Lasse Marhaug on playing with KA, September 2020
Since my “instrument” is electronics, both analogue and software-based, I probably feel a closer connection to what KA is doing than if I was more of a traditional musician, meaning playing an an “actual” instrument that is made for playing music based on scales, notes and rhythms. I have no musical training, so I’ve only ever worked with processing sounds through electronic means. Early on I found that one thing electronics could offer that other instruments didn’t was that it could play itself. I could set up feedback-systems that fed the electronics back into itself, which would make it self-play and create repeated sound patterns that I could influence or just leave to run by themselves, to see where it went. This could be done with both analogue and digital means, it worked in the same way and give similar results. When recording I would often set up systems like this and record them, responding to them intuitively, and leave the critical judgment for later when I was to edit and shape the recordings into something I could present as a piece of work. KA seemed to be more sophisticated than these feedback systems, programmed to specifically work in this way.
The main problem with improvising with an interactive human-machine is quite simply that it’s not a human. No matter how many parameters and possibilities we give the machine (limited only by the technical abilities of how much it can process, which is going to at least double every year from now on) it’s still a machine that will only do what we tell it to do. It can only imitate and copy. The key to improvising and composing music is the human ability of abstract and conceptual thinking. A machine can mimic our behavior, but it cannot understand conceptual thinking. It does not understand why something is absurd or interesting. It is completely democratic in terms of sounds and music. It has no taste.
But, given enough machine-power a software like KA can probably mimic our behavior in such a complex way that we don’t see the patterns of mimicking, and we start/choose to believe the machine is making something original, but it’s really just re-shuffling what we feed it. This is why it’s better to look at a machine like KA as an extended and complex mirror. Something that takes what we do and shows it to us in a way that we can benefit from. During the process of experimenting with KA I kept thinking feeding sounds into the system and having it mixed and dictated by previous recordings reminded me of photography. My iPhone, Facebook and Instagram often shows me old photos, things I did years ago, telling me these are my memories, even though I probably wouldn’t think of them by myself. This leads me to the photographer Sally Mann, who once said that “Photographs doesn’t preserve memories, they create them”. She also said “Photographs open doors into the past, but they also allow a look into the future.” I think this is how KA also works - it mirrors and processes recordings (memories) into sound objects for the future. And it helps me to think about it like that, because KA is quite frustrating as a thing of the present. It always seems distant. The greatest aspect of playing with other (human) improvisers is that the music is the immediate interaction, and no matter how fast KA processes, it either seems like it is behind or ahead of me, it’s never here in the moment. It completely lacks the human chemistry, which is crucial to improvisation. That sense of interacting with someone who has an opinion about what you do. KA doesn’t judge, won’t say if the session was a good or bad experience.
This idea I got of mirrors/memories/photographs lead me to think that using a turntable was the most interesting thing to use with KA. Like photographs records carries memories/documents from the past. I have been using turntables as part of my music practice for 30 years. Influenced by artists like Christian Marclay rather than hip-hop/beat-based music I found that I could manipulate a previously recorded time lineage in interesting ways. Often it felt like anti-music, that I was destroying music, but of course it was new music since it was done in a musical context. I liked how immediate the turntable as. To jump from one memory to another in the movement of a hand. So it seemed that the most interesting aspect of testing what KA could do was to feed it with my turntable playing, then I would also have the advantage of dealing with memories. I did a session with Morten Qvenild in my studio, with him setting the parameters without telling me. This was by far the most interesting interaction, and led me to believe that for KA to be an actual useful tool for me I would need to not be able to set the parameters. It should simply be a tool that did something without telling me. A room to enter rather than a house to build. I also thought it shouldn’t have a start and ending point. That it could be a server or an online site that you just plugged into for whatever time you wanted to use it. Something that was constant. And rather than the sound banks being set folders on my computer it could be a vast online archive that I had no control over. So that I had no idea what it would throw back at me. Maybe this is pushing it too far into the realm of random sound processing, but if such an artificial room was created with a musical approach then it could work for me. I remember reading that Derek Bailey during the last years of his life, living in Barcelona, would have open house every Wednesday evening and that anybody who wanted could drop by and play with him. Maybe my wish for KA is coming back to that. To visit Derek Bailey whenever I wanted. At least I find the prospect of everything being defined and available on my own computer rather sad and uninspiring. KA needs to be surprising and unexpected.
