Synchronize Dual Hands for Physics-Based Dexterous Guitar Playing

Abstract

We present a novel approach to synthesize dexterous motions for physically simulated hands in tasks that require coordination between the control of two hands with high temporal precision. Instead of directly learning a joint policy to control two hands, our approach performs bimanual control through cooperative learning where each hand is treated as an individual agent. The individual policies for each hand are first trained separately, and then synchronized through latent space manipulation in a centralized environment to serve as a joint policy for two-hand control. By doing so, we avoid directly performing policy learning in the joint state-action space of two hands with higher dimensions, greatly improving the overall training efficiency. We demonstrate the effectiveness of our proposed approach in the challenging guitar-playing task. The virtual guitarist trained by our approach can synthesize motions from unstructured reference data of general guitar-playing practice motions, and accurately play diverse rhythms with complex chord pressing and string picking patterns based on the input guitar tabs that do not exist in the references. Along with this paper, we provide the motion capture data that we collected as the reference for policy training. Code is available at: https://pei-xu.github.io/guitar.

SIGGRAPH Asia 2024. Video: https://www.youtube.com/watch?v=r_y0P2pIeF8&list=PLLfEynalFz6j0X5Kiut0U3GLRxt3Oz_oa. Project Webpage: https://pei-xu.github.io/guitar

References in corpus (13)

• DeepMimic: Example-Guided Deep Reinforcement Learning of Physics-Based Character Skills
• Character Controllers Using Motion VAEs
• GRAB: A Dataset of Whole-Body Human Grasping of Objects
• ASE: Large-Scale Reusable Adversarial Skill Embeddings for Physically Simulated Characters
• Gesticulator: A framework for semantically-aware speech-driven gesture generation
• Audio to Body Dynamics
• ControlVAE: Model-Based Learning of Generative Controllers for Physics-Based Characters
• Temporally Guided Music-to-Body-Movement Generation
• Semantic Gesticulator: Semantics-Aware Co-Speech Gesture Synthesis
• Composite Motion Learning with Task Control
• A GAN-Like Approach for Physics-Based Imitation Learning and Interactive Character Control
• AdaptNet: Policy Adaptation for Physics-Based Character Control
• Synthesize Dexterous Nonprehensile Pregrasp for Ungraspable Objects

Cited by in corpus (1)

• ELGAR: Expressive Cello Performance Motion Generation for Audio Rendition