# These Smoke Simulations Have A Catch! 💨 ## Метаданные - **Канал:** Two Minute Papers - **YouTube:** https://www.youtube.com/watch?v=QrsxIa0JDi4 - **Дата:** 27.02.2022 - **Длительность:** 7:09 - **Просмотры:** 85,988 - **Источник:** https://ekstraktznaniy.ru/video/13644 ## Описание ❤️ Check out Weights & Biases and sign up for a free demo here: https://wandb.com/papers 📝 The paper "Spiral-Spectral Fluid Simulation" is available here: http://www.tkim.graphics/SPIRAL/SpiralSpectralFluids.pdf ❤️ Watch these videos in early access on our Patreon page or join us here on YouTube: - https://www.patreon.com/TwoMinutePapers - https://www.youtube.com/channel/UCbfYPyITQ-7l4upoX8nvctg/join 🙏 We would like to thank our generous Patreon supporters who make Two Minute Papers possible: Aleksandr Mashrabov, Alex Balfanz, Alex Haro, Andrew Melnychuk, Angelos Evripiotis, Benji Rabhan, Bryan Learn, Christian Ahlin, Eric Martel, Gordon Child, Ivo Galic, Jace O'Brien, Javier Bustamante, John Le, Jonas, Jonathan, Kenneth Davis, Klaus Busse, Lorin Atzberger, Lukas Biewald, Matthew Allen Fisher, Michael Albrecht, Michael Tedder, Nikhil Velpanur, Owen Campbell-Moore, Owen Skarpness, Rajarshi Nigam, Ramsey Elbasheer, Steef, Taras Bobrovytsky, Thomas Krcmar, Timothy Sum Hon Mun, Torste ## Транскрипт ### Introduction [] Dear Fellow Scholars, this is Two Minute  Papers with Dr. Károly Zsolnai-Fehér. After reading a physics textbook on the laws of  fluid motion, with a little effort, we can make   a virtual world come alive by writing a computer  program that contains these laws, resulting in   beautiful fluid simulations like the one you  see here. The amount of detail we can simulate   with these programs is increasing every year,  not only due to the fact that computer hardware   improves over time, but also, the pace of progress  in computer graphics research is truly remarkable. And, this new paper promises   detailed spiral-spectral fluid and  smoke simulations. What does that mean? ### Details [0:47] It means that the simulations can be run in  a torus, spheres, cylinders, you name it. But wait, is that really new? When using  traditional simulation techniques, we can just   enclose the smoke in all kinds of domain  shapes where the simulation takes place.    People have done this for decades now. Here  is an example of that. So, what is new here?    Well, let’s have a look at some results  and hopefully, find out together. Let’s start with the details first.   This is the new technique. Hmmm.    I like this one. This is a detailed simulation,  sure, I’ll give it that. But we can already create   detailed simulations with traditional techniques,  so, once again, what is new here? You know what,   let’s compare it to a traditional smoke simulation  technique and give it the same amount of time to   run, and see what that looks like. Wow. That is a  huge difference. And, yes, believe it or not, the   two simulations run in the same amount of time.   So, it creates detailed simulations. Checkmark. ### Robustness [2:02] And, it has not only the details, but it has  other virtues too. Now, let’s bring up the heat   some more. This is a comparison to not an older,  classical technique, but a spherical-spectral   technique from 2019. Let’s see how the new method  fares against it. Well, they both look good,   so maybe the new method is not so much  better after a…wait a second. Ouch!    The previous one blew up. And the new one,   yes, this still keeps going. Such improvement in  just about two years. So, it is not only fast,   but it is robust too. That is super important for  real-world use. Details and robustness. Checkmark. Now, let’s continue with the shape of the  simulation domain. Yes, we can enclose the   simulation within this domain where the spherical  domain itself can be imagined as an impenetrable   wall, but it doesn’t have to be that way.   Look. We can even open it up! Very good. Okay, so it is fast, it is robust, supports  crazy simulation domain shapes, and even better,   it looks detailed, but are these the right  details? Is this just pleasing for the eye,   or is this really how the smoke should  behave? Well, the authors tested that too,   and now, hold on to your papers, and…look! I could  add the labels here, but does it really matter?    The two look almost exactly the same. Almost  pixel perfect. By the way, here you go. So, wow, the list of positives just keep on  growing. But we are experienced Fellow Scholars   here, so let’s continue interrogating this method.   Does it work for different viscosities? At the ### Interaction [4:02] risk of simplifying what is going on, the  viscosity of a puff of smoke relates to how   nimble the simulation is. And, it can handle  a variety of these physical parameters too. Okay, next, can it interact with other objects  too? Some techniques look great in a simple,   empty simulation domain, but break down when  placed into a real scene where a lot of other   objects are moving around. Well, not this new  one. That is a beautiful simulation, I love it. So, I am getting more and more convinced with each  test. So, where is the catch? What is the price   to be paid for all this? Let’s have a look.   For the quality of simulations that we get, ### Quality [4:48] it runs in a few seconds per frame. And  it doesn’t even need your graphics card,   it runs on your processor. And  even that, this is blazing fast.    And, implementing this on the graphics-card could  very well put this into the real-time domain,   and boy, getting these beautiful smoke puffs  in real time would be an amazing treat. So, once again, what is the price to be paid  for this? Well, have a look. Aha! There it is.    That is a steep price. Look. This needs  tons of memory. Tens of gigabytes.    No wonder this was run on the processor, this  is because modern consumer graphics cards   don’t have nearly as much memory on board, so  what do we do? Well, don’t despair - not even ### Conclusion [5:40] for a second. We still have good news. And the  good news is that there are earlier research works   that explore compressing these datasets down,  and it turns out, their size can be decreased   dramatically. A perfect direction  for the next paper down the line!    And, what do you think? Let me know in the  comments below what you would use this for. And, just one or two more papers down the line,  and maybe, we will get these beautiful simulations   in our virtual worlds in real time. I can’t wait.   I really cannot wait. What a time to be alive! Thanks for watching and for your generous  support, and I'll see you next time!