Computational Analysis and Classification of the Pannisai system of Indian Classical Music

This pilot study explores the computational classification of Pannisai, an ancient Tamil musical system used to sing Thevaram hymns (5th–7th century), through modern deep learning techniques. By collecting datasets from Othuvars (traditional Pannisai practitioners) and the Tamizhisai Kalloori (Institute of Tamil Music), the research applies methods similar to raga classification to map the Pannisai system to contemporary Indian Carnatic music. In addition to classifying three Pannisai forms using CNN, RNN and LSTM the study compares the Pannisai framework with the swaras of Carnatic ragas. This comparison provides insights into the structural and melodic nuances of both systems, highlighting their interconnectedness. The promising accuracy of initial experiments underscores the potential of computational approaches in preserving and analyzing this intricate musical tradition, offering a bridge between ancient and modern musical frameworks, thereby opening a novel computational perspective on musicological studies.