ICMC and me at 50
My personal history and connections to Computer Music
Introduction
In June 2025 I attended the international computer music conference (ICMC) which celebrated it’s 50th anniversary and took place in Boston. That summer alsohappend to be my 50th birthday. The last few years after the COVID pandemic I have not been traveling to academic conferences as much as I used to do. I decided ICMC 2025 was a conference I did not want to miss. This theme was an homage to the working spirit of David Wessel (1942–2014), who organized the inaugural ICMC at Michigan State University in 1974 and has been a major influence in how I think about computer music.
ICMC has been instrumental in shaping my academic career in many different ways and I decided to reflect on this personal history as well as summarize some of my experiences and thoughts I had during that conference trip. Quickly this turned into a longer trip down memory lane than expected. Even though I started writing it shortly after the conference, I did not finish it, got distracted with other stuff, and only got back to it in the Fall of 2026 so here we go.
High school and University in Crete
As a high school student I studied classical piano and music theory,
played keyboards in rock bands, and worked as an alto saxophonist with
the municipal marching band of my home town of Heraklion in Greece. I loved playing and learning about music and I could see myself becoming a professional musician although I knew I had a lot of hard work ahead of me. I was also very fascinated with computers and computer programming , to a large degree thanks to my uncle Andreas. He introduced me to computer programming using the absolutely amazing Turbo Pascal 3 (fitting an entire compiler, IDE and debugger in a single floppy disc was an incredible feat). Andreas had invited me to spend the Xmas holidays in Athens at his house and during those 10 days I wrote my first computer programs and was hooked. At the time, I had no idea that these two interests of mine, computer programming and music, could be combined and that this combination would be the focus of my entire career as an academic. It is important to realize that back then most personal computers had very limited hard disk space, memory, and their ability to play and manipulate sound was mostly limited to the blips and blops of computer game chip tunes.
After high school, I studied Computer Science at the University of
Crete. The Computer Science Department, which recently celebrated 40
years of operation, had been recently established. The professors,
many of which had received their PhDs in the United States, were
young, full of energy, and gave us an excellent education in Computer
Science. These were exciting times in Computer Science with personal computers becoming more and more powerful every year and the web emerging (the first internet node in Greece was at the Univerity of Crete). During my undergraduate years I continued studying and
practicing music. This included every two weeks traveling to Athens
during the weekend (an 11 hour ferry boat ride) to study saxophone
with my amazing saxophone teacher, Theodore Kerkezos.
It was during one of these weekend visits to Athens that I first became aware that there was such a thing as computer music . Part of my weekend
routine was to visit the largest music store in Athens (Nakas). There were
multiple floors with the top one being used to show the latest music
gadgets including my favorite ones: digital synthesizers. I read all
the info I could find, played and experimented with the keyboards, and dreamed about buying one. I distinctly remember the glossy flyers advertisting the different configurations of the Roland Synthesizers, the D10, the D20, and the flagship D50. I really could not afford any of them but at least I could entertain the possibility of buying a D10. It is hard to describe to readers today how cool and futuristic these machines looked back then and how fascinated I was by them.
On the same floor there were also crates with books mostly covering the Music Instrument Digital Interface (MID) communication protocol, practical music technology advice, and commercial music software . It was during one of these weekend visits that I discovered a fascinating paperback book that had a more academic focus with the title Foundations of Computer Music. I immediately bought the book and read it.
My first encounter of Computer Music
Published in 1985 the book is an edited collection of articles from 1970s and early 1980s from the Computer Music Journal https://direct.mit.edu/comj published my MIT Press, the oldest academically oriented publication focused on Computer Music. Many of these pioneering works were first presented in the early years of the International Computer Music Conference (ICMC). The first ICMC took place in 1975 so we are the same age. Later on I became an associate editor for the Computer Music Journal, a position that recently returned to.
