Nikita Braguinski

Biography

NIKITA BRAGUINSKI INTERVIEW


This interview was first published on the Music Journalism Insider website.

March 13, 2022


***I’m Todd L. Burns, and welcome to Music Journalism Insider, a newsletter about music journalism. Click here to subscribe! https://www.musicjournalisminsider.com/

Nikita Braguinski is a musicologist and historian of technology. He studied musicology at the University of Cologne and wrote his PhD in media theory at the Humboldt University of Berlin. He was a visiting postdoctoral fellow at the Max Planck Institute for the History of Science, a postdoctoral fellow in the music department at Harvard University, and most recently a postdoctoral researcher at Humboldt University. His new book is Mathematical Music: From Antiquity to Music AI.***

CAN YOU PLEASE BRIEFLY DESCRIBE YOURSELF AND WHAT YOU DO?

I am a musicologist with a special interest in connections between music, technology, and mathematics. I studied musicology in Cologne, Germany, wrote my PhD in Berlin on early video game music, and now continue to study the many things that make music for us (or maybe instead of us).

When people hear I study music technology, they first of all think of recording technology, like the vinyl record or an mp3. But this is in fact not at all what I study most. Instead, I personally find it much more interesting to look at technologies that create, or seem to create, what we call music. So, I look at how artificial intelligence (AI), and especially machine learning or deep learning, is used to generate music. But I also find it equally fascinating to look at the ways people ingeniously employed all the other possibilities that they had before the computer came, like shuffled stacks of cards, slabs of wood, paper circles, and, of course, the famous rolling dice.

To contact me about my new book Mathematical Music, or any other topic connected to my research, please send an email to n.braguinski.uni [at] gmail.com

WHAT ABOUT THIS AREA OF STUDY IS SO INTERESTING TO YOU?

First of all, I am fascinated by the tension between music as a deeply personal experience and practice, and music as an almost fully formalized, standardized, and even mechanized set of “rules”. I think, the fact that music created in this semi-automatic way can be quite credible for a lot of people deserves a lot of attention, and could lead to deeper insights into what, ultimately, constitutes music. Of course, the issue of automation of human labor is a very urgent one, and it also has motivated me to study possible future scenarios around musical AI.

CAN YOU BRIEFLY SUMMARIZE YOUR BOOK?

My book Mathematical Music. From Antiquity to Music AI is a concise history of the ways in which mathematics has been used to create music. And since modern AI, such as machine learning, is based on a lot of mathematics, my book also looks at musical uses of artificial intelligence. I wrote the book for the non-specialist general audience, with the hope that it would help readers make sense of today’s music technology by situating it in a larger historical context.

The story of music, mathematics, and technology that I present in my book ranges from ratios in antiquity to random combinations in the 17th century, 20th-century statistics, and contemporary artificial intelligence. It provides a panorama of how thought processes involved in the creation of music became gradually mechanized. In the last chapters, I also take a look at what possibilities the near future of music AI might hold for listeners, musicians, and society.

The book is scheduled to appear in print on March 14, 2022. More information is available on the publisher’s website, and I am also happy to answer questions regarding my book (see my email address above).

HOW DID YOU FIRST FIND OUT ABOUT THE SUBJECT / REALIZE IT WAS SOMETHING YOU WANTED TO PURSUE?

In a way, this work is a continuation of one example that I found absolutely fascinating during my PhD study. One of the case studies in my dissertation was a relatively simple algorithm from an early video game which nevertheless produced quite credible, and almost endless, variations on a jazzy theme (the game is called “Ballblazer”, and the generated music appears when the game starts). I thought, “If a well done, but basically simple program can do this, what does it say about us, the music, and technology?”

Over the years that followed, I looked for inspirations and answers in different fields of study, such as the history of science and technology, or media theory. I also taught this topic at the university, which gave me a sense of what people need to know to discuss it. So, this new book is a quintessence of many strands of study and work that I have been doing in recent years.

WHAT WAS THE MOST SURPRISING THING THAT YOU FOUND IN YOUR RESEARCH?

There are many, otherwise I would not want to write a whole book about it. One relatively recent case that really got me thinking about the future of music technology is the project called Jukebox, created by the researchers working at the OpenAI organization. What is so special about it, is not only the quality of its musical output, but the fact that it skips musical notation, and goes straight to audio, with very credible results. Normally, people think about music, especially Western music, including popular genres, in terms of notation. That means, there is almost always an intermediary step before something audible is created, like a notated melody, or even just a set of chords on which a band agrees verbally. Mostly, music AI systems also operate on the level of notation. Here, however, audio is analyzed and generated directly, without notation. This makes it such an interesting case to look at, because there is a chance that such an approach could catch some of the underlying structure in music that never gets notated and is just being passed on from musician to musician without being acknowledged, or understood, in any clear way.

CAN YOU POINT US TO ANY FURTHER RESEARCH ON THE TOPIC THAT YOU THINK IS RELEVANT?

In this book, I have focused especially on popular music, both historical and contemporary. But there is a whole world of very interesting technological experiments in art music, and of artistic and academic commentary on these experiments. For this topic, excellent overview literature exists, with which one can embark on a life-long journey of listening and, possibly, even creation.

One very good book on this topic is Nick Collins and Julio d’Escriván, The Cambridge Companion to Electronic Music (Cambridge: Cambridge University Press, 2017). Those interested in learning to apply programming in a practical musical scenario should take a look at Gerhard Nierhaus, Algorithmic Composition. Paradigms of Automated Music Generation (Wien: Springer, 2009). Finally, people especially interested in a critical analysis of modern technology would find Mark Coeckelbergh, AI Ethics (Cambridge, MA: MIT Press, 2020) helpful. An especially good introduction into modern AI which I highly recommend is John D. Kelleher, Deep Learning (Cambridge, MA: MIT Press, 2019).

WHERE WOULD YOU LIKE TO SEE MORE RESEARCH DONE AROUND THIS TOPIC?

I believe there should be more research that shows clearly how the current hype around musical AI is misleading because it tends to ignore all the history that has gradually shaped us, the notion of music, and the technologies that we use.

WHAT SORT OF MUSIC SHOULD WE CHECK OUT TO GET A BETTER SENSE OF THE TOPIC?

A great place to begin is the album Bach by Design by David Cope from 1993. It’s a recording of some of Cope’s computer-generated imitations of historical musical styles, played back automatically through a mechanical piano. It is a very good starting point for a listening journey of 20th-century computer music.

Books

News