❤️ Check out Weights & Biases and sign up for a free demo here: https://www.wandb.com/papers
❤️ Their mentioned post is available here: https://wandb.ai/wandb/getting-started/reports/Debug-Compare-Reproduce-Machine-Learning-Models--VmlldzoyNzY5MDk?utm_source=karoly
📝 The paper "Complementary Dynamics" is available here:
https://www.dgp.toronto.edu/projects/complementary-dynamics/
🙏 We would like to thank our generous Patreon supporters who make Two Minute Papers possible:
Aleksandr Mashrabov, Alex Haro, Alex Serban, Alex Paden, Andrew Melnychuk, Angelos Evripiotis, Benji Rabhan, Bruno Mikuš, Bryan Learn, Christian Ahlin, Eric Haddad, Eric Lau, Eric Martel, Gordon Child, Haris Husic, Jace O'Brien, Javier Bustamante, Joshua Goller, Lorin Atzberger, Lukas Biewald, Matthew Allen Fisher, Michael Albrecht, Nikhil Velpanur, Owen Campbell-Moore, Owen Skarpness, Ramsey Elbasheer, Robin Graham, Steef, Taras Bobrovytsky, Thomas Krcmar, Torsten Reil, Tybie Fitzhugh.
If you wish to support the series, click here: https://www.patreon.com/TwoMinutePapers
Meet and discuss your ideas with other Fellow Scholars on the Two Minute Papers Discord: https://discordapp.com/invite/hbcTJu2
Károly Zsolnai-Fehér's links:
Instagram: https://www.instagram.com/twominutepapers/
Twitter: https://twitter.com/twominutepapers
Web: https://cg.tuwien.ac.at/~zsolnai/
#wiggles #jiggles
Оглавление (2 сегментов)
Segment 1 (00:00 - 05:00)
Dear Fellow Scholars, this is Two Minute Papers with Dr. Károly Zsolnai-Fehér. I hope you like wiggles and jiggles, because today… we are going to a see a lot of them. You see, this technique promises to imbue a rigged animation with elastoplastic secondary effects. Now, if you tell this to a computer graphics researcher, they will be extremely happy to hear that this is finally possible, but what does that really mean? This beautiful fish animation is one of the finest demonstrations of the new method. So, what happened here? Well, the first part means that we modeled a piece of 3D geometry, and we wish to make it move, but in order to make this movement believable, we have to specify where the bones and joints are located within the model. This process is called rigging, and this model will be the input for it. We can make it move with traditional methods, well, kind of. You see that the bones and joints are working, but the model is still solid. For instance, look…the trunk and ears are both completely solid. This is not what we would expect to see from this kind of motion. So, can we do better than this? Well, hold on to your papers, and let’s see how this new technique enhances these animations. Oh yes, floppy ears, also, the trunk is dangling everywhere. What a lovely animation! Prepare for a lot more wiggles and jiggles. Now, I love how we can set up the material properties for this method, of course, it cannot just make decisions by itself because it would compete with the artist’s vision, the goal is always to enhance these animations in a way that gives artists more control over what happens. So what about more elaborate models? Do they work too? Let’s have a look and find out together! This is the traditional animation method. No deformation on the belly, nostrils are not moving too much, so now, let’s see the new method. Look at that. The face, ears, nose, and the mouth now show elastic movement. So cool! Even the belly is deforming as the model is running about. So we already see that it can kind of deal with the forces that are present in the simulation. Let’s give this aspect a closer look. This is the traditional method…we are moving up and down, up and down. Alright, but look at the vectors here, there is an external force field, or in simpler words, the wind is blowing. And, unfortunately, not much is really happening to this model. But… when we plug in this new technique, look, it finally responds to these external forces. And as a reward, we get more wiggles. Ha-haa! So, what else can this do? A lot more. For instance, it does not require this particular kind of rig with the bones and joints. This hedgehog was instead, rigged with two handles, which is much simpler, but unfortunately, when we start to move it with traditional techniques, …well… leg is kind of pinned to the ground, where the remainder of the model moves. So, how did the new technique deal with this? Oh yes, the animation is much more realistic, and things are dangling around in a much more lively manner. The fact that it works on many different kinds of rigged models out there in a wild bolsters the usability of this technique a great deal. But we are not done yet. No-no! If you have been holding on to your papers so far, now, squeeze that paper, because we can take a scene, drop in a bunch of objects, and expect a realistic output. But wait a second…if you have been watching the series for a while, you know for a fact that for this, we need to run an elaborate physics simulator. For instance, just look at this muscle simulation from earlier. And here is the key, this animation took over an hour for every second of video footage that you see here. The new method does not need to compute a full-blown physics simulation to add this kind of elastic behavior, and hence, in many cases, it runs in real time. And it works for all kinds of rigs out there in the wild, and we even have artistic control over the output. We don’t need elaborate models and many hours of rigging and simulation to be able to create a beautiful animation anymore. What a time
Segment 2 (05:00 - 05:00)
to be alive! Thanks for watching and for your generous support, and I'll see you next time!