The 36 articles in the Foundations of Computer Music book include seminal papers by some of the biggest pioneers in the field and people who over time became my personal heroes. I bought my paperback copy probably around 1998 and still have it. I barely understood the articles but thanks to my undergraduate Computer Science classes I managed to at least understand parts of each article and slowly tried tried to learn more about Computer Music. The books included articles on classic topics written by pioneers of the field such as Frequency Modulation (FM) synthesis by John Chowning, Granular Synthesis by Curtis Roads, Synthesis by Analysis and Resynthesis by James Moorer, The POD System for Polyphonic Sound Synthesis and Composition by Barry Truax, Computer Music Notation Systems by Donald Byrd, Timbre Space as a Musical Control Structure by David Wessel, Hearing Musical Streams: Perceptual Organization and Music by Stephen McAdams and Albert Bregman, the SSP Score-Editing Tools by W. Buxton and others. I would learn much more about all these topics and names over the years and after writing this paragraph I decided to work on a separate post discussing this book in more detail. These articles opened up whole new worlds of possibility of what one could do using computers and even though my understanding of them was limited and my access to some of the hardware they described non-existent. By the time I read the book some of the software described was outdated but I could see incredible potential and fascinating questions to explore. I was very curious and decided I would like to somehow pursue graduate studies in Computer Music.
It is hard to comprehend today how revolutionary these early ideas were given the state of computer hardware and software at the time. Here is a video showing the Scriva, the SSSP Graphical Score Editing Tools by Bill Buxton and others at the University of Toronto at the time described in the Foundations of Computer Music Book followed by an interview with Barry Truax discussing real-time granular synthesis using the PODX system developed at Simon Fraser University (an additional reason that I like this videos is because both of these projects took place in Canada where I have lived and work the last 22 years).
During the last year of my undergraduate studies in Computer Science I tried to learn more about Computer Music and processing digital sound signals but back then this was a very niche research area and none of the Computer Science faculty at the time being there were able to help me out. The closest I got I was doing my senior thesis under the supervisor of the amazing Dr. Apostolos Traganitis exploring how pitch detection and pitch shifting could be done using codebook-excited linear prediction algorithms that are used for speech compression and synthesis in mobile phones. The senior thesis later received some kind of academic award from Ericsson.
To give you a sense of the hardware at the time, Apostolos provided me with an extra external Sun Microsystems hard drive to store the few minutes of speech data that I was playing with as the standard Sun workstations we had access to at the University did not have enough hard disk space to store a reasonable amount of audio data for what I was doing. In that project I learned a lot about how to write code for processing audio samples but I still wanted to use computers for more music purposes.
1997 was the last year of my undergraduate studies. Even though part of me considered seriously the possibility of pursuing a career as a professional musician in Greece (everyone I played with back then is now a professional musician) I knew that I wanted to pursue graduate studies in Computer Science and somehow figure out how I could combine my two passions for computer programming and music creation.
Applying to graduate school
With a little bit of help from the primitive early search engines of the time but mostly by directly going to webpages of Computer Science Departments mostly in the USA but also in Europe I identified the handful of places whereI could find academic groups that worked on Computer Music that would be interested in someone like me with a computer science background. I still remember the names and places.
I applied to work with Roger Dannenberg at Carnegie Mellon University, David Wessel at Center for New Music and Audio Technologies at the University of California in Berkeley, Barry Vercoe at the MIT Media Lab, Perry Cook at Princeton University, Paul Hudak at Yale University, Perry Cook at Princeton University, and Julius Smith at the Center for Computer Research in Music and Acoustics (CCRMA) at Stanford University. I also applied as a backup to a few more computer science departments without any computer music activity to work in computer vision which was what I had focused as an undegraduate.
I was very fortunate to be accepted to pursue a PhD in Computer Science under the supervision of Perry Cook at Princeton (all the other computer music place I applied did not accept my application). Over the years of my PhD and after I connected directly and indirectly in different ways with all the names mentioned in the previous paragraph. I spent an amazing year and half in Pittsburgh after my PhD doing a PostDoc with Roger Dannenberg. The work of the late David Wessel has always been an inspiration for me especially his deep respect for improvisation and embodiment in Computer Music. David was instrumental for starting the International Computer Music Conference (ICMC) series in 1974, one year before I was born. The ICMC 2025 theme was “Curiosity, Play, Innovation” as a homage to David Wessel. I never met Paul Hudak but I have read his books and papers, and I love functional programming and Haskell. I never met Barry Vercoe but I became friends, frequently interacted with several of his graduate students such as Paris Smaragdis, Dan Ellis, Eric Scheirer, Youngmoo Kim, and Keith Martin and was influenced by their work.
