[Bf-committers] Envmaps

[Alexander Ewering]

On Thu, 28 Aug 2003, Douglas Bischoff wrote:

> Beautiful! Is this commit-able? What'd you do to get it to handle this 
> properly?

Sure, it should be commitable... I was just argueing with ton about the
additional memory needed by an Object* pointer in the VlakRen struct :)

First, the old bug that kept objects with envmaps from showing up in
other envmaps was caused by Blender's automatic function to hide objects
with envmaps from envmap renders (because if they were rendered, the
surface would just show the _inside_ of the object).

The problem was that _all_ objects were turned off, and then all envmaps
were rendered.

I've added an Object pointer to the render face struct (VlakRen) so that
the envmap code knows what object a face belongs to, and can selectively
hide objects from render _per_ envmap.

And then, the recursion... I've simply put the make_envmaps stuff in
a for() loop, and made it re-render the envmaps several times (and ignore
the fact that they are already rendered). I have to work a bit more on this,
but it basically works.


| alexander ewering           instinctive new media
| ae@instinctive.de       http://www.instinctive.de 
|
| fon: +49-2393-220558         fax: +49-2393-220559