Breaking Good: Fracture Modes for Realtime Destruction SIGGRAPH Asia 2022



1University of Toronto, 2CSU Fresno, 3UCLA, 4MIT, 5University of Maryland, 6Adobe Research


Drawing a direct analogy with the well-studied vibration or elastic modes, we introduce an object's fracture modes, which constitute its preferred or most natural ways of breaking. We formulate a sparsified eigenvalue problem, which we solve iteratively to obtain the n lowest-energy modes. These can be precomputed for a given shape to obtain a prefracture pattern that can substitute the state of the art for realtime applications at no runtime cost but significantly greater realism. Furthermore, any realtime impact can be projected onto our modes to obtain impact-dependent fracture patterns without the need for any online crack propagation simulation. We not only introduce this theoretically novel concept, but also show its fundamental and practical advantages in a diverse set of examples and contexts.


Supplemental Video


  title = {Breaking Good: Fracture Modes for Realtime Destruction},
  author = {Silvia Sellán and Jack Luong and 
	Leticia Mattos Da Silva and Aravind Ramakrishnan and 
	Yuchuan Yang and Alec Jacobson},
  year = {2022},
  journal = {ACM Transactions on Graphics}


This project is funded in part by NSERC Discovery (RGPIN2017–05235, RGPAS–2017–507938), New Frontiers of Research Fund (NFRFE–201), the Ontario Early Research Award program, the Canada Research Chairs Program, the Fields Centre for Quantitative Analysis and Modelling and gifts by Adobe. The first author is supported by an NSERC Vanier Scholarship and an Adobe Fellowship. The four middle authors were supported by the 2020 Fields Undergraduate Summer Research Program.

We acknowledge the authors of the 3D models used throughout this paper: MakerBot (Fig. 1, CC BY 4.0), HQ3DMOD (Figs. 6 and 19, TurboSquid 3D Standard Model License), Freme Minskib (Fig. 7, CC BY-NC 4.0), 3Demon (Fig. 9, CC BY-NC-SA 4.0), Reality_3D (Fig. 11, CC BY 4.0), Alex (Fig. 14, CC BY-NC-SA 4.0), Falha Tecnologica (Fig. 18, TurboSquid 3D Standard Model License), LeFabShop (Fig. 16, CC BY-NC 4.0), The Database Center for Life Science (Fig. 15, CC BY-SA 2.1) and Gijs (inset in Section 3.5, CC BY-NC 4.0).

We are grateful to the anonymous peer reviewers for their in- sightful suggestions. We would especially like to thank Reviewer #3 for inspiring the shockwave-based impact projection in Section 3.3. We would also like to thank Chris Wojtan, David Hahn and Klint Qinami for early experiments and discussions of sparse-norm frac- ture models; Eitan Grinspun, David I.W. Levin, Oded Stein and Jackson Phillips for insightful conversations; Rinat Abdrashitov for providing an implementation of his algorithm mentioned in Sec- tion 3.6; Qingnan Zhou for providing the 3D models used in Fig. 5; Xuan Dam, John Hancock and all the University of Toronto DCS staff that kept our lab running during the hardest of times.