Princeton and ICMC 1997
In the Fall of 1997, I arrived at Princeton to start my PhD studies. I remember vividly exciting the shuttle from LaGuardia on a gorgeous September sunny day with the trees starting to change colors and seeing the Princeton campus for the first time. I was thrilled to start this new chapter of my life except for one small frustration. The International Computer Music Conference (ICMC) in 1997 took place in Thessaloniki, Greece. I had read about the ICMC and knew that Perry Cook, my PhD supervisor, would go but did not want to take the time off from classes and being a teaching assistant in my first term as a graduate student. Amazingly the website for the conference is still around: https://sites.music.columbia.edu/icmc97/ — including ftp instructions for submitting papers. I remember reading most of the papers of ICMC 1997 and hearing stories about it from Perry. The papers included some very influential works and I wished I could have met the people behind them — I would meet all of them later on:
Open Sound Control (OSC) was introduced in: Wright M, Freed A. Open soundcontrol: A new protocol for communicating with sound synthesizers. In ICMC 1997 . OSC is widely used to this day by a variety of computer music and audio software.
PureData, the open source alternative to Max was also introduced in: Puckette M. Pure Data. In ICMC 1997. PureData or Pd is still actively developed, widely used, and over the years I have had many amazing conversations with Miller Puckette about computer music and learned a lot from his papers and writings.
ICMC 1998 in Michigan
The next fall ICMC took plan at the University of Michigan in Ann Arbor. This time I was able to go and this was the first academic conference I attended in my life. There were a few graduate students from Princeton going so we rented a van — on the drive there I tasked Fig Newtons for the first time and I was hooked. The fall colors were on full display and I loved the university town feel of Ann Arbor. Many papers from that conferences influenced me and introduced me to new people that in many cases would become friends and influence my research.
Get George Tzanetakis’s stories in your inbox
Join Medium for free to get updates from this writer.
Enhanced Vocal Performance Tracking Using Multiple Information Sources by Lorin Grubb and Roger Dannenberg — I can’t recall if I talked to Roger Dannenberg during that ICMC but I had already started reading his papers and being amazed at the pioneering work he was doing. This type of work would later become part of the research field of Music Information Retrieval (MIR).
An Audio-based Real-time Beat Tracking System and Its Applications — by Masataka Goto, and Yoichi Muraoka. This work was truly transformational — it was one of the first true MIR papers, it had an amazing 3D video with polygon dancers following the beat of Gonna Make You Sweat (Everybody Dance Now) which was a refreshing change from the more academic computer music pieces, and I got to meet the incredible, energetic Goto-sensei who since then has been one of the most important MIR researchers in the world.
The GUIDO Notation Format: A Novel Approach for Adequately Representing Score-Level Music — GUIDO is not used as much any more but it was a textual encoding of music notation similar to Lilypond. Although I did not interact with the authors back then much later I got to know Holger Hoos and Keith Hamel at UBC when I moved to Canada to start my job as an assistant professor at the University of Victoria.
Interpolation Error in Waveform Table Lookup another excellent implementation/systems paper by Roger Dannenberg exploring the trade-off between interpolation and memory requirements in wavetable lookup. A good reminder of how much memory constraints mattered back then.
Machine recognition of timbre using steady-state tone of acoustic musical instruments was another important influence to my future research and introduced me to another friend and colleague, Ichiro Fujinaga who later would move to McGill University and become a major figure in the nascent field of Music Information Retrieval (MIR). This was an early paper on single tone instrument classification using spectral features and a nearest neighbor classifier. One of the first tasks I set for myself after the conference was implementing something similar with my own code and that started me down the path of leveraging machine learning techniques for audio and music tasks.
NBody: Interactive Multidirectional Musical Instrument Body Radiation Simulators, and a Database of Measured Impulse Responses by Dan Truman and Perry Cook was the Princeton submission that year. Dan drove the van and I had helped them set up the microphones for the measured impulse responses.
