I think that I have the idea of subsurface scattering down fairly well now. In terms of modeling an octopus or ant, the transparency and internal scattering have more effect than the surface. This example is a blue star behind a simple icosahedron sphere converted to mesh and set with an index of refraction of 1.15 and a transparency of 0.814 with sub surface scattering in the red only. I added some mirror reflection to simulate the effect that a layer of chitin has on the appearance and with the other techniques of hair, particle, and proper meshing of the model, I think I can get a complete GENE to phenotype modulation working efficiently so that I can merely modify a specific gene effect and get the various species of ants.
I have looked at some ant models in cad and they all suffer from texturitis. They think that by applying a surface texture it can make the model realistic and it does not. It is also time consuming to match a texture to a surface, especially when the surface could change based on the changes in the model itself and its deformation.
I have looked into the internal structure and how the muscles bind and it is obvious that there are some specific redundant elements that can be duplicated for realism sake. I also set up an IPO curve for a rotating light source so I could just select one of 100 orientations of the light for the scene. That seems to be a big advantage for experimenting and seeing which methods produce the best results. It looks a bit like a gum drop or amber ( which makes me think I will make an ant in amber when I am done).
0 comments:
Post a Comment