Uploaded July 2017 | Updated September 2026, 19 hours ago
Trying out the volumetrics capability of the new Eevee real-time renderer in Blender 2.8. It supports procedural textures so I threw these complex shaders at it and at 128 samples and 128 shadow samples performance quickly dropped down to one minute per HD frame on an average GPU.
On most systems this will cause the display driver to fail, because the OS will terminate and restart it after a few seconds when it's not responding anymore. On Windows you have to make changes to the Registry to avoid this. Yuck!
Fun fact: both scenes have 0 vertices. :)
Trying out the volumetrics capability of the new Eevee real-time renderer in Blender 2.8. It supports procedural textures so I threw these complex shaders at it and at 128 samples and 128 shadow samples performance quickly dropped down to one minute per HD frame on an average GPU.
On most systems this will cause the display driver to fail, because the OS will terminate and restart it after a few seconds when it's not responding anymore. On Windows you have to make changes to the Registry to avoid this. Yuck!
Fun fact: both scenes have 0 vertices. :)
![Arecibo Observatory Collapse Simulation [8K]](https://i.ytimg.com/vi/AqBvyx_zSK4/mqdefault.jpg "Arecibo Observatory Collapse Simulation [8K]
This is a structural simulation of the collapse of the Arecibo Observatory in 2020. The simulation implements the three major failure events which ultimately led to the progressive failure causing the platform to fall into the dish.
The simulation is based in part on actual physical data such as mass and strength for the larger components, and in part on plausible estimations. What appears to be multiple simulations at different points in time is actually one large global simulation. Deviations from reality are to be expected due to the fact that the simulation software runs completely autonomous, only the main initial failure points have been triggered by a predefined script. Nothing is animated by hand.
01:06 Cable M4N (Auxiliary cable from Tower 4) socket failure
01:40 Cable M4-4 failure
02:22 Cable M4-2 failure
Failure history:
2020-08-10 Tower 4 auxiliary main cable north comes loose from the socket (#301)
2020-11-06 Tower 4 main quadruple cable south broke, one of four (#101)
2020-12-01 Tower 4 main cable total failure and collapse
Some stats for nerds:
Base 3D model - 1.87 million polygons
Simulation model - 12,900 elements + 136k constraints
Simulation time - ~16 h
Credits:
Simulation & video by Kai Kostack
http://kostackstudio.de
Observatory 3D model provided by Dr. Rhys Taylor
http://www.rhysy.net
Additional data obtained from Phil Perlialts website
http://www.naic.edu/~phil
Music: Kai Engel - Run (CC BY)
Made with Blender + BCB + Fracture Modifier
https://blender.org
https://inachuslaurea.wordpress.com
https://blenderartists.org/forum/showthread.php?343637-Custom-Build-Blender-Fracture-Modifier
The BCB structural simulation software has been developed at the Laurea University of Applied Sciences, Finland. Written within the scope of EU Inachus FP7 Project (607522): Technological and Methodological Solutions for Integrated Wide Area Situation Awareness and Survivor Localisation to Support Search and Rescue (USaR) Teams
Island Observatory Failure")
