This is basically a MR description but this commit covers a lot of things so I'll include it so you don't have to look at the MR:
This allows us to use BC6H and BC7 textures as is, provided the used OSG revision and hardware support them
DDS images are unflipped by default, non-DDS bottom-left convention images are flipped (TGA, BMP, PNG, JPG)
3D DDS images always used top-left convention if they were used in post-processing
Numerous things flipped in this commit to guarantee the same visual results are:
- Terrain UV and blend maps
- Water UV, normal map and simple water ripple alignment
- NiTriShapeData and BSTriShape UV (unflipped), NiUVController vertical offset
- NIF particles are NOT flipped, they remain flipped for Morrowind compatibility
- GUI widget texture UV and convention
- Cursor
- Character preview inventory doll and race menu head UV and display
- Loading screen scene image UV
- Local and global map UVs (fog of war, unexplored and explored maps)
- Saved game screenshot UV
- Video widget UV
- Rain particle alignment
- Sun and moon UV
- Tangent space handedness (bitangent direction)
- Parallax vertical offset
Things where this is a breaking change
- Collada geometry UV (Example Suite and anything else affected must adopt the changes)
- Compressed non-S3TC images that should but can't be flipped (this is KTX which we don't technically support yet)
- Any forward pass shader replacements that currently still use the original tangent space handedness and parallax convention
This PR aims to solve all issues with `Groundcover` view distance handling in a satisfying way while preserving other optimisations that benefit other features. The main idea here is not to rely on `ViewData` updates for distance culling calculations because we can not accurately determine distance against lazily updated views. Instead, we perform an accurate measurement in a cull callback we can run every frame in `Groundcover` itself. While we do have to add some code to handle this feature, it is quite loosely coupled code that could be useful elsewhere in the future. These changes should address a performance regression @akortunov experienced.