Introduction
Online performance has served as one way for musicians to connect with audiences during COVID-enforced restrictions on gatherings. While online spaces cannot replace face-to-face meetings, it is still possible that their idiosyncrasies can create productive spaces and encounters for performers and audiences. One strand of my research looks into these realities that offline and online performance are currently grappling with, particularly focusing on the situation in Mainland China.
Originally from various parts of China, Li Song, Liu Jia, and Tan Shuoxin are three musicians now living in the cities of London, Karlsruhe, and Cologne respectively. Their geographical separation has led them to form the research project group [ _ _ _ ], through which to develop forms of what they term, “algorithmic computer music collaboration”. In this wide-ranging interview (conducted in March 2021, via Zoom and Jitsi) we talk about their collectivity in which each member’s knowledge and skills are woven together to produce their various outputs. They discuss the technologies they use and develop for their practice, the theoretical and mathematical models that underpin them, their historical influences, and reveal how their modes of practice follow metaphors and realities of intertwinement to create new communicative networks for these times of enforced distancing.
Network, connections
Li Song (LS): I’m Li Song, from Xi’an in Shaanxi Province in China. Now I’m living in London, working as a software engineer. I have been doing computer, electronic music performances since about 2013. Before I came to London, I lived in Beijing for maybe five years where I started playing more with other improvising musicians, and now I’m collaborating with Shuoxin and Liu Jia on this project, which is network computer music.
Tan Shuoxin (TSX): I’m Tan Shuoxin and I come from Beijing. I just came here to Cologne, Germany, for my Masters study, although I don’t study in Cologne but in Düsseldorf, at the Robert Schumann University, Institute for Music and Media. I study Epistemic Media which is not only about sound but also other forms of knowledge, for example, from philosophy and mathematics and other disciplines.
I got to know Li Song very early on, about ten years ago, because we were colleagues in the social networking company, Douban. Outside of work we were also both freelance independent musicians, and we did some things about experimental music and electronic music, although we never had a chance to collaborate together. Then last year I got to know Liu Jia, also via Douban, through Li Song. He told me he has a friend who also lives in Germany, who studies composition and music informatics, maybe I should get to know her? So I wrote a DouYou [Douban’s direct message] to Liu Jia.
I think this, this way of getting to know each other is very interesting – it’s like an unseen network of human relationships. Then we set up a chat group on Telegram, because at that point Song and Jia are in Europe but I was still in China because of the pandemic. We chat a lot. But there is a turning point I must mention: when I finally arrived in Germany in September last year, we all planned to meet. Li Song had his visa to come over to Germany, but then came the lock down [due to COVID].
LS: There was an Algorave in Karlsruhe, where Jia is, and we were all going to meet up during this event (although we were not playing together yet). But more restrictions came, so I couldn’t go.
TSX: The situation became very bad. But I think this is a turning point because from this point on we started to think about how can we live, and how can we communicate, how can we create, in this situation? That is how this project came about.
Liu Jia (LJ): Actually, Song and I have never met in person. I originally come from Qingdao (the city of beer!). Very early on my brother was learning piano and at some point I started to repeat whatever he’s doing, so my mom said, “Okay, it seems like you have more talent!” So I went to learn piano instead of my brother. Anyway, my brother didn’t enjoy it. I played piano to accompany our choir at school. One time we played a piece by a local composer in Qingdao, and I was just improvising the accompaniment somehow, and it worked, so the composer immediately suggested that I should learn composition. I started to learn with him and then I got into the Shanghai Conservatory of Music where I stayed for five years just studying composition at Bachelor level – it’s really a long course. Compared to Song and Shuoxin, I come more from the instrumental composition side, in the western tradition, with the education system borrowed from Russia which includes very strict Solfège and so on…
But in the last two years of my course we had an electronic music course and I immediately got into it. That was a seed planted which explains what I’m doing nowadays, although at the time I couldn’t really understand how it worked technically. I first came to Germany in 2015 to continue studying composition here in Karlsruhe. I found out that there was another course called “Basics of Programming”, that I had a lot of interest in. Back then my German was so bad that I thought [the programming language] Python was actually so much easier to understand! That was the course I could understand the best. After my Masters concluded I thought that if I have so much interest, why not take another Master in Music Informatics? It was there that I got to know a lot of stuff, like how digital sound systems work. I also got to know live-coding and about the Algorave.
Then, last Summer, I saw that someone on Douban had posted a CD compilation from Algorave Tokyo, and there’s a Chinese name on it that I had never heard of. I listened to it and found it very interesting, so I just wrote to Li Song on Douban, telling him I thought what he did was very nice. That’s how we got to know each other.
