An Iterative Approach to Simulation of Organic Growth on Surfaces Using Displacement and Procedural Textures
GrowthNodes is a Blender plugin for generative content creation and simulation of organic growth processes on polygonal surfaces. It can be utilized for both destructive and non-destructive content generation.
Simulation is stored as a series of shapekeys which allows fine grain control over the baked data and enables easy transferring of blend files from one user to another.
I was inspired by Computational Growth by Deskriptiv and Wanderers by Mediated Matter and fortunately was assisted to create a fully functional growth simulation toolkit as part of a 3 months development project. However the codebase is larger than what I can maintain alone and is available in a separate branch.
Branches & Features
The master branch contains only the features that I was responsible for since their maintenance and future development proves to be easier. An exception to this are excerpts of code borrowed for nodetree management from Animation Nodes developed by Jacques Lucke whose addon was inspirational in how we approach user interactivity with GrowthNodes.
The master branch includes the following nodes and features:
- Mesh objects
- Procedural Textures (Blender Internal Textures)
- Image Textures
- Vertex Groups
- Geometry Displacement
- Geometry Dissolve
- Geometry Subdivide
- Slope Detection
- Crease Detection
- Integer, Float, and Boolean Operations
The experimental branch (UMOG) has many additional features that require compilation, external python modules and are not fully tested and are often hacky. However this branch deserves much attention as it offers many advanced and optimized features beyond what Blender could provide. These are:
- GPU computed Gray-Scott Reaction Diffusion
- Support for both 2D and 3D (solid) textures.
- Cython optimizations
- An experimental engine implemented in Cython without calling Blender's Mesh API with up to 50x performance boost for select operations.
GrowthNodes uses Blender's PyNodes API to facilitate user interaction. Functions and data blocks are represented as self-contained nodes. Such an approach allows for scalable and customized node systems that surpass all static GUI implementation.
GrowthNodes can behave intelligently with regards to existing geometry. User is able to select specific regions based on geometric attributes such face slopes or the angle of crevices. Furthermore one can introduce additional detail by subdividing specific regions on the fly and apply growth only to selected regions. Essentially these are simplified yet powerful utilities to obtain dynamic topology.
- Intersection prevention
- Correct texture preview
- Additional parameters fo targeting geometry
- Persistent vertex groups
- Mesh sequence export
If you have any suggestions feel free to contact me @hiradsab on twitter.