Can a computer replace the music composer?

Eduardo R. Miranda on creativity, music and artificial intelligence

A lot of people get pretty arrogant when they talk about artificial intelligence in the arts ...

This is the end of creativity! That kills the music! AI will overtake us one day!

In reality, the opposite is true. Technologies that work with automation and algorithms help us still to get more creative. From DAWs to MIDI controllers, many of the music technologies we use every day involve some kind of 'intelligence'.

This leads to such difficult (and extremely fascinating) questions as: What is creativity? What is intelligence And how can we make them usable for music production?

In search of answers, I spoke to one of the key figures in the intersection of AI and music: Eduardo R. Miranda.

LANDR: TELL US A BIT ABOUT YOUR BACKGROUND.

ERM: I studied computer science first, and then composition in Brazil. I then moved to the UK to do my PhD. I did my PhD at the University of Edinburgh with a PhD on AI-assisted sound design.

After teaching computer music at the University of Glasgow for a few years, I went to the SONY Computer Science Laboratory in Paris as a researcher.

During my time in France, I also taught computer science at the American University of Paris and composition at CCMIX - the Iannis Xenakis Center for Music Composition.

In 2003 I finally settled in the UK to set up the Interdisciplinary Center for Computer Music Research at Plymouth University, where I currently hold a research professorship.

LANDR: HOW DID YOUR INTEREST IN AI AND MUSIC COME FROM?

ERM: After graduating in computer science, I went back to university to study music. During one of my visits to the library, I came across a double issue of the French trade journal La Revue Musicale with the title Xenakis and music stochastics stumbled. It contained an article about the work of the Romanian-born and Paris-based composer Iannis Xenakis, which I didn't really know much about before.

It turned out that he is one of the most important composers of modern classical music.

Iannis Xenakis

My French wasn't that good at the time, but I immediately discovered Venn diagrams, set theory, logical formalisms, and probability formulas - things that I was familiar with from my previous studies.

That was a moment of revelation: I realized that I could combine my computer science knowledge with my expertise in music. I was intrigued by the idea that computers could be programmed to generate music, and I immediately went deeper into the field of AI, which was still in its infancy at the time.

LANDR: HOW WOULD YOU EXPLAIN "ARTIFICIAL INTELLIGENCE" TO SOMEONE WHO HAS ABSOLUTELY NO IDEA?

ERM: In general, artificial intelligence is defined as the art of programming computers to perform tasks that are classified as intelligent. The problem with this is that intelligence is a pretty difficult concept to define.

Until recently, intelligence was associated with rational thinking, logic and mathematical reasoning, etc. Based on this definition, the field of AI has experienced immense popularity over the past 50 years. A variety of methods have been developed to program computers to mimic intelligence.

Nowadays, however, intelligent behavior is more associated with creativity, emotions, and intuition than with math or logic. And it is now widely recognized that in addition to humans, animals also have a certain form of intelligence that does not require logic or mathematical thinking.

Unsurprisingly, the research community is struggling to make headway in the field of AI because such an intelligence concept is much broader.

LANDR: HOW DO YOU RESPOND TO PEOPLE'S FEAR THAT AI MAY REPLACE HUMAN CREATIVITY AND LABOR FOR ONE DAY?

ERM: Even in prehistoric times, people developed technologies that make work easier, and in some cases even replace it. This is nothing new. It is inevitable that here and there one or the other will also be replaced in the future.

"I prefer to see AI as something that does not destroy humanity, but instead makes use of it."

However, I prefer to see AI as something that does not destroy humanity, but instead makes use of it. For example, I'm interested in developing AI systems that help me be creative. I'm not interested in AI systems that automatically compose complete pieces of music.

Photo: Courtesy Plymouth University

I find pieces of music that are completely generated by a computer unappealing. On the other hand, I find AI systems fascinating, which help me to create music that I would otherwise never have been able to create. I often see computer-generated music as seeds or raw material for my compositions.

