views:

44

answers:

1

I'm starting to design my object graph for an OpenGL ES 1.1 app on Android.

What's advisable when it comes to actually drawing my triangles and controlling their size? Should I store all the vertices in units relative to each other and then multiply them by a value I pass during a draw(GL10 gl)?

Would this approach have any impact when it comes time to do frustum culling?

+1  A: 

When you have full control over all your model/vertice data you shoul try to create each one in the correct size to each other and in relation to your design concept. Thereby you don't need to call any glScale functions which take up processing power. Frustum wise it doesn't make a difference whether you scaled before or not. By scaling you merely modify your modelview matrix.

Soulreaper
Right now I've got a static "size" variable that I'm defining on my rendering classes that is then multiplied by each point when setting up my vertex buffer array.Both variables are static and are only calculated at class load time, but this way if I increase the generic "size" value in the class (or hook it onto a more broad design), the vertices scale accordingly and the calculations are cached. Does this make sense?
Omega
I see your design concept. Whether that is required is up to your implementation i'd say.
Soulreaper