[Radiance-general] rvu: too many modifiers in ambient list

Lars O. Grobe grobe at gmx.net
Mon Jul 11 19:54:20 CEST 2005


Hi!

Once again, maybe I misunderstood, maybe I just wrote a misleading 
description what I wanted to achieve by including the material 
definitions (at least the colorpict modifier's definition) into each 
object file.

I have radiance-files describing geometrical objects, like e.g. a plane 
of marble. This "object" is placed by replmarks, which brings it to its 
place moving and rotating it in space. Now, as the object is made of 
marble in my example, it also has a surface structure which is to be 
represented by a picture, used by the colorpict modifier. Both, 
geometry and image, are placed at (0;0;0) on the x-y-plane originally, 
and translated by replmarks.

So far, I either used instances with replmarks or included the object 
geometry with the colorpict modifier's definition. As such, the xform 
operations performed by replmarks were also applied to the colorpict.

Now, having only the alias definition left in the object file, the 
image map is not translated together with the object itself any more. 
So the object moves somewhere in space, while the image map, staying at 
(0;0;0) on xy, visually disappears because it no longer touches its 
geometry.

Does that mean that I have to stay with the redundant definition of 
colorpict modifiers? Using only instances gives bad overhead with such 
simple geometry, objects with only the alias included don't show the 
mapped image when moved or rotated. Using object geometry with included 
colorpict definition gives my the "ambient list" message that I 
initially described when beginning this thread. For future projects: 
using u/v with mesh primitives would prevent all that trouble, right?

So for now, I think I will have to stay with the redundant material 
definitions included in object (geometry) files.

Thank you for all the help, CU Lars.




More information about the Radiance-general mailing list