LS: I just posted this track because I thought it would be interesting to people on Douban. Because there’s many musicians on Douban, and people who’re interested in new music. Then I received this message from Liu Jia and I got to know that she is learning music in Germany. At that time Shuoxin was still in Beijing, but was supposed to move to Germany very soon, so I introduced them. In the end the three of us got together, like a mini group.
Music, source code, sets, collaboration
ES: Li Song, you came from the computer side of things, but have you ever formally studied music?
LS: When I was studying Computer Science in the local university in Xi’an, I did all the basic computer stuff, but it didn’t have anything to do with music or sound specifically. But I was quite interested in music at the time, and I tried some courses online. For example, I did an online course in sound design, that introduced different oscillators, filters, and how various things worked. But at the end the teacher introduced this tool called SuperCollider, saying that you can use this to program to make sound. I thought this must be interesting because I already know programming from my university course, and I have this basic knowledge of sound, I can try to use this too. Combining the programming side of things with the sound was much more interesting for me. That’s when I started learning by myself, through online courses.
ES: It’s interesting that you all bring different things to the table. What are you doing which defines your group?
LJ: I think we are way more than what we present in our performances. We are first of all like friends. We have a group on [the messaging app] Telegram, and the information and communication going on there is really rich. We share what we’re doing, photos, topics on music, and the composers and musicians we like; we are constantly exchanging.
A big part of our communication actually has to do with technical things about SuperCollider. This is a very important matter, and the reason we got so close, so quickly. I think the open-source approach, of sharing code, and the live-coding community, really influenced the way I want to communicate with people. I can hardly imagine that if Song were someone writing contemporary music and I heard his piece somewhere, I would’ve contacted him on Douban. I think this exchange doesn’t exist so well in some other circles. Since I knew that Song and Shuoxin both use SuperCollider, we suddenly had a lot to communicate, to be curious about: What do you use? How do you organise your code to produce sound? So, I would say our group is more than just what we present. This project we just played at the bunker space in Hong Kong is not even the starting point; it is just part of our research and presentation.
LS: We started not just by playing music, but by reading about music. We decided to meet online every Monday night, to read the classic texts about electronic music, like Pierre Schaeffer’s Treatise on Musical Objects. Also, as Jia says, the fact that we’re using this programming language is a very big part of our collaboration, we only need to share text between each other. We wrote the code together for this project, entirely within SuperCollider; I sent the code to Shuoxin and Jia; and they changed it and sent it back; and then I need to send it on to you. I couldn’t imagine how this would work if there’s no text to work on – how would we do it?
ES: It’s very collaborative?
LS: Yes, it just makes more sense to edit the text, and exchange it like a score. Or like source code – this is very much about open-source culture.
TSX: This was my first experience, actually, of collaborating musically with other artists. I have always had this personal problem of communication with other people, so I just do everything by myself. But with Song and Jia I have almost never had this problem. I think first it is the effect of the special situation we are in [due to COVID], and not only because we share the same interests. Every day you can find a hundred people who share the same interests with you, but you aren’t interested in getting to know them all.
I think the key point is the differences between us other than the common interests we share. In mathematics, it is like a ‘power set’ in Set Theory: we are the three elements in a set, and the subsets of this power set present all the possible combinations of these elements. There are {me}, {Li Song}, and {Liu Jia} individually; there are: {me, Li Song, Liu Jia} together; then {me, Li Song}, {Li Song, Liu Jia}, {me, Liu Jia}. And last there exists an empty set { }. 😀
LS: I feel that too. Jia and Shuoxin learned music in the academy more than I did; but I was playing experimental music in Beijing like Shuoxin; and Jia and I know more of the technical side, the programming.
ES: Like a Venn diagram?
LJ: Yes, maybe! This really reflects very well in our performance, which is the “shadow” of this arrangement, the alias or mirror of it. You see that how we perform is basically constantly sending messages to each other, but sometimes just two of us are playing ping-pong [with our messages] and the third person is ignored. Sometimes one is sending messages to him/herself while ignoring the other two. These different combinations, the dynamic of how they are organised, and how to deal with this balance and imbalance, is reflected in our performance.
Dice, hubs, performance
ES: So, you’ve mentioned the messaging, but can you explain how the performance works in general?
