|
![](/i/fill.gif) |
On Sat, 22 Jan 2000 22:30:59 +0100, Fabien Mosen wrote:
>Am I right if I say :
>The required skills to program a shader are almost equal to
>the skills required to program a new texture/pattern into the
>core Povray code ?
Pretty much, yeah. Maybe more, if the trend goes toward self-
antialiasing shaders a'la those that Larry Gritz has made.
>After all, and if I understood well, renderman shaders must
>be compiled with an ordinary C++ compiler, and they are -sort of-
>dynamically linked to the renderer (who gives some data to
>the shader, and get a pixel color or some other data in return).
Not exactly. They're compiled into bytecode (IIRC) by the shader
compiler, which is not really a C compiler. They're then executed
by the renderer pretty much as you say.
>About UV-stuff, the renderers that have shaders usually have only
>a few types of primitives : nurbs, bicubic things,... And we know
>that some Pov privitives CAN'T be UV-mapped : julias,...
Right. RIB only allows for a small subset of objects with easy UV
mappings. But even some of those objects don't have UV code written
yet in MegaPOV. One could always warn that a shader that expected
uv coordinates wouldn't work properly on a non-UV-enabled object.
--
These are my opinions. I do NOT speak for the POV-Team.
The superpatch: http://www2.fwi.com/~parkerr/superpatch/
My other stuff: http://www2.fwi.com/~parkerr/traces.html
Post a reply to this message
|
![](/i/fill.gif) |