Authors: Asoke Kumar Datta, Ranjan Sengupta, Nityananda Dey
Title: "Objective Analysis of Shrutis From The Vocal Performances of Hindustani Music Using Clustering Algorithm"
An objective analysis of shrutis from the 'alap' from 150 songs performed by 53 eminent musicians and scholars covering 21 different ragas are presented in this paper. The total time of singing analysed is nearly 8.5 hours. The audio samples are taken from recorded performances. Pitch extraction is done using a specially developed PDA/VDA (phase-space algorithm). The experiment handled 1.36 GB of data to extract the steady pitch states. The steady states are determined from the extracted pitch sequences. The ratio of the steady states with the tonic of each artist in each song was then found out. The complete data of 150 songs was then folded into the middle octave and distributed over 1200 bins of one-cent width each. The shruti positions were then ascertained using a clustering algorithm. The non-cognitive approach utilized here for the extraction of shrutis is unique in nature.