Date of Publication :28th October 2023

Abstract: Using an innovative methodology, the audio recordings utilized in the Indian Classical Dance (ICD) version of Kathak have been identified and categorized. This footwork for the Kathak Ladi was created from an audio dataset. This audio dataset was examined for classification purposes, and feature extraction was performed. To identify the beginning of the beats, beat detection is used. Beat tracking is then applied to the audio signals. Then, based on rhythms and mfccs, Ladi audios are identified and categorized into seven classes—L2, L3, L4, L5, L6, L7, and L8—using a machine learning-based approach. The Convolution Neural Network (CNN), which is now in its experimental stage, can produce results that are more accurate. Our study recommends more subclassification of each class based on the underlying variations in the quantity of bols and taals in the related audios of Kathak dance.

Reference :