Re: Object graphics output police (iImage) [message #44091 is a reply to message #44090] |
Fri, 13 May 2005 09:41   |
Karl Schultz
Messages: 341 Registered: October 1999
|
Senior Member |
|
|
On Fri, 13 May 2005 09:34:03 -0600, David Fanning wrote:
> Randall Skelton writes:
>
>> Can someone explain if there is a way to 'fix' the iTools problem
>> demonstrated in the code below. The problem is that the image plotted
>> in the iMap call does not export to a vector postscript cleanly
>> (File->Export->'To File'->myFile.eps). Instead of getting a white
>> masked area outside of the map projection, I get black in the eps file.
>> None of the bitmap outputs suffer the same. I've tinkered with the
>> CHANNEL and CLIP_PLANES keywords but was unable to make this go away :(
>
> Sadly, this is pretty much what I would have expected
> in this case. The truth is, the vector implementation
> is just limited when it comes to textured overlays,
> which is what I think this is. :-(
>
I think that is correct. The image uses an alpha channel to mask out the
parts not occupied by real data.
Vector devices don't have depth buffers and have other limitations. Many
do not support alpha, and that's what is happening here. Not all
PostScript versions support alpha. The IDL docs do mention that there are
restrictions and that vector output is most valuable for simpler graphics.
Why are you (Randall) using vector output here anyway? If you are
generating output where image data is covering 100% of the scene, you are
going to end up putting about the same number of bits into the PS file for
bitmap and vector output anyway. So there is no size advantage to using
vector output. In fact, in a case like this one, the vector output file
might be larger because the code to draw the lines and text would have to
be in the file along with the image bits.
Karl
|
|
|