Raga recognition using fibonacci series based pitch distribution in Indian Classical Music. (October 2020)

• Record Type: Journal Article
• Authors: Sinith, M.S.; Tripathi, Shikha; Murthy, K.V.V.
• Is Part Of: Applied acoustics. Volume 167(2020)
• Publication Date: 2020-10
• Subjects: Indian Classical Music -- Raga recognition -- Fibonacci series based pitch distribution -- Hidden Markov Model; Acoustical engineering -- Periodicals; Periodicals; 620.2
• DOI: 10.1016/j.apacoust.2020.107381
• Languages: English
• ISSNs: 0003-682X

Abstract

Abstract: Fibonacci Series and Music are related in many ways. Even though Indian Classical Music (ICM) is known to be based on Just Intonation rather than equal tempered, not much attention has been given to inflections on the notes, which are integral part of the notes themself. In this work an interesting relation between Fibonacci series and Just Intonation in ICM has been found. Further, discrete version of pitch contours based on a table derived using Fibonacci series is introduced for ICM. These pitch contours can effectively represent the melody and can be a tool for music information retrieval. A pitch distribution based on Fibonacci series is also introduced. Raga recognition is performed using this distribution and a recognition rate of 95.3 percent is obtained for the typical 17 ragas considered for this experiment. The recognition rate is higher compared to reported work in raga recognition.