LANDR: HOW DO YOU THINK ARTIFICIAL INTELLIGENCE HELP US UNDERSTAND HUMAN CREATIVITY BETTER?

ERM: Researchers working in the field of AI often proceed by isolating certain aspects of intelligent behavior and then building models to mimic them. Using such models, we can conduct experiments aimed at understanding these aspects in detail.

“Personally, I'm less interested in using AI to understand creativity. Rather, my interest lies in making AI usable for my creativity. "

The difficulty here is that creativity is such a difficult phenomenon to pin down. Personally, I'm less interested in using AI to understand creativity. Rather, my interest lies in making AI usable for my creativity.

LANDR: HOW DO YOU THINK KI COULD BE USED TO DEVELOP THE WORLD OF MUSIC INTERFACES BETTER?

LANDR: You once mentioned in an interview how frustrating you find it that there is so little ingenuity when it comes to the design of synthesizer interfaces. To date, we mainly use different types of keyboards when interacting with complex machines such as synthesizers.

ERM: I discussed this problem in detail in my 2006 book "New Digital Musical Instruments: Control and Interaction beyond the Keyboard".

A bit has changed since then: A steadily growing community of users is now developing new musical interfaces. In recent years, numerous musician-friendly programming tools (Max, Pure Data, Python, etc.) and DIY electronics kits (Arduino, Raspberry Pi, etc.) have been developed that have given musicians broader access to technologies with which they can develop their own Be able to build controllers. Very exciting, the whole thing.

Gesture-controlled gloves developed by musician Imogen Heap

However, most of these controllers fail to recreate the feel of a real acoustic instrument. For example, when you play the violin, you feel the vibrations of the strings and the instrument in your hands and on your body.

It is only a matter of time before these users master AI. Then they will be able to develop active musical controllers.

I'm not only thinking of musical controllers with expanded and improved haptics, but also of intelligent musical instruments that have the ability to produce reactions, improvise, and so on. Like the Voyager by George E. Lewis.

LANDR: HOW CAN AI HELP PEOPLE EXPRESS YOUR MUSICAL CREATIVITY BETTER?

LANDR: You have been working for a long time on a brain-computer interface that enables people with neural disabilities to make music. How do you think AI can help people - especially those who live with a disability - better express their musical creativity?

ERM: We now have relatively robust technology that we can use to scan electrical activity in the brain. This activity is called an electroencephalogram, or EEG. The holy grail of brain-computer interface technology is to develop effective methods for training people to produce specific patterns of electrical activity at will, which can then later be detected in the EEG.

The technology we developed in my lab enables people to control musical algorithms with their EEG. However, this control is still quite limited and only works for simple things, such as: B. toggling a few switches to play melodies, or moving faders to increase tempo or volume.

Such control does not involve musical thinking in any way, but the desire to flip a switch and move a fader, which could basically control anything but music.

My current research is how I can improve this scenario. Ultimately, my goal is to recognize musical thinking in the EEG and then use it to generate music. We are still in the dark, but I am convinced that it will be possible in the future.

LANDR: WHICH OF THE PROJECTS YOU ARE WORKING ON IS YOUR FAVORITE?

ERM: In addition to the aforementioned project, I'm currently working on developing new types of music computers. Computer science has played a crucial role in the development process of the music industry over the past 80 years. It is highly likely that future technological developments will continue to have an impact on music.

I support research on bioinformatics and music. My team and I have already developed an interactive biocomputer in a feasibility study that operates with living, organic components that have been cultivated on a circuit board.

I have already composed two pieces with this system, Biocomputer Music and Biocomputer rhythms, both of which have already been played live several times.(You can see a short documentation about it above).

***

Eduardo R. Miranda doesn't even think about quitting. He recently did an e-book called Mind Pieces: The Inside Story of a Computer-Aided Symphonyreleased. In it he gives an insight into his ideas on how to combine technology and intuitive musical creativity. You can find more information about his workhere.

Annika loves strange stories and dazzling characters. She writes about music and everything else she can get her hands on.