- From: Alex Menzies <amenzie@us.ibm.com>
- Date: Mon, 26 Jun 2006 16:01:55 -0600
- To: www-svg@w3.org
- Message-ID: <OF1CF7A4C3.D8A0E871-ON87257199.007888C5-06257199.007E32EB@us.ibm.com>
I have a question about the correct behavior of the Gaussian Blur filter
when a very small stdDeviation is used. The 1.1 SVG specification
suggests using 3 box blurs with a size given by:
boxSize = floor (s * 3 * sqrt( 2 * pi) / 4 + 5)
in order to approximate a Gaussian Blur for s >= 2.0. I am curious if
anyone has some suggestions about how to handle the case when s < 2.0. One
might be tempted to just assume the effect of the blur with a low value
for s to be negligible, but such is not the case. The following code
shows an (extreme) example in which the light blue circle is blurred with
a stdDeviation of 0.001 and scaled to a size where the blurring effect is
noticeable (when rendered in Opera9beta, Batik, or IE Adobe plugin). The
yellow circle shows what an unblurred edge looks like for comparison.
Thanks!
-Alex
<svg xmlns:xlink="http://www.w3.org/1999/xlink"
xmlns="http://www.w3.org/2000/svg">
<defs>
<filter id="smallBlur">
<feGaussianBlur in="SourceGraphic" stdDeviation="0.001"/>
</filter>
</defs>
<circle cx="0" cy="0" r="1" fill="lightblue" filter="url(#smallBlur)"
transform="scale(1200)"/>
<circle cx="0" cy="0" r="1" fill="yellow" transform="scale(1100)"/>
</svg>
Received on Monday, 26 June 2006 21:58:39 UTC