com
on 6 Nov 2009 at 3:16
I plan to close this issue soon unless a better example can be provided. The
example
does not demonstrate a performance problem as much as lack of support for
advanced
SMIL features or problems with the test file. Compared to other browsers, I
think the
frame rate for SVG Web looks reasonable. Perhaps Ken's recent SMIL overhaul has
helped.
Original comment by grick23@gmail.com
on 4 Jun 2010 at 3:15
As noted in comment #8, there is no specific, confirmed problem to fix here
that is not accounted for in other open issues.
Original comment by grick23@gmail.com
on 10 Jun 2010 at 4:03
Original issue reported on code.google.com by
raks...@gmail.com
on 28 Oct 2009 at 4:39Attachments: