OK, when creating a mesh (triangles) and defining my simple shape - a rectangle for example - what is the best approach on sizing. That is, should I make a big mesh so it is big when placed in the scene Or should I make the mesh small and then scale it or what ever during the scene
One instance of this is putting up a "flat" background image into the scene to serve as the backdrop. I made a "large" mesh (-5, 4, 0 to 5, -4, 0) and then placed in in the scene. Of course the farther (Z axis) I needed to place it in the large the mesh needs to be. I can use scaling to then enlarge the mesh, but detemining the scale to use was trial and error for me.
Am I doing something incorrectly, or the hard way (aside from using a tool like Blender)
Thank you.
Mesh size versus scaling
MDesigner
Man, I'm learning all kinds of good stuff writing this game. Thanks for the information.
So just to make sure I understand you correctly.
1) I define the mesh vertices using the a format different than I am using now (PositionNormalTextured) to the format TransformedTextured. I would make the triangle as 0,0 to 800,600
2) I would set the VextexFormat to Transformed and set the EnableAutoDepthStencil to false before rendering the background.
3) Reset depth buffer after call for remainder of the rendering.
Thanks.
Fanel SAVA
Yep, though the reset of the depth buffer is done when you call Clear. The EnabelAutoDepthStencil jsut makes sure that nothing gets changed when you draw the background. Hence everything will always get drawn over the top of it.
jasontang
As always, it depends... but guessing scalle transforms to fill a screen is not a good thing
Scaling by setting the world transform is effectly free since the GPU has to do the multiplication anyway. But you will get into problems with your lighting becuase the normals all get scaled so you have to set the NormalizeNormals renderstate to counter that which is extra overhead.
If you want a backdrop that filles the screen then the best way is just to use transformed coordinates and write it using screen coordinates with the zbuffer turned off. Then turn the zbuffer back on and everything will be in front of it - no worries about scaling. If your backdrop needs to be able to mvoe in 3d as you pan/tilt etc then you should use a skybox technique which will do much the same thing in 3d.