ICMCs after 2000
During the next 28 years the world of music technology and computer music would change dramatically. My 20MB hard disk in 1992 could only hold a single music track of about 3 to 4 minutes. In 1998 this capacity had increased to something like 2GB or about 100 full tracks. The infamous GTZAN dataset released around 1999/2000 was 1000 30-second clips and about 1GB size. The wide adoption of perceptual audio compression (.mp3 files) increased the size of personal music collections and Napster pioneered the idea of digital music distribution. Digital Sampling of musical instruments which in the 1980s was prohibitively expensive became commonplace. Many milestones followed including the iPod, iTunes, streaming services such as Spotify, and the increasing use of recommendation software. Advances in Artificial Intelligence profoundly influenced music technology tools with techniques such as auto-tuning, audio fingerprinting, beat tracking, audio source separation, and music transcription moving from academic curiosities to being widely used by music practitioners.
During the same period, a large number of different specialized conferences appeared focusing on specific areas of computer music that previously would have been included in ICMC. Examples include my main conference ISMIR (the International Conference of the Society for Music Information Retrieval) focusing more on the analysis and extraction of information from music, the New Interfaces for Musical Expression conference (NIME) focusing on various new devices for playing music, the Digital Audio Effects conference (DAFX), Sound and Music Computing (SMC) and many others. I have published in all of them while still occassionally going to ICMC.
One of the most memorable ICMC wast he join ICMC and SMC (Sound and Music Conference) in Athens in 2014. After missing it in Thessaloniki I finally got to experience the conference in my home country. Traditionally ICMC was relatively balanced between the technical program and the music program but in the past 10–15 years the technical program has become less emphasized and typically contains works that fall between the cracks of the other children conferences that are more focused to specific research areas like ISMIR or NIME. Lately I find myself enjoying reading these papers more as they are more creative and open ended compared to the technically clear but mos incremental papers that get published in other conferences. I have also always enjoyed the concert aspect of ICMC and some of fondest music listening memories I have are waking up after a jet lag induced nap in the dark listening to some kind of weird multi-channel tape pieces with alien sound sources moving in space around me.
ICMC 2025
When I heard that ICMC 2025 would be in Boston and it would be the 50th year anniversary I decided I had to go. I did not regret the decisions and really enjoyed my time there. Boston is a city I have always liked, the weather was great, and I got the chance to catch up with old friends and colleagues. I have added to the story a few random photos from the trip.
The person I spent most time with was my old PhD office mate and good friend Georg Essl. Georg is a research professor in Mathematical Sciences at the University of Wisconsin-Milwaukee. We both entered the PhD program at Princeton in 1997 and he was part of the Princeton crew that drove in a rented van to ICMC 1998 in Ann Arbor, Michigan. After his PhD he worked in Ann Arbor for something like 6 years before moving to Milwaukee. His paper had the great title “Topology-Preserving Deformations of Streaming Audio via Box Snake Surgery” . As is typical of Georg’s research he brought some really interesting mathematical concepts from topology into audio processing. I had not seen him in person for more than a decade so it was nice to reconnect and chat multiple times during the conference.
I also enjoyed the keynote speech by Richard Boulanger. Csound was the dominant programming language for computer music during the 1980s and 1990s and it is still widely used. It started as an implementation of the venerable Music V computer music language by Max Matthews in the C programming language. Being written in C made it easier to port to different architectures and more effiicient. The keynote was marked by a lot of memories of Barry Vercoe, a towering figure in computer music and the creator of Csound, which was released in 1985. I never met Barry in person but have known many of his students and Richard did a great job of remembering him and it was very touching to see him get emotional about his good friend of many years.
The other keynote was by Cheng-Zhi Anna Huang 黃成之. I had been following Anna’s research for a while and I was really glad to hear that she was joining MIT and academia. She is an amazing researcher with solid musical and technical background. Her keynote “Algorithms and Interaction for Human AI Creative Partnerships” covered some of her work with generativity AI and how she is interested in making it more of a creative partner rather that the text-prompting followed by generation of a full music track paradigm that is currently dominated in AI music generation. This resonated with a lot of my thinking about this area and I look forward to seeing what she and her research group come up with in the next few years.
