This is the second week of creating a new work every day (see my last post for full explanation), and it’s noticeably harder having less time. Almost halfway through and my stance so far is: this was a good idea for me, maybe less so for the listener. It has helped me manage my mental health; stretch all those creative muscles, and clear my brain of random ideas to make room for the more important ones.
Day 8: Reflection
I spent a good deal of time reflecting on days 1 – 7 so there is no new musical output for day 8. Read that post here.
Day 9: Improvisation #1 by Jennifer Walshe (The Text Score Dataset 1.0)
I’d been playing more guitar and felt like looking at someone else’s work. I’ve got a lot of text scores on file but love these for many reasons. Walshe published ‘The Text Score Dataset 1.0’ after working with a machine learning model trained on a dataset of existing text scores, and I had the pleasure of hearing her speak about it during an online Darmstadt talk. She continues to do inspiring things with AI where the idea is incredibly simple and effective. A huge amount of work went into them and I hope they get performed more. Download The Text Score Dataset 1.0 [Opens PDF]
This is also a little bit of ‘research’ for my PhD, in which I’m looking at AI’s relationship with art and disability.
Day 10: podie #2
This is one of many turntable improvisations I’ve done with an Erik Satie solo piano record that is getting quite dramatically more scratched and warped each time. My turntable performances and recordings (so far) are all ‘live’ and unedited; I decided to take advantage of the format and record an acoustic guitar improv alongside it. I quite like it. Despite being a ‘trained’ (ew) guitarist, I’m not a great free improv guitarist.
Day 11: TEMPI TEXT
AI-generated text scores using common Italian tempo markings as prompts. I’ve been making the most of playing around with GPT-3 since it became available. It is endlessly fascinating, it will be a significant aspect of my PhD research & works but I’m still figuring out how best to use it and, most importantly, how I feel about it.
Day 12: Looks Like Rain
This is probably the day I’m least happy with so far, and made me consider what I’m going to do with these works when the 30 days are over because the quality varies hugely. I’m committing to creating and sharing regardless of how inspiring, creative or convincing the work is because a quality recording/composition is entirely not the point of this.
This is another virtual field recording from The Witcher 3: Wild Hunt. Originally I was thinking about how every time I – or anyone else – say the words “it’s gonna rain” I get the Steve Reich tape loop piece stuck in my head, and it would be cool to re-make it with a video game character. Geralt of Rivia has exactly the right voice for this exercise. This is mostly a joke; one that only I will find amusing. It does beg the question: why would anyone listen to this all the way through?
Day 13: Tracking
This is a draft of a piece that I’m planning for my PhD which will look at AI and mental health. For a first iteration/realisation I’ve used a Max patch to control 10-tone microtonal sine wave chords. I’d like to arrange it for instruments but need to figure out a score that would be open to performer interpretation, in line with the main research aims of my project.
This was my first day back in my studio space for a couple of weeks. The last 2 sessions I’ve spent there I wasn’t feeling well (NB: a purposeful understatement), I wasn’t productive and the isolation of the space was very overwhelming so I didn’t go back for a while.
Day 14: Tracking (score)
A version of a score for ‘Tracking’ with different pitches. It begins to highlight points I’m trying to make or explore with the piece but is not a “good” score. To be continued! For any instruments although I am biased toward strings.
Megan Steinberg 