A Survey on the Modelling and Classification of Raagaas in Hindustani Classical Music

Abstract

Raagaas are the basis of Hindustani Classical Music. They are made up of a group of 12 well structured characters called swaras. Combinations of these swaras give birth to different Raagaas. Though classical music is gaining popularity, Raagaas are quite under-studied from the point of view of Computer Science. There have been attempts to model them, classify them and generate them using several techniques. In this paper, we study and summarize various such approaches. This includes Machine-Learning based approaches like the use of LSTMs, CNNs and HMMs for classification and LSTMs and GANs for generation. The focus of our survey is on various problems solved by the previous work and their subsequent merits and drawbacks. We also include the survey methodology, which is followed by the future directions and conclusion, where we highlight some problems and solutions that could be applied to obtain better results on tasks like Raagaa Classification.

Introduction

The text reviews the problem of classification and generation of Raagaas in Hindustani Classical Music from a computer science perspective. Raagaas are melodic frameworks built from combinations of swaras (notes), each associated with specific moods, tempos, and times of performance. Because Raagaas are defined more by flexible musical properties than strict rules, identifying them is difficult even for experts, making automated classification and generation especially challenging.

The paper explains that most computational approaches to Raagaa identification rely on analyzing audio signals (e.g., spectrograms) and extracting features to predict the Raagaa. Human experts, by contrast, identify Raagaas through swaras and musical characteristics such as chalan and note emphasis. A key technical challenge is mapping frequencies to swaras, since Hindustani music uses a relative scale based on a chosen tonic, unlike Western music’s fixed frequencies.

Through an extensive literature survey, the paper reviews methods used for Raagaa classification and generation, including context-free grammars, probabilistic models, HMMs, random forests, LSTMs, GANs, and other deep learning approaches. Many studies use MFCC-based audio features, while others rely on swara sequences. The review also compares Hindustani music research with work on Western music, where MFCCs, VAEs, CNNs, and LSTMs are more appropriate due to fixed-scale structures.

Key findings highlight several limitations in existing research:

• Lack of scale invariance, a critical issue in Hindustani music that most models ignore.

• Overreliance on MFCC features, which do not accurately capture Raagaa-defining swara structures.

• Absence of standardized datasets, making comparisons across studies difficult.

• Limited Raagaa coverage, with many studies focusing mainly on Yaman and Bhairavi.

• Interpretability issues in deep learning models such as LSTMs, which function as black boxes.

Conclusion

In this paper, we summarize the techniques proposed and applied by several papers for the problem of Classification of Raagaas in Hindustani Classical Music. Using our methodology for the survey, we were able to select and filter high-quality papers with diverse proposals. Some of the categories of papers included in our survey are as follows: . For each of the selected papers, we highlight the problems identified by the authors, the proposed solutions and methodologies, and the performance of the solution when compared with other papers of the same category. We also present some unique problems related to the classification task that were not addressed/partially addressed by the authors of these papers. After presenting the literature review and the key findings, we conclude this paper by proposing potential future directions for the problem of Raagaa classification in Hindustani Classification Music.

