View Full Version : Nice Physics Simulation Video
I wasn't sure at first if this was a real video or not, and it's several years old. Wow.
also see http://www.youtube.com/watch?v=3Ilojin4vQ8
02-18-2012, 05:52 AM
What is this? ...amateur hour? Obviously they neglected to take into account the nine independent constants needed to describe the elastic behavior of wood: three moduli of elasticity E, three moduli of rigidity G, and six Poisson’s ratios μ. Geez, the error is simply glaring!
02-18-2012, 06:37 PM
@erco: I've tried replicating a lot of this guy's work, and I subscribe to his channel. It's amazing the physical simulations he accomplishes with simple computer simulation software. He does a lot of his work with blender, which I also work a lot with. Here's my attempt at KEVA plank simulation. http://www.youtube.com/watch?v=VnYPDKf2MNo&feature=fvst
My computer isn't good enough to process more than 2000 planks, but I may have to try this out on the Mac to see if I can get more (without crashing)!
Very impressive, uC! Be proud, that's excellent work! I'll have to look into some of these simulation & modeling programs. Things have come a long way since CSMP in my college days. :)
02-18-2012, 10:42 PM
@erco: I've tried replicating a lot of this guy's work...
02-18-2012, 10:45 PM
Wonder how that would hold up to my geodesic sphere made of Popsicle sticks .... jk .... :-) .... that's pretty cool !!
It would be interesting to see a model of a man of war ship, and the bullets (with a bit more velocity, but not much) come crashing through the side. Most injuries in those days was done by flying spears of wood ripped from the sides of the ship.
Master & Commander... a great flick.
02-19-2012, 06:05 PM
Wow, thanks for all the positive reception to my video! You've inspired me to get out the project files again and run another (larger) simulation. I found in doing this that an iMac can handle 10,000+ KEVA planks without crashing while simulating. 1 hour of simulating and 6 hours of rendering later, I'm halfway done and will have a new video on YouTube as soon as I can edit the files. Thanks again guys. :)
@SRLM: Though there are algorithms for physics calculations, there are none (that I know of) on the market for the simulation of splintering wood. Most of what you see in movies is not a true simulation, but a well faked one. Each splinter is modeled and textured by hand, about 10 separate models are made, and then they are duplicated and emitted using a particle simulator. Particle simulators are not as accurate as physics simulators, but they do the job for cinematic effect.
The only other simulators I know of (besides physics simulators) are shatter simulators (which either cost money or don't work right) and fluid simulators.
I wasn't thinking of making any sort of additional simulator. I just wanted him to build a model of a ship using the KEVA planks, and position the camera inside a gun deck when the cannon balls (bullets) come crashing in. A "splinter" would be a plank.
02-19-2012, 09:26 PM
I might be able to do that. How large would the ship have to be in proportion to the planks? And the bullets in proportion to the ship?
If you're looking for a ship to base it off of, try searching google images for toothpick model ships. Most of the images are pretty poor (small), but you can get the general idea. It would probably be best to make the cannon balls a bit bigger than scale (as relative to the cannon bore) since the planks will be a bit bigger by scale.
Note that this particular ship is of a design from a little bit earlier than the great cannon ships of the 18th century...
02-19-2012, 11:53 PM
@SLRM: The problem with the physics sim is that, unlike a real ship, the blocks are only held together by gravity, meaning that the ship may just crumble apart when I run the simulator. I'll try it though. It may take a while to make the model.
02-20-2012, 12:41 AM
Here's my originally mentioned new physics sim video, simulating a "table bump". Hope you like it! :)
BTW, if anyone finds the editing annoying, I'll upload a straight-up version.
02-20-2012, 01:05 AM
I wasn't joking about a sphere made of Popsicle sticks ... :-)
This was for a "planet" project at school I was helping my daughter with ... The picture is just of the sphere sitting in a lamp as an artistic possibility filling the 'gaps' with wax or onion skin paper.. The design has longitude lines and an equator to hold everything in place, but if you repeated the longitude line pattern at every point to re-enforce the structure, I wonder how it would hold up to the simulations mentioned in this thread?
02-20-2012, 01:18 AM
Beau, that is really cool! (I did think you were kidding about it at first) I don't think it would hold up to any physics tests, though. :)
Nice, uC! You got that great new video generated with audio track in record time!
02-20-2012, 07:10 AM
"...I don't think it would hold up to any physics tests, though." ... Why not? ... I understand that these simulations rely on 'gravity' . Why can't you define the gravity to the center of an object? ... a planet if you will. Instead of bullets, fire asteroids. re-enforce the structure with more "longitude" / "equator" bands than what I have put together and if constructed right it should be a very withstanding structure.
02-23-2012, 01:22 PM
@Beau: The physics engine does not allow for the breaking and bending of materials. It is, simply put, a game engine. It's main intention is to be used for the design of video games and simple simulations. It is not a fully functional and totally reliable physics simulation program, it's just good enough to make some cool videos.
That being said, it will all change when Phymec releases his "Phymec Tools" package for Bullet Physics, which will allow me to do MUCH more complicated (and cool!) simulations. See his demo here: http://www.youtube.com/watch?v=FIPu9_OGFgc&context=C3943095ADOEgsToPDskL0C1gs8Imh5ZaFoKYdlcaO