Is there a way to use the Subdivision Surface modifier on my mesh while keeping the silhouette more or less intact? The Catmull-Clark algorithm reduces the volume and the form of the mesh a lot. I want to create Level of Detail versions that should blend into each other based on procimity/the space they take up in the camera view in my game engine. The following picture hopefully clarificates what I am going for.
The black Wireframe is the original unmodified model. The displayed mesh is the outcome of subdividing with Catmull-Clark algorithm. The red outline I painted in is kind of what I need. The silhouette shouldn't "change", just get a little more smooth/detailed with each level of subdivision.
I tried to Shrinkwrap the subdivised mesh back to a lowpoly copy of itself, "Apply Base" in the Multiresolution Modifier but nothing helps, the volume and silhouette is always compromised. Maybe there is a way to achieve this with some clever combination of modifiers?