> 
>         Actually I would favor having one executable for each post process
> filter (with a common library of helper functions probably). It would be
> much more flexible and extensible (you wouldn't need to recompile the
> whole thing to add a new filter) and probably easier to use too.
> 

> 
> PS: Shouldn't this thread go to unofficial.patches?
> --

Somewhat off-topic for this group my message will be, so I set followups
to unofficial.patches (this is the very first time I use this feature,
so probably I just make mess of it ;-)

With all this post processing I thought that for implementing it similar
approach to my shader patch could be used: user writes post-processing
filter in "shading language" (C-alike syntax with common graphical
operations and types: color, point, normal, vector), compiles it into
byte-code and it is "added" to scene in POV-Ray script for
post-processing. This way it will be easy to create new filters and
change existing ones. This will not be as fast as built-in processing
(but I think that during post-processing it will not be issue), but it
allows greater flexibility. RenderMan (R) uses similar approach with the
imager shaders (
http://www.pixar.com/products/rendermandocs/toolkit/RISpec/section12.html#Imager.shaders
)
I had no time to examine post-processing and I don't know, what is
needed (in terms of POV-Ray internal variables/values/etc.) for running
post-processing filter, so if someone can sum it up or give me a link to
document, then I can say, whether it could be done easily with some
modifications of my (not-yet-released, i.e. vaporware;-) shader patch.

Post a reply to this message