The other day I had a thought, which I expressed on the fediverse in the form of a poll, because I was curious if other people thought the same way I did: when someone makes a MIDI version of an existing song, is it a performance, a transformative work, both, or neither?

Of course there's a universe of nuance and discussion I'm leaving out, I just wanted a gut check. But that carries a number of biases and assumptions I have that may not be universal. Maybe people who understand music better than I do don't have a gut opinion about it.

In my totally naive opinion, I'd call it a performance. When I conceptualize someone in the mid-90s making a MIDI of, for example, "Smells Like Teen Spirit", I picture someone sitting down in front of their windows sequencer or composition program with the tablature or sheet music in front of them. I always sort of assumed sheet music was available for a lot of popular songs, for people who want to learn how to play them, I figured you just had to order it from a special catalog for musicians that I've never had an opportunity to see. Maybe I'm off-base. But I see the act of translating a song from sheet music to MIDI notation as analogous to sitting in front of a piano with sheet music and playing a song. I don't usually think of this kind of encoding as performance, but it kind of is, and I think that's neat.

I know that people tried to make MIDIs by ear, because there are a ton floating around that are terrible and don't sound anything like the song they claim to be. I see transcribing or playing a song by ear as a superpower, something almost impossible that maybe a few people in the world with magical golden ears can do. I can imagine learning to identify single notes in a given scale, with a lot of patience and discipline and study, but figuring out which notes go into a chord? Picking apart the notes from all the different instruments playing in tandem? You might as well ask me to pick salt out of seawater with a pair of tweezers.

I still see these MIDIs as a performance, just a bad one. Someone strumming an out-of-tune guitar in not quite the right time signature. You wouldn't call that a transformative work, or at least I wouldn't.

Of course, if you make a remix or a new arrangement of a song in MIDI form, that's a transformative work, the same as it would be if you did that with traditional instruments. But for someone making a Smash Mouth midi in the year 2000, I imagine their goal was to re-create it as accurately as possible.

As I write this, I realize that there are plenty of examples of popular songs with multiple serviceable midi versions that sound drastically different. They can't all be working from the