Nathan Kopp wrote:

> What I'm wondering about is this:
> Instead of having post-processing built into the core of POV, it is provided
> (with the package distribution) as a separate tool.  The core POV-Ray would
> output a data file containing the 3d data, along with the image file.  The
> post-processing tool would then be used to read in the original image and
> the 3d data and produce a new image file.

This is a very good idea, indeed, and this would correct the main
problem with current implementation, where you have to re-render the
whole image to change the post process !!

> This post processing tool _could_ be considered to be part of the rendering
> tool, since it is part of the package.  

It could be a separate executable, something like "POV-Process" maybe :)
When rendering, POV-Ray would write every information (depth,
normals,..)
to suited files, and you just have to indicate them to "POV-Process",
along
with a processing script (much like the current post_process {..} ).

> It definately could not be
> considered to be a "paint program such as PhotoShop(tm)".  So, would such an
> implementation be acceptable under the current rule?

Of course, if it's part of the package, no objection could arise.

Fabien.

Post a reply to this message