Adobe's New Simulation: Bunnies Everywhere! 🐰

Dear Fellow Scholars, this is Two Minute  Papers with Dr. Károly Zsolnai-Fehér. Today is going to be all about simulating  virtual bunnies. What kinds of bunnies?    Bunnies in an hourglass with granular  material, bunnies in a mixing drum,   bunnies that disappear, we’ll  try to fix this one too. So, what was all this footage? Well, this is  a followup work to the amazing Monolith paper.    What is Monolith? It is a technique that helps  fixing commonly occurring two-way coupling issues   in physics simulations. Okay, that sounds great,   but what does two-way coupling mean? It means that  here, the boxes are allowed to move the smoke,   and the added two-way coupling part means that  now, the smoke is also allowed to blow away   the boxes. This previous work makes  can simulate these phenomena properly. It also makes sure that when thrown at the  wall, things stick correctly, and a ton of   other goodies too. So, this previous method shows  a lot of strength. Now, I hear you asking, Károly,   can this get even better? And the answer is yes,  yes it can! That’s exactly why we are here today. This new paper improves this  technique to work better in cases   where we have a lot of friction. For instance,  it can simulate how some of these tiny bunnies   get squeezed through the hourglass, and get  showered by this sand-like granular material.    It can also simulate how some of them remain stuck  up there because of the frictional contact. Good. Now, have a look at this, with this one, with  an earlier technique, we start with one bunny,   and we end up with…. wait a minute. That  volume is not one bunny amount of volume.    This is what we call the volume dissipation  problem. I wonder if we can get our bunny back   with the new technique. What do you think? Well,  let’s see…one bunny goes in, friction happens,   and…yes! One bunny amount of volume comes out of  the simulation. Then, we a bunch of them into a   mixing drum in the next experiment where their  tormenting shall continue. This is also a very   challenging scene because we have over 70 thousand  particles, rubbing against each other. And,   just look at that. The new technique is so robust  that there are no issues whatsoever. Loving it! So, what is all this simulation math good for?   Well, for instance, it helps us set up a scene   where we get a great deal of artistic freedom. For  instance, we can put this glass container with the   water between two walls, and, look carefully! Yes,  we apply a little leftward force to this wall.    And, since this technique can  simulate what is going to happen,   we can create an imaginary world where only  our creativity is the limit. For instance,   we can make a world in which there is just a  tiny bit of friction to slow down the fall. Or, we can create a world  with a ton more friction,   and now, we have so much friction  going on that the weight of the liquid   cannot overcome anymore, and thus, the cube  is quickly brought to rest between the walls. Luckily, since our bunny still exists, we can  proceed onto the next experiment, where we will   drop it onto a pile of granular material. This  previous method did not do too well in this case,   as the bunny sinks down. And if you think  that cannot possibly get any worse, well,   I have some news for you. It can. How? Look!   With a different previous technique, it doesn’t   even get to sink in, because it crashes when it  would need to perform these calculations. Now,   let’s see if the new method  can save the day, and…yes!    Great! It indeed can help our little  bunny remain on top of things. So, let’s pop the Scholarly question -  how much do we wait for such a simulation? The hourglass experiment takes  about 5 minutes per frame,   while the rotating drum experiment takes about  half of that, 2. 5 minutes. So, it takes a while.    Why? Of course, because many of these scenes  contain tens of thousands of particles,   and almost all of them are in constant frictional  interaction with almost all the others at the   same time, and the algorithm mustn’t miss  any of these interactions. All of them   have to be simulated. And the fact that through  the power of computer graphics research we can

simulate all of these in reasonable amount of time  is an absolute miracle. What a time to be alive! And as always, you know what’s coming. Yes, please  do not forget to invoke the First Law of Papers,   which says that research is a  process. Do not look at where we are,   will be two more papers  down the line. Granular materials,   and frictional contact in a matter of seconds  perhaps? Well, sign me up for this one! Thanks for watching and for your generous  support, and I'll see you next time!