Is there a ratio of pixels to distance units in Carrara when you have created a terrain from a heightmap JPG? This seems especially problematic if the size/height of the terrain primitive in altered in Carrara. I find it gets in the way of building a decent terrain shader.
I often use Daylon Graphics’ Leveller tool but also frequently tweak things up in Photoshop to create my heightmap. Once I have my Carrara terrain, I use an opaque white plane object that is larger than the terrain so that I can note interesting height values for the Z axis.
What I still find confusing is the Blend feature of the terrain shader. I understand what it is for conceptually, but it is a bugger to find values that work for me without doing a lot of test renders.
The first image was early in my work on this scene. I have surface replicated stones along the riverbed. The surface of the riverbed and the banked slopes should be looking muddy and brown. But I was getting a lot of “Blend leak” with the grass part of the shader extending far down the bank (a little bit would be nice, but not that much).
The second image is more or less done—I’ll take what I have learned and go on to make better scenes like this. Although I cured the “Blend leak,” my only choice was to remove it entirely; so there is an unnatural hard cut-off between the muddy slope and the grassy plane. One thing I am a bit proud of is combining the replicated rocks with a texture map of sandy gravel that is constrained with altitude to keep it at the river bottom. I think that turned out rather nicely.
Critiques and suggestions for working with terrains and terrain shaders are gratefully welcomed.
