[scilab-Users] Re: [Scilab-Dev] Scilab 5.2 Beta available
Stéphane Mottelet
stephane.mottelet at utc.fr
Fri Nov 6 15:30:40 CET 2009
Stéphane Mottelet a écrit :
> Pierre LANDO a écrit :
>> Stéphane Mottelet a écrit :
>>> Hi all,
>>>
>>> just a question/remark. Since version 5, we have a powerfull rendering
>>> engine for graphics, but I still do not understand one particular point
>>> about interpolated shading. I know that Scilab still works with an
>>> indexed color scheme, but why interpolated shading still works by
>>> cuting triangles into constant color patches ?
>>
>> Maybe for being compatible with scilab 4, but I'm not sure at all...
>>
>>> This generates artifacts when exporting figure to ps,pdf,svg,...
>> You don't have the same result because it's not JoGL who make the
>> exported files (in vectorial format).
>> You better should be happy to have something viewable when you export
>> in vectorial format :o)
> Well, I thought that most "modern" vectorial formats (pdf, svg,...)
> implemented natively interpolated color patches, but after a quick
> search on
> the net, I realized that these format only take into account
> gradients. At least
> with svg, this could be a way to emulate interpolated shading for
> triangles.
People have already given the answer :
http://www.svgopen.org/2005/papers/Converting3DFaceToSVG/index.html#S3.
S.
>
>> (see http://bugzilla.scilab.org/show_bug.cgi?id=4866).
>>
>>> I think that JoGL should be able to use hardware bilinear shading,
>>> even with indexed colors (instead of TrueColor).
>> You are right, (modulo drivers problems), actually the 'indexed
>> color' is see by JoGL as a 1D-texture, nothing simplest as activating
>> bilinear shading for the OpenGL texture.
>>
>> But, it's look like that some people don't want to see in a graphic a
>> color who doesn't exist in the colormap (very old trauma related to
>> VGA display and is 0x13 mode).
> Ok, but, Matlab has no problem to render interpolated colors with an
> indexed color visual. This should be doable in Scilab ?
>
> S.
>
>
More information about the users
mailing list