I'm finding that average normals adjusts all selected vertex normals. I'm looking for something to blend between the 1st and last selected vertex normals, leaving these alone, but blending any intermediate normals. The result, I expect, would be a smooth transition from the first to last selected vertices. I am not interested in a solution that requires first locking the 1st and last vertex normals, then doing the average - there's only so many hours in a day.
If the average normal tool/function used the first and last vertex normals in it's calculations but did not change those normals, it would do what I want to accomplish.
If nothing exists, I assume the process would look something like:
a) get selected vertices,
b) for each vertex, get attribute normal
c) calculate normal deltas between all normals
- each component averaged on it's own, or is there some function to average
d) update all selected vertices _except_ first, last
By vertex, do you really mean vertex, or do you mean the set of vertices at the same location? For instance if you selected two corners of a cube, would you consider those two vertices or would you consider those six vertices? From your image, I see two sets of vertices selected each with 5 coincident vertices. I'm not sure if two of these 10 vertices represent the vertices with the normals you want to keep unchanged.
Can you describe a little bit of the effect you want to end up with? That might help me to understand what the script is supposed to accomplish.
Ah. I mean the collection of vertices then taken collectively per unique 3d position:
If the average normal tool/function used the first and last vertex normal [collection]s in it's calculations but did not change those normals, it would do what I want to accomplish.