Another old friend at ICMC 2025 that I had not seen for many years was Matthew Wright. I got to work with Matt as a PostDoctoral Fellow at the University of Victoria in the academic year 2007–2008. Matt is a fantastic musician, programmer, and researcher and we had worked on a variety of interesting project on analyzing and visualizing pitch and micro-timing information for computational ethnomusicology. Matt had worked for many years at CNMAT with the late David Wessell who founded the ICMC and to whom ICMC 2025 was dedicated to. There was a whole set of presentation by friends and colleagues of David Wessell that was very touching and inspiring. The website for the event is at https://icmc2025.sites.northeastern.edu/wessel-festschrift/ and the 150-page book can be read here: https://bpb-us-e1.wpmucdn.com/sites.northeastern.edu/dist/a/8176/files/2026/02/David-Wessel-Festschrift-FINAL-6.1.25.pdf — a beautiful collections of stories by many of the amazing people that David worked with over the years.
There were also a couple of papers I found interesting. This is not an exhaustive list as there were some parallel sessions and I did skip some to visit friends. Alexandre François presented some interesting work on efficient low latency spectral analysis — Resonate . The building blocks are resonating phasors of different frequencies — similra to a phase vocoder — but the process operates with no buffering and allows arbitrary frequency grids and indvidual setting of the time sensitivity. You can also perform time-stretching and pitch shifting similar to a phase-vocoder. The paper won the best paper award of ICMC 2025. Definitely something I need to check out in practice in the future.
Eric Lyons is a composer and researchers at Virginia Tech that has done a lot of interesting work especially with spatial sound. His paper “Dynamic Generalized Parametric Control of Digital Music Instruments” was more conceptual and providing an interesting way of framing gestural control of digital music instruments informed by his extensive experience in the topic.
Another interesting paper was by Jocelyn Ho, Margaret Schedel, Sofy Yuditskaya and titled “Women’s Labor: Weaving it all Together”. It explored how weaving can be used as a musical instrument. I have known Margaret for a long time and she is a powerhouse composer, performer, researchers whose research frequently challenges in interesting ways existing structures and practices.
“CAFFEINE: Collaborative Affordable Framework For Experiments in Interactive Networked Electronics” by Scott Petersen, Alejandro Mayagoitia describes an open source/open hardware framework that is affordable and enables all sorts of interesting networked interactive computer music experiments. Although I have done a lot of research with sensors and actuators that is similar in nature I have never looked at it from a beginner perspective and a focus in affordability. I could see something like this becoming part of early courses introducing computer music in the future at the University of Victoria.
I should also mention the work I presented “A Real-Time Gesture-Based Control Framework” by Mahya Khazaei, Ali Bahrani, and George Tzanetakis. In this work we showed how the Google Mediapipe framework can be used in conjuction with models trained in Python to perform gesture control using hands and bodies with Max/MSP. My research group has a long history of working on gesture control with early work using custom-made bulky hardware and non-real time motion capturing system and then over the years incorporating new hardware such as the Wii-controller, the Kinect depth-camera, and real-time computer vision using marker motion capture. It has been amazing to see how powerful and accurate the data we get from Mediapipe is with no markers and just using a standard web-camera.
Last but not least, I would like to mention “Arco: A Flexible Audio Processing Framework” by the guru of computer music and my postdoc supervisor Roger Dannenberg. Roger has been thinking about computer music languages and frameworks for over 40 years. Arco is his latest computer music baby and is informed by all his experience. It allows distributed network audio processing and utilizes O2 the extension to Open Sound Control he has proposed. Again, another work that I need to look into more detail. Unfortunately Roger could not make it in person and presented his work remotely.
Summary
During the last 50 years computer music grew from a exotic niche academic curiosity to being an integral part of all aspects of music creation, distribution, and consumption. Many of the technologies that are used in modern Digital Audio Workstations can trace their origins to work that was presented at some ICMC or another. I am grateful I got to experience this evolution from a front row seat and be part of it. ICMC 2025 rekindled my interest in going to conferences as it was the right size for socializing and meeting with people and there was enough interesting research work presented and music played that kept me engaged and connected. Boston is great city with a lot of cool computer music activity happening and I look forward to visiting it again in the future.