References

[1] Gupta, K. (2022). Improvised Sequence Generation In North Indian Classical Music (Doctoral dissertation, International Institute of Information Technology, Hyderabad). [2] Adhikary, S., Manasa, S. M., Shreesha, S. K., & Bhat, S. (2023, April). Automatic music generation of indian classical music based on raga. In 2023 IEEE 8th International Conference for Convergence in Technology (I2CT) (pp. 1-7). IEEE. [3] Dighe, P., Karnick, H., & Raj, B. (2013, November). Swara Histogram Based Structural Analysis And Identification Of Indian Classical Ragas. In ISMIR (pp. 35-40). [4] Dighe, P., Agrawal, P., Karnick, H., Thota, S., & Raj, B. (2013, July). Scale independent raga identification using chromagram patterns and swara based features. In 2013 IEEE International Conference on Multimedia and Expo Workshops (ICMEW) (pp. 1-4). IEEE. [5] Pendyala, V. S., Yadav, N., Kulkarni, C., & Vadlamudi, L. (2022). Towards building a deep learning based automated indian classical music tutor for the masses. Systems and Soft Computing, 4, 200042. [6] Das, D., & Choudhury, M. (2005, January). Finite state models for generation of Hindustani classical music. In Proceedings of international symposium on frontiers of research in speech and music (pp. 59-64). [7] Joshi, D., Pareek, J., & Ambatkar, P. (2021, February). Indian Classical Raga Identification using Machine Learning. In ISIC (pp. 259-263). [8] Humse, K., HS, R. K., Ramachandra, L. S., GR, Y., & Puttegowda, K. (2025). Automated raga recognition in Indian classical music using machine learning techniques. Journal of Integrated Science and Technology, 13(1), 1011-1011. [9] Paschalidou, S., & Miliaresi, I. (2023). Multimodal deep learning architecture for hindustani raga classification. Sensors & Transducers, 260(2), 77-86. [10] Sharma, S., Ghisingh, S., & Mittal, V. K. (2017, September). Component characterization of western and indian classical music. In International Symposium on Signal Processing and Intelligent Recognition Systems (pp. 57-69). Cham: Springer International Publishing. [11] Tiwari, P., & Jha, S. (2024, June). Music Generation with Long Short-Term Memory Networks from MIDI Data of Classical Music. In 2024 IEEE International Conference on Information Technology, Electronics and Intelligent Communication Systems (ICITEICS) (pp. 1-4). IEEE. [12] Sambaragi, L. M., Naik, P. P., Kakatkar, N. N., Chikkamath, S., SR, N., & Budihal, S. V. (2024, March). Music Generation: A simplified approach. In 2024 3rd International Conference for Innovation in Technology (INOCON) (pp. 1-6). IEEE. [13] Briot, J. P., & Pachet, F. (2020). Deep learning for music generation: challenges and directions. Neural Computing and Applications, 32(4), 981-993. [14] Lim, Y. Q., Chan, C. S., & Loo, F. Y. (2020, July). Style-Conditioned Music Generation. In ICME (pp. 1-6). [15] Shi, Y. (2025). A CNN-Based Approach for Classical Music Recognition and Style Emotion Classification. IEEE Access. [16] Nanni, L., Costa, Y. M., Lumini, A., Kim, M. Y., & Baek, S. R. (2016). Combining visual and acoustic features for music genre classification. Expert Systems with Applications, 45, 108-117. [17] Singh, J. (2022). An efficient deep neural network model for music classification. International Journal of Web Science, 3(3), 236-248. [18] Feng, Y. (2025). Computational Modeling and Structural Generation of Piano Music in the Classical Style. [19] Simmermacher, C., Deng, D., & Cranefield, S. (2006, July). Feature analysis and classification of classical musical instruments: An empirical study. In Industrial Conference on Data Mining (pp. 444-458). Berlin, Heidelberg: Springer Berlin Heidelberg. [20] Goulart, A. J. H., Guido, R. C., & Maciel, C. D. (2012). Exploring different approaches for music genre classification. Egyptian Informatics Journal, 13(2), 59-63. [21] Goulart, A. J. H., Guido, R. C., & Maciel, C. D. (2012). Exploring different approaches for music genre classification. Egyptian Informatics Journal, 13(2), 59-63. [22] Pal, S., Kar, T., Kanungo, P., & Pattnaik, T. (2023, November). Composer Identification in Classical Genre. In 2023 IEEE 3rd International Conference on Applied Electromagnetics, Signal Processing, & Communication (AESPC) (pp. 1-6). IEEE. [23] Moysis, L., Iliadis, L. A., Sotiroudis, S. P., Boursianis, A. D., Papadopoulou, M. S., Kokkinidis, K. I. D., ... & Goudos, S. K. (2023). Music deep learning: deep learning methods for music signal processing—a review of the state-of-the-art. Ieee Access, 11, 17031-17052. [24] Walczy?ski, M., & Grzyba?a, P. (2020, September). Selected methods of parametrization in problem of automatic classification classical music from the Renaissance era against the classical works from other eras. In 2020 Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA) (pp. 151-156). IEEE. [25] Zhang, K. (2021). Music style classification algorithm based on music feature extraction and deep neural network. Wireless Communications and Mobile Computing, 2021(1), 9298654. [26] Kesimal, K., Reyhanl?oglu, M. E., & Ozcan, H. (2025, June). Deep Learning versus Machine Learning for Music Genre Classification: A Dual-Dataset Evaluation. In 2025 9th International Symposium on Innovative Approaches in Smart Technologies (ISAS) (pp. 1-6). IEEE. [27] Samsekai Manjabhat, S., Koolagudi, S. G., Rao, K. S., & Ramteke, P. B. (2017). Raga and tonic identification in carnatic music. Journal of New Music Research, 46(3), 229-245. [28] Li, M. (2025, August). Development of Intelligent Music Genre Classification Using Stacked LSTM Networks. In 2025 Third International Conference on Networks, Multimedia and Information Technology (NMITCON) (pp. 1-7). IEEE. [29] Xie, X. (2024, October). A Hybrid CNN-LSTM Architecture for Enhanced Music Genre Classification. In 2024 5th International Conference on Machine Learning and Computer Application (ICMLCA) (pp. 72-75). IEEE. [30] Du, P., Zhou, Z., & Li, F. (2025). GuitarNeXt: An Advanced Convolutional Neural Network Architecture for Music Genre Classification. Traitement du Signal, 42(4). [31] Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural computation, 9(8), 1735-1780. [32] Baum, L. E., & Petrie, T. (1966). Statistical inference for probabilistic functions of finite state Markov chains. The annals of mathematical statistics, 37(6), 1554-1563.

Copyright

Copyright © 2025 Pushkar Joglekar, Madhur Vaidya, Soham Nimale, Hrishikesh Potnis, Sarthak Pithe. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.