LJ: It’s not that hard to understand, actually. The essential difference between us as a group performing together compared to a conventional band, is that we don’t split our sound across various instruments – we are intertwined. For example, we have these three digits: 0, 1, and 2 as our IDs: 0 represents me, 1 Li Song, and 2 Shuoxin. Then we play a game: Imagine each of us has a three-sided die which also shows 0, 1 and 2. The rule is anyone receiving a message needs to send a new message. For example, I start the game by sending a message – I throw the die and let’s say I have a 1, then the message goes to Li Song since 1 is his ID. Then, Li Song also throws the die, for example he has a 2, then he needs to send a message to Shuoxin. Then Shuoxin throws the die, and the message chain goes on and on. Each message would trigger a sound and the sound’s property depends on who sent the message, who the target is, and what the wait time is that the target person has set to respond. But the interesting part is that we can change how the dice appears, in real-time; you could say that my die now has 10 sides and 7 of them show 0, and 3 of them show 1 – so I have a 70 percent possibility that I could send a message to myself and 30 percent to Li Song. But the next second I decide that all sides show 2 – so the message always goes to Shuoxin.
What I am explaining here is like the model of a Markov chain, but we are not stuck in this chain because our initial thinking was how could we create an algorithm that could itself be collaborative? It’s easy to imagine writing an algorithm by yourself, but how could you create an environment for people to collaborate within? You could have the power to write your own algorithm and put it into a bigger structure. This is what we were originally interested in, so we can say that metaphorically the Markov chain is one of our interfaces in the performance, where we plug our own algorithms into the whole bigger one.
TSX: Here I would like to mention the influence of The Hub, and the aesthetics of technology. This group,The Hubfrom the ‘80s, is very important for us. I think they were the first “network computer music ensemble”. The technology which they used to exchange data is still wired, not wireless. From this point I would say the aesthetics of their works had a strong relation to the technology of that time. They built up such a system to exchange both local and global information in real-time – exactly like a hub. Their performance was like a kind of game, each player asserting control and anti-control, by sending each other different data via the wired network. They each play an electronic instrument which doesn’t decide its own sound, but is controlled by the data received from other members of the group. And our approach is very strongly influenced by this arrangement.
LS: The hub is the device that connects them all to each other. Everyone sends a message to the hub, and the hub forwards it to all the other people, sharing the data. We have set up something similar with a wireless network.
TSX: There are two aspects to this arrangement that we are interested in. One is time, the other is place. For The Hub,there is not so much of a localisation issue because they performed in the same room most of the time. But for us a lot of ideas come out of this [issue of] localisation. For our first performance we played on a virtual platform, provided by TOPLAP. But for the second performance with you in Hong Kong, how can we describe our place, our localisation, in the bunker of this performance? This is very interesting. We write the same sound system, but we don’t trigger the sound by ourselves – we trigger the sound which is located in another place. Where does this sound come out? – finally it comes out in the bunker. Who hears the sound? At which place? It’s not only about the time, but also about the place.
Distance, sonification
ES: For the HK performance did you adjust your programme to take that aspect of distance into account?
LJ: Our script is constantly changing, and its sound has a lot of flexibility, but I think the distance certainly influenced how we played. After we rehearsed for the first time in the bunker space and listened to the sound we met up and had some new ideas. We liked the noise, and we would like to have more of it, so we mixed in more noise from the bunker. It’s all about the frequencies.
LS: Technically we didn’t change much because of the distance, but it had more effect on how the performance itself occurred, the structure of the performance. Because of the distance to Hong Kong the messages will take a longer time to travel there, and it will have a small impact on how it sounds. That’s one of the things about networked music: sometimes if the messages drop, the sound will change.
LJ: When I re-listened to our performance, sometimes it’s clear that the connections are also not always static. Sometimes you can hear that there are four or five messages arriving at the same time and so you hear the sounds also come at the same time. That’s a totally different effect for this performance, and it’s part of the network. That’s why it’s very exciting.
LS: It reminded me of when I collaborated with my friend, Tadashi Usami a few years back, we also collaborated over the network [between Tokyo and Beijing]. I think he mentioned that a weird internet connection is highly recommended. It brings very interesting results because it’s very unpredictable and exciting.
LJ: I’m also working on this idea with a study project with my colleagues at my university. This project will sonify this round-trip time into frequencies. If the internet is really slow you can hear that everything is shifting into a higher pitch.
We should also mention Patrick Borgeat at this point, who provided this possibility for our connection. The software we’re using for this is called OSCjunction, which as far as I know Patrick programmed also because of the pandemic. He teaches SuperCollider and Creative Coding at my school and needed to find a good way to teach remotely, while letting the students listen to the sound produced by his code simultaneously, and OSC messages (Open Sound Control protocol) work well for this. But this software also creates a lot of possibilities for musicians and artists like us to collaborate through the network.
