xmlgraphics-batik-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From bugzi...@apache.org
Subject DO NOT REPLY [Bug 47065] New: Intermediate (alpha) values are clamped inside an feComposite
Date Tue, 21 Apr 2009 16:03:25 GMT

           Summary: Intermediate (alpha) values are clamped inside an
           Product: Batik
           Version: 1.7
          Platform: PC
               URL: http://wiki.inkscape.org/wiki/index.php/Advanced_Gradi
        OS/Version: Windows Vista
            Status: NEW
          Severity: normal
          Priority: P2
         Component: SVG Rasterizer
        AssignedTo: batik-dev@xmlgraphics.apache.org
        ReportedBy: th.v.d.gronde@hccnet.nl

I have been experimenting with simulating non-SVG gradients using (among other
things) feComposite. For this purpose I'm using values for k1-k4 that can cause
intermediate (during the computation of one feComposite) values to go outside
the range [0,1]. For non-alpha components this doesn't seem to present a
problem (see the second attempt at simulating a conical gradient), for alpha
components this goes completely wrong (see the third and last attempt).

I want to stress again that I'm not talking about the values between filter
primitives (I'm carefully making sure that they stay within the [0,1] range),
I'm talking about the computation k1*i1*i2 + k2*i1 + k3*i2 + k4. If, for
example, this would be negative without k4, but k4 IS used to make sure it
doesn't, then this works just fine with a non-alpha channel, but it doesn't
work with the alpha channel.

This seems odd and is frustrating because the alpha channel is very suitable
for these kinds of computations, as you then don't have to worry about using
premultiplied values.

Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

To unsubscribe, e-mail: batik-dev-unsubscribe@xmlgraphics.apache.org
For additional commands, e-mail: batik-dev-help@xmlgraphics.apache.org

View raw message