Virtual Reality Dance Tracks from Skeletal Animations

Date：Dec 2023

Teams:PUC-Rio,

Writers:Leonardo Mastra,Luiz J. S. Silva,Alberto Raposo,Vinicius da Silva

PDF:Virtual Reality Dance Tracks from Skeletal Animations

Abstract

This paper presents a novel approach for automatically generating Virtual Reality (VR) dance tracks, focusing on translating the movement of animated 3D models directly derived from music. Our method capitalizes on the recent advances in automated synthesis of animated 3D models from music, using this data to bridge the gap to fully automatic VR dance track creation. We introduce a novel plugin for the Unity game engine, facilitating the conversion of dance animations from the Choreomaster dataset into dance tracks for the VR game Synth Riders. This approach aims to guide dance while supporting creative movement interpretation and minimizing movement restriction. The paper offers a comprehensive review of the current state of VR dance experiences and music-to-animation synthesis, and an in-depth explanation of our plugin and its key components. Our approach has potential to enhance the creation of dance experiences in VR, reducing resource dependency and increasing versatility in dance styles. This research is a step towards fully automated VR dance track creation, potentially impacting realms of music, dance, and fitness in VR.