On Thu, Feb 21, 2008 at 10:23 AM, Drew Smathers <[EMAIL PROTECTED]> wrote:

>  Yes, my use of the term resize was little too liberal (maybe analogous
>  to "canvas" resizing in gimp/photoshop vs. image resizing?).  The
>  effect I observed were some missing "black holes" in the mapped
>  texture (the unfilled portions of the larger texture, I presume).  If
>  you have any suggestions on how to gracefully handle this in a pyglet
>  application, I would be very interested and appreciative.  Otherwise,
>  I will continue to force myself to only use power-of-2 images when
>  creating objects in blender.

Only recent(ish) nvidia graphics cards will accept a texture in
GL_TEXTURE_2D with non-power-2 dimensions.  (Some other graphics cards
will accept such a texture but run unusably slowly).  So pyglet never
creates such textures.

In pyglet 1.1 (or SVN trunk) you can specify for the
GL_TEXTURE_RECTANGLE_* target to be used, which permits non-power-2
dimension textures without performance penalty.  The texture
coordinates on these textures are not normalized to [0, 1] though (a
decision made by OpenGL, not pyglet), so Blender-exported texture
coordinates will be incorrect.

It's a fact of all realtime 3D graphics (OpenGL and DirectX) that
textures must have power-of -2 dimensions (besides the performance
benefits, it's the only way to do mipmapping).

Alex.

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"pyglet-users" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at 
http://groups.google.com/group/pyglet-users?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to