Faster method for rendering anisotropic textures?

Is there any faster metho开发者_高级运维ds for improving texture quality on oblique angles than using anisotropic texture filter?

In my previous question i asked if it takes more memory than normal mip mapped textures, but apparently it doesnt, so i am hoping this could be optimized.

So is there any way to optimize the rendering speed ?

No. The alternatives such as supersampling are significantly slower. Your best bet on OpenGL is using the anisotropic filter.

Regarding the question of pre-calculating it. Yes, you theoretically can precalculate the anisotropic filtering. It would almost certainly never be useful to do so. Because the sampling and filtering of a texture are dependent on your view of it, it would only be something that you would do if the result were intended to be perfectly static onscreen with only a view that is known in advance. If that's true, then you probably wouldn't be drawing it in 3D with OpenGL in the first place.

Most modern hardware can do that sort of filtering pretty quickly, but if you really are getting unacceptable performance because of it, you can try disabling it to optimise the speed. I imagine that you will likely be more disappointed in the quality of the rendering that impressed with the increase in speed.