Abstract
Bharatanatyam is the most popular form of Indian Classical Dance. Its Adavus are basic choreographic units of a dance sequence. An Adavu is accompanied by percussion and vocal music and follows a specific rhythmic pattern (Sollukattu). In this paper, we first characterize the audio, video, and sync events of Adavus to succinctly represent the Adavus. Then, we present simple yet effective algorithms to detect audio and video events and measure their synchronization. The audio, video, and sync event detection achieve 94%, 84%, and 72% accuracy, respectively. A comparison of our audio event detection against a well-known method by Ellis shows significant improvement. We also create an annotated repository of Sollukattus and Adavus for research. There are several applications of the characterization and beat detection including music/music video segmentation, synchronization of the postures with the beats, automatic tagging of rhythm metadata, etc. Characterization of events or repository of Bharatanatyam Adavus has not been attempted before.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
