the cmrc blog
tmrw> label just released a new piece I have secretly been working on during lockdown:
I was asked to do a release on floppy disk by the Berlin-based label, which of course appealed to me so I tried to remember pieces of music that had been released in this format in the past. Then, I thought of Brian Eno’s ground-breaking album Generative Music I, which was released in floppy disk in 1996 and included Eno’s music that was created with SSEYO’s Koan software. This triggered my imagination and led me to create a generative retro-conceptual piece where I examine my deep love for MIDI, hard drives as archives, algorithms infiltrating and transforming files, and listening requiring time and patience.
This is what I wrote about the piece, which hopefully will give you an idea of my thinking behind this release:
Midlockdown listening :: HDArchive_(date.today)
Hard drives can become personal archives, storing precariously digital snapshots of our life through various data types (from images and videos, to text and audio) in a multiplicity of formats. These files reflect fragmented stories of our own individual histories and having a detailed look at someone else’s hard drive is like having a glimpse at their brain. Our personal digital archive contains memories and revisiting old forgotten folders and files can remind us what we have done in the past and who we once were. The way that each of us organise our digital information is intriguing — hard drives have structures that may give us a clue to the way we organise ourselves, our house, stuff, time, work and personal life, not to mention our thoughts and emotions.
As a musician and composer, I often find myself accumulating files of musical data in my hard drive. They contain not only the music that I listen to, but also my own music, whether that is completed compositions and performances or just unfinished drafts and sketches. My hard drive is a record of my creative process, which involves generating, sonifying, listening and manipulating materials stored as data. Throughout the years I have collected many MIDI files in my computer. I find the format extremely charming: it is an interesting symbolic representation of music, it’s not a score, it’s not a performance, it may be quantised to a particular grid or not, it may contain human expression or it may represent disembodied musical ideas, and what’s fascinating to me is that it always omits musical information because of its limitations as a somewhat outdated protocol. MIDI files come in many different guises that can be interpreted as different forms of musical abstraction that remain symbolic — MIDI is data, and to become music, it needs interpretation.
During this lockdown, because of the isolation I’ve experienced, inevitably loneliness has creeped in and I’ve felt the need of putting sound into the space I inhabit for hours, to keep me company. The type of music I find most appealing to listen to during this time usually has long duration and large structures that can last for hours as the slowness of the day passes. As a former pianist, in these circumstances I find myself drawn to long piano pieces, like Finnissy’s The History of Photography in Sound, or Ives’ Concord Sonata or to recordings of complete works like the Well-tempered Clavier, or Nancarrow’s Studies for Player Piano, or Beethoven late piano sonatas. All of these works I think have something in common, they indulge in long duration, demanding a particular type of listening that requires time and patience. At the same time, these are pieces of improvisatory nature that have such fluency and invention that they convey a feeling of endless creativity, which is at the same time absorbing and terrifying.
When I was asked to do a release in floppy disk by tmrw> label, partly due of the storage limit and aesthetics of the format, I thought I would take the opportunity to, instead of composing a conventional piece or album, write a computer program that would, like an irresponsible DJ, infiltrate and play selections from my own digital musical archive. On Monday, 20 April 2020 at 19:37:50, I unleashed the algorithm into my hard drive, and it spread through its folders and sub-folders searching for all my MIDI files, tracing and choosing small samples from them and infecting, in a selected few, digital mutations and disorders. The result is a 6 hours, 1 minute and 32 seconds long piece made of short fragments of data extracted from my personal MIDI file collection.
I find the sounding result of this experiment thought-provoking as while you experience it, questions surrounding digital creativity, aesthetics and data sense making emerge. While listening to the output of the program, I can’t help but asking myself if what I’m listening to is data or music, or both. Which makes me think that a gap exists between data representing musical information, and our understanding of this data as music. I also find interesting that while arguably this algorithm is not creative, at some points while listening to the music it reminds me of the endless improvisation, variation and creativity of some of the piano works mentioned earlier. The large structure of the music reveals the way in which I organise my hard drive, which contains a mixture of my own compositional biography, including scored compositions, failed ideas, algorithmic experiments and encoded messages, with music that I collect by others which the computer program might have tampered with. The sense I get from listening to the entirety of this monstrosity is that, in the end, what prevails is not the individual characteristics or cultural contexts of the musical fragments or the compositional techniques that were used to create them, but the weight of the history, structure and aesthetics of the archive.
Midlockdown listening :: HDArchive_(date.today) includes 2 floppy disks with 2 large MIDI files containing the result of running the algorithm in my computer. In addition, I have included a track list of all 1389 MIDI files found in my hard drive, detailing the time when they occur in the MIDI files. In the floppy you will also find the code of the computer program so that you can sample your own personal MIDI archive from the contents of your hard drive. To do this, you will first need to modify the code and replace my paths with yours. Then you will start by running the Python script (you can use any Python IDE to do this, e.g PyCharm) which will generate a list of all your MIDI files. Then, you will need to run the SuperCollider code, which might require you to download and install SuperCollider, as well as installing the SimpleMIDIFile Quark, by evaluating Quarks.gui and following the instructions. Once you’ve done this, you will be able to evaluate the SuperCollider code and play the result in the MIDI device of your choice.
I hope you enjoy the results of this project and the sound of my dusty archive. Hopefully it makes for a useful playlist while on lockdown or whenever you need to kill some time