- Book author
- Ambuja Salgaonkar
- Makarand Velankar
This book is intended for researchers interested in using computational methods and tools to engage with music, dance and theatre. The chapters have evolved out of presentations and deliberations at an international workshop entitled Computer Assisted Music and Dramatics: Possibilities and Challenges organized by University of Mumbai in honour of Professor Hari Sahasrabuddhe, a renowned educator and a pioneering computational musicologist (CM) of Indian classical music.
The workshop included contributions from CM as well as musicians with a special focus on South Asian arts. The case studies and reflective essays here are based on analyses of genres, practices and theoretical constructs modelled computationally. They offer a balanced and complementary perspective to help innovation in the synthesis of music by extracting information from recorded performances. This material would be of interest to scholars of the sciences and humanities and facilitate exchanges and generation of ideas.
The Natyashastra (dramatics in Sanskrit), the 2500-year-old surviving Indian compendium on theatrical art, defined music as having three forms: vocal, instrumental and dance. It has referred explicitly to music in half of its chapters. Though music and dramatics have evolved along with society, the fact remains unchanged that they still share the traditional framework to a great extent. An international workshop, “Computer Assisted Music and Dramatics: Possibilities and Challenges (CAMAD’19)” held on February 25–27, 2019 in the Department of Computer Science, University of Mumbai, focuses on this definition.
AI has been driving traditional human centred activities to partly or completely automated processes. Automation has been attempted for a range of tasks, from identification and composition to judgement and appreciation of art performances, which have been considered the exclusive forte of human intelligence. The emphasis of CAMAD’19 was on developing computational models, as far as possible, to study music and allied fields. The selected 15 papers of the workshop are getting published under Springer’s book series Advances in Intelligent Systems and Computing.
The first book in the Springer series on computational music science is from 2010. Soon enough the fourth title, Computational Musicology in Hindustani Music, was published in 2016. Though not part of the same series, we are privileged to publish the second book on computational musicology in Indian classical music by Springer. The 2016 book illustrates fundamental aspects like the role of statistics and introduction to a computational research platform, while in the present volume, readers will find the application of AI and ML to musicology. Topics like structural analyses, entropy, comparison of ragas and machine-assisted composition of music are thrust areas even now, and they will be researched in the future as well. Here they have been extended to the domains of instrumental music and dance. The therapeutic use of Indian classical music had been touched upon in the concluding chapter of the earlier book. This topic is important for society. For want of substantive objective proof, we could not consider two papers in this domain that were presented in CAMAD’19, including one by Guru Prem Vasantji.
Here, a set of 56 songs analysed by employing a regression-based learning model is shown to have captured the mood of a song with an accuracy of 73%. This study emerged out of a keynote speech, “Role of Prosody in Music Meaning,” proposing features like tempo, melody and instrumentation to delve into classical music. A paper on music composition using PSO, and another on automating Bharatanatyam choreography, propose to direct and evaluate processes in accordance with the traditional framework without human intervention. These initiatives are worth pursuing. Exploring the Bharatnatyam Margam using graph theory has been considered for the first time. The paper on developing a musicality scale provides an elaborate computational process for checking a poetic piece for its potential to become a song. Such research is opening up new areas for computational exploration to heighten Indian music and dance forms in tune with the forthcoming technoscience era. They may attract the young generation. Yet another keynote speech “Epistemology of Intonation” was on the theoretical foundations of music, the computation of 22 shrutis and an interpretation of how they are instrumental in creating musical mood. An attempt to chart out directions for advancing computational musicology in the Indian context, by linking past and present knowledge and by employing associated technology, is made in “Computational Indian Musicology: Challenges and New Horizons”. Such information presented by practicing stalwarts will develop deeper insights into the subject.
The distribution and ordering of the papers in suitable sections were shaped by the review process. The three papers in the first category entitled Computer Assisted Musicology address the extraction of musicological information from concert performances by processing their sound signals with the help of known techniques. Their outcome may yield feedback for future performances of seasoned musicians or could be employed in training novice learners. The five papers in the second category, Machine Learning Approaches to Music, are about the application of machine learning for processing metadata to extract pragmatic information. Research on the computation of aesthetics has evolved as an extension of the research on automating Bharatanatyam steps and sequences. These two papers, along with one on the automatic creation of tabla compositions, form the third category entitled Composi- tion and Choreography. The fourth category is Interfacing the Traditional with the Modern. The papers here are suggestive of future trends. New perceptions of traditionally available information have been put forth. For example, music production has become an attractive career path for music-loving engineers.
A review paper on the evolution of research in musicology and another on the industrial applications of musicology research, and two more on the automatic classification and clustering of ragas from CAMAD’19, are not included in this book since the authors could not submit the final copy. In the context of non-Indian theatre, there are citations in the literature to the computation of the ontology of a play for purposes of action analysis, modelling suspense and dramatic arc in order to predict the success potential of a story, as well as the estimation of dramatic tension as a function of goals, obstacles and side effects. However, no paper on computational dramatics was submitted to CAMAD’19, though we had hoped to hear about research on the computational aspects mentioned in the Natyashaastra. Kathak dance guru Rajashree Shirke and acclaimed Marathi folk music performer Ganesh Chandanshive, fascinated as they were by the idea of applying computation to their domains of the performing arts, presented computable ideas in their respective domains. In the absence of implementations, at this stage, we are not able to include these papers in these Proceedings, as also another paper related to dramatics. We look forward to collaborations between the authors of these papers and interested researchers in the domains of music, computer science, mathematics and cognitive science to take these ideas to their logical conclusion.
As many as 18 authors have contributed papers to this volume. Interestingly, about a third of them are professional musicians, while the remaining are AI or computer scientists, about half of whom are formally educated in Indian classical music. For about a third of the authors, these papers are part or extensions of their doctoral research. In a way, this mix testifies to the variety and quality of the content.
Mainly because of circumstances due to the pandemic, this publication project took almost three years for its completion. Substantiating the writings with statistically proven results was a major task. This kind of experience is no less than that of completing a doctoral thesis. It is a matter of professional satisfaction that all the authors took the observations and suggestions in the right spirit and willingly revised their drafts, sometimes more than once. Our thanks are due to all of them for being with us, patiently and passionately, throughout this journey. Special thanks are due to Professor Hari Sahasrabuddhe for guiding us all through and Professor Jayant Kirtane for editing all the drafts for enhancing their readability. This project would not have been completed without the consistent support of Professor Vivek Patkar who critically went through each and every draft and provided clear and constructive feedback. What to say about Mr. Srijan Deshpande? After completing his own paper, he volunteered to help us organise the material under four categories from the practitioners’ perspective. We thank the authorities of the University of Mumbai, the then Pro Vice Chancellor Professor R D Kulkarni, in particular, for offering grants to host an array of world-class speakers at CAMAD’19. Thanks to our family and friends at our homes and our professional homes for providing all the required support. We have no words to express our gratitude and appreciation for the patience, consideration and guidance that we experienced at Springer. They agreed to produce this volume without expecting any financial support from our side.
We are sure that this book about the advances in computing and musicology specific to Indian music will receive due attention from researchers across the globe with varying perspectives. The proceedings should help discover fresh avenues and break new grounds to cater to emerging tastes and exploit advanced technologies. The readers of this volume would expect the next one. Ye dil mange more ... .
Mumbai, India Ambuja Salgaonkar Pune, India Makarand Velankar