Re: Export Gif animation bug?
- To: mathgroup at smc.vnet.net
- Subject: [mg39159] Re: [mg39148] Export Gif animation bug?
- From: Dale Horton <daleh at wolfram.com>
- Date: Fri, 31 Jan 2003 19:57:18 -0500 (EST)
- Sender: owner-wri-mathgroup at wolfram.com
The color problems are in the colormap of the GIF. There are
ConversionOptions that control the colormap. Either can be used to fix this
problem.
By default, a single colormap is used for the entire animation. You can
create a colormap for each frame with
Export[..., ConversionOptions -> {"GlobalColorReduction" -> False}]
Or you can override the automatic colormap with one of your own. Here's an
example using "web safe colors".
Export[..., ConversionOptions -> {"ColorReductionPalette" ->
Table[RGBColor[i, j, k], {i, 0, 1, 0.2}, {j, 0, 1, 0.2}, {k, 0, 1, 0.2}]}]
At 03:37 AM 1/31/2003, Don Taylor wrote:
>I think I've found a bug in exporting GIF animation
>(Mathematica Version 4.0.2.0 under Windows 98 with MS patches)
>
>I don't know what the problem is with this. I've trimmed down what
>I'm doing to what seems to be the bare minimum that shows the bug.
>Tiny changes in the value of z or in the step size for the table
>will make the background of the gif dark black or brilliant green,
>change the resulting file size by a factor of 16x, etc. These three
>examples show a sample of the range of behavior that I've seen with this.
>
>This simulates some axes and then draws a moving point which is
>colored green at the bottom of each cycle and black elsewhere.
>
>All these examples are identical code, except for the value of z.
>It takes a couple of minutes to generate each example.
>
>111 kbytes and what seems to be pretty much a correct plot
>
>z = 1; Export["1.gif", Table[Graphics3D[{
> Line[{{0, 1, 0}, {0, -1, 0}}], Line[{{0, 0, z}, {0, 0, -z}}],
> Line[{{0, 0, 0}, {4Pi, 0, 0}}],
> If[Sin[t] <= -0.8, RGBColor[0, 1, 0], RGBColor[0, 0, 0]],
> Point[{t, Cos[t], z Sin[t]}]},
> PlotRange -> {{0, 4Pi}, {-1.05, 1.05}, {-1.05z, 1.05z}},
> ImageSize -> 800, Boxed -> False], {t, 0, 4Pi, Pi/16}], "GIF"]
>
>1598 kbytes! with bright green graph and lots of artifacts scattered
>across the entire graph as it displays
>
>z = 2; Export["2.gif", Table[Graphics3D[{
> Line[{{0, 1, 0}, {0, -1, 0}}], Line[{{0, 0, z}, {0, 0, -z}}],
> Line[{{0, 0, 0}, {4Pi, 0, 0}}],
> If[Sin[t] <= -0.8, RGBColor[0, 1, 0], RGBColor[0, 0, 0]],
> Point[{t, Cos[t], z Sin[t]}]},
> PlotRange -> {{0, 4Pi}, {-1.05, 1.05}, {-1.05z, 1.05z}},
> ImageSize -> 800, Boxed -> False], {t, 0, 4Pi, Pi/16}], "GIF"]
>
>72 kbytes with entire graph nearly completely black
>
>z = N[\[ExponentialE]]; Export["e.gif", Table[Graphics3D[{
> Line[{{0, 1, 0}, {0, -1, 0}}], Line[{{0, 0, z}, {0, 0, -z}}],
> Line[{{0, 0, 0}, {4Pi, 0, 0}}],
> If[Sin[t] <= -0.8, RGBColor[0, 1, 0], RGBColor[0, 0, 0]],
> Point[{t, Cos[t], z Sin[t]}]},
> PlotRange -> {{0, 4Pi}, {-1.05, 1.05}, {-1.05z, 1.05z}},
> ImageSize -> 800, Boxed -> False], {t, 0, 4Pi, Pi/16}], "GIF"]
>
>I searched newsgroup postings and Wolfram and didn't find any mention
>of such behavior. I presume this is a bug in the gif export code.
>
>If anyone can confirm that this is a bug and that it has been fixed
>that would be good. If anyone can describe a dependable workaround
>for the moment that would be even better, waiting a good part of
>an hour for an animation to grind out and then discovering that it
>is another "black cat in a dark room" is getting a little old.
>
>Thank you
-Dale