LS: OSCjunction is a tool which synchronises the messages between all of us during the performance. I think many people have similar ideas, but Patrick’s version is really general and user-friendly. When I was performing with Tadashi Usami he created a similar tool called OSC Hub, but it was not that reliable compared to OSCjunction.
LJ: How it works is that somewhere there’s a server that we use. No matter what message we’re sending, it is first sent to this server, and then that server sends it to all the connected nodes.
LS: Then we can program the logic of how we react to those messages. Each one of us receives the messages via OSCjunction, and these then trigger the instructions we wrote to make certain sounds.
LJ: There’s actually a lot of hidden consequences due to this linked communication between us. It’s totally possible that if that central server had an electricity failure, we wouldn’t be able to perform. You also had all the technical problems at the Hong Kong end. If one tiny thing doesn’t work at some location, then the others don’t matter anymore because it breaks the whole system.
Topology, braids, play, chat
ES: Shuoxin, how does this relate to your studies on epistemic media?
TSX: My point of view is always conceptual. From the beginning I always talked a lot [with Song and Jia] about what I’ve learned from the seminars with my professor Julian Rohrhuber. He is an interesting intellectual! As one of the contributors to SuperCollider, he is not like the 99% of the users who use it only as a tool for music composition. He uses it to embody his research and meditations on philosophy, mathematics, logic, political economy, and so on.
My own research recently has been focused on topological matters which is a useful way to understand the network. The network is like a topological structure. It is virtual, and there is no relation to dimensions, only arrows from one place to another, or from one person to another. Between different places there are just the projections of things. There are also some aspects to do with set theories, and so on. At the beginning when we were reading and talking about conceptual things, I introduced this subject of topology. The others also think it is very interesting, but it is also very abstract. So how can we create some concrete ideas to realise this concept? To do this we settled on this network collaboration.
ES: In terms of the visual display, you talked about a “braid”. Is that related to topology?
TSX: The braid is designed and coded by Liu Jia. To me the braid is also a part of this topological concept. Because topology can have different presentations, like knots, chains, or braids. We studied these different presentations together and Jia was very impressed by this braid, because it can be a very good visual presentation of our sounds – the three lines represent three frequencies of sine waves, like the three of us weaving the braid together.
LJ: Yes. When Shuoxin introduced the Lacanian topology, she mentioned that Borromean rings could have different presentations. We did a sketch together trying to understand the concept. The Borromean rings and the braid are isomorphic. It is pretty fascinating to think about structures from very different perspectives. Plus that I really love the computer game “Braid” by Jonathan Blow. Then I just named this part of our performance “braid”.
ES: In the performance it seemed that these lines were reflecting frequencies?
LS: It’s a simple logic: each of us controls one of the lines that represent the frequency. But there’s a limitation of the maximum difference between the highest frequency and the lowest frequency. So, you have this competition between the three of us; you try to increase your frequency but eventually you reach the boundary of this limitation.
TSX: At first, we set the frequency limitation to a very, very narrow range, for example we can only change the frequency within 10 Hz , and none of us can escape from this range. We don’t communicate orally during the performance, we just decide by ourselves when to open up this frequency range. So for a long time we just play in this very narrow range, but at an uncertain point someone decides to open this range up. This interaction is interesting.
LJ: It’s almost an experiment or a study in group dynamics. We talked together about how we could play together with these three lines. I thought that there should be some rules: you can’t just control the frequency however you want. We can define a maximum difference, so that we can’t leave each other too far behind. There’s something interesting that can happen if you have these constraints. It’s like a game.
LS: Yes, it’s very much a game.
LJ: It’s also interesting to mention that the ways of communication within the performance are different from each other. It’s not just one model. We have the first model I mentioned, the Markov chain-based logic. We also have this different logic in the braid, where each of us controls a sine wave individually, but within a global constraint. That’s more of an idea of how to present or how to play with the sound materials. But these models certainly play very different roles when we’re playing together, they are different modes of collaboration. It’s interesting how we organise these different modes simultaneously, as musicians, but also like game players.
TSX: I think the visualization is very good too. It is a very direct reflection of this human relationship during our real-time performance. Because there is control and anti-control, autonomy and dependency, and it all happens during the performance. From this visualization, for example of the braid, we can observe this relationship and how it can be visualized.
LJ: We all love sine waves, that’s also a thing for us. So we want to play with them.
ES: The chat that is going on is another communication device. Do you see all these different modes—the braid, the code, the Markov chain relationships, and the chat—as different representations of your interactions?
LS: Yes, I agree. They’re all different modes of exchange, communication.
