Marcel Hampel

Testing my minimal 3D reaction-diffusion setup in Jupyter. Visualization is slow af.. musing on the best way to port this in #blender. Suggestions welcomed.pic.twitter.com/1rOGSQg7Rj

First obvious approach: displacement + alpha, via exported video texture (not interactive, can't handle 3D data)pic.twitter.com/kOeCTsLPvH

Other approach: particle system with density controlled by texture (still not interactive nor 3D)pic.twitter.com/GwZkgbRQOH

Other approach: moving plane with particle system emitter controlled by video texture. Not interactive, but can render 2D systems evolution in 3D, or render a still of a 3D volume via exported 2D slicing.pic.twitter.com/0EvLkDVyAU

Simple script to edit in-memory image used as texture, with the simulation running as a frame-change handle. Semi-interactive, but still stuck to 2D, as I have no idea how to edit 3D texture data (volumes) programmatically. Code https://github.com/5agado/data-science-learning/tree/master/graphics/reaction_diffusion …pic.twitter.com/soGDYZpdXu

Testing 3D reaction-diffusion in Blender using stacked-planes + alpha from video-texture. Simulation is done via my custom Python codepic.twitter.com/s5UyVyYNhw

'COEFF_A': 0.12, 'COEFF_B': 0.06, 'FEED_RATE': 0.02, 'KILL_RATE': 0.05 #b3dpic.twitter.com/HL3EtN502q