Hi Kevin,
Sorry for the delay. At this point, given what appeared to be a fix on my end for the problem with the previously attached .tgd, I'm not sure what more I can do. I'll try to answer your questions below but I'm not sure it will help much.
Fake Stones are applied as a displacement function over the existing terrain. While logically you would expect that if they were applied last their shape should not change even if the underlying terrain does, that misses some of how this particular shader works. Fake Stones internally uses a 3d voronoi noise variant which keys off of the computed normal of the terrain (normally provided by the Compute Terrain) to generate the stones. It's done this way so that the stones maintain a proper shape and don't get stretched on steep parts of the terrain.
Generally speaking displacement is additive as well so your Fake Stones will be affected by anything above or below them as they are fundamentally a displacement function. Only a population of rock objects would avoid that, but unfortunately those are currently broken.
You shouldn't need to set a vertical offset for a population. As long as it's properly referencing the planet and terrain that it's sitting on it should work. There are some issues with this, but generally speaking it should work and any offset you apply will be a kludge and unlikely to work well, or at least in many situations.
A Noise Variation of 2 actually isn't particularly high. It's the default for a Power Fractal terrain in fact. But here in combination with your other settings it seems to make significant negative altitudes. Which is fine, it's just that they tend to be a little more problematic to deal with. You could try applying an overall positive offset to the terrain displacement to remedy this if you like - the terrain shape will stay the same, it will just rise a couple hundred meters.
You're right that your settings are producing some really cool rocks.
Finally, what I was basically saying is look at the size of the features on your terrain and use that to determine if you need to change the scale at which the normal is computed in Compute Terrain. It's just something you'll have to experiment with I think.
I know that's probably all not much more help, but I hope you can get it sorted out. Some of these things will continue to improve with time and I believe the final release of TG2 should be easier for you to work with in situations like this. In the meantime keep experimenting and reading and asking questions.
- Oshyan