LJ: In the chat message we can address someone directly. I could ask a question, like: “Shuoxin, how’s the weather in Cologne?” If we view our message chain (the text on the top of the screen, which presents the sound) metaphorically, it could basically also work like that: to send a message to Shuoxin, I just need to address my target to 2. But this could also happen dynamically, musically. I could let the message wait 10 seconds or two milliseconds. This is communication having similar and different structures, happening in different time scales. This is all coming into our way of thinking, of how to organize sound, immediately, because we have this mechanism for the wait time, and so we can organize the pattern. We tell our algorithm how to wait: first half a second, then one second, then wait piseconds, whatever. This is also a way of pattern organisation and collaborative pattern organisation.
LS: This wait time is an important part of our Markov chain because that is what is bringing the rhythm to the sound. The music is actually about these waits. It’s such a primitive concept,the fact that one waits a certain time and then passes it on opens up all the possibilities of sound.
Space, place, sound
ES: How did you feel your Hong Kong performance went, in terms of realising all those ideas? What have you learned from this performance and how do you think you will develop?
LS: What I learned is that software is so fragile, it’s just not stable enough to do very complex things. That’s one of the things we need to work on, to make it more robust.
LJ: For me, the most important thing is we had this space which we were not really in. The sound there is really fascinating. When we usually do algorithmic music, or electronic music, the sound is really pure, we use very simple sound sources. But it sounded so rich in the bunker. I think that this is something we couldn’t easily replace – the physical space and the sound. Another important thing is of course, you, Edward. Because you are the first person who came into our “space”. If we are ids 0, 1, and 2, then you are id 3, right? That’s really inspiring, that we can think about how to invite other audience members into our space, and how could they interact with us? What could we do with more ids? That’s what I’m thinking about.
LS: What’s also very interesting is that what each of us hear is different. It’s down to our own perceptions. We don’t hear the ambient sounds, and even if we hear the environmental sound it may still be different because of the network delay. But in fact everyone is listening to a different thing.
TSX: I think we are the first artist project where not only the audience is absent, but also the performers.
LS: My duo project “no performance’ with Zhu Wenbo [in Beijing] did something like this before. At the start of gigs, we tried hiding in different rooms or playing instruments across the street outside of the venue. We were there, but not there in a way.
ES: I think this is another consequence of the COVID situation. Part of it is how to perform remotely if you can’t be in a live house (or wherever). Then that brings in the whole set of questions about who are you performing to? Where are they? Is it important to know you or they are? Can you work with this somehow? It was also very important for me to emphasise that the bunker isn’t just any space, it is the space where the sound sounds, even if it immediately gets live-streamed out again.
LJ: I think a very important thing that the audience brings is of course themselves. As the audience, when you physically go to a concert, you see and listen to things from your own unique perspective, just like everyone else who is present. There can be many people in the same room, but there can only be one person in each position. But with the live-stream, although every audience member is physically in a different place, the live-stream signal they get is potentially exactly the same – as if their ears were all where we placed the microphones, and their eyes were where we placed the camera. This means that we made a lot of decisions for the audience, which they would have made themselves had they been on-site.
TSX: This situation reminds me of [the artist] Marcel Duchamp, who took a thing out of its original context, and brought it to another context that it didn’t belong to. A lot of sound artists also use this concept. For example, Bill Fontana made a sound project by radio called Soundbridge, transmitting the sound from Cologne train station mixed with the sound beneath the San Francisco Golden Gate Bridge. In its acoustic aspect it is similar for us, in that we take something out of its original context.
LJ: The difference is that we create sound ourselves; we don’t just take something and put it into another situation. We did an experiment like this over Christmas when Shuoxin and I were walking in Karlsruhe Schloss Park. Shuoxin was playing some sine waves and we’re both doing field recordings, and we called up Song to broadcast what’s happening in the park where he was. That was a pretty interesting experience.
Next
ES: You haven’t done any more performances since Hong Kong, so what’s next for the group?
LS: We have a couple of things coming up. In May there will be another online live-coding performance organized by “On the Fly”. We’re also thinking about developing different sections of our project into separate pieces. For example, the braid itself is very interesting and complex, and has the potential to evolve into different things.
LJ: We are also doing more research and further reading together, for instance of some more texts by Pierre Schaeffer and several from The Hub, and about Claude Shannon’s information theory.
LS: So we are learning things together, reading things, and trying to use our time together to plan ahead. After the last two performances we have been thinking about maybe making a release of the recordings. After a couple more performances I think we will have enough material for this.
Exactly! And one day we will be sitting in the same room and playing gigs! I’m looking forward to it …
[Afterword]
ES: My sincere thanks to Li Song, Liu Jia, and Tan Shuoxin for their enthusiasm developing their performance for Hong Kong and for taking the time to share their knowledge in this interview.