A review on tonic estimation algorithms in indian art music

• Published in: Multimedia tools and applications 2024-04, Vol.83 (13), p.38443-38463
• Main Authors: M.A., Aiswarya, Rajan, Rajeev
• Format: Article
• Language: English
• Subjects: Algorithms; Artists; Computer Communication Networks; Computer Science; Data Structures and Information Theory; Error analysis; Group delay; Multimedia Information Systems; Music; Pitch estimation; Special Purpose and Application-Based Systems
• ISSN: 1573-7721; 1380-7501; 1573-7721
• DOI: 10.1007/s11042-023-17161-4

Tonic is one of the basic concepts in music. The tonic pitch is normally entrenched by the main performers, depending on their vocal range and the range of the accompanying instruments. Hence, it varies among artists as well as performances. Tonic identification is a key task to be performed since it plays an important role in solving other problems like raga recognition, tuning analysis, etc. In this paper, we review some of the latest tonic identification approaches in Indian art music. We study the performance of each method in the context of music tradition using the Indian art music dataset. We analyzed the performance of the cepstral-based method of tonic pitch estimation with and without optimal selection of frames, non-negative matrix factorization (NMF), frequency-ratio method, and group delay-based algorithm. Furthermore, we show that the modified group delay-based methods outperform the conventional tonic estimation methods. We also present a detailed error analysis of each method which will be helpful for future research.