I would like to understand how anisotripic filtering works in gpu. Specifically what is the minimum state in which the texture needs to be in for anisotropic filtering to work. by state of the texture I mean the content(mip level data) of the texture. For example if I try load just to 0th mip for the texture and use anisotropic filtering will it work or the full mip chain needs to be loaded for anisotropic filter to work correctly(no artifacts in output). The information is needed as I am working with sampler feedback streaming so I want to make sure I do not miss any cases which might lead to artifacts(namely I will start the sparse texture will just the least detailed tail mips loaded,so will aniso filter work correctly with it or not)?
$\begingroup$
$\endgroup$
0
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Anisotropic filtering (and any other filtering process) will respect the base/max level settings for the texture they're fetching from. So as long as those settings are properly provided, it will never fetch garbage data if the mipmap pyramid is incomplete.
Anisotrpoic filtering will be more effective (ie: look better given a particular sample count) if more of the pyramid is available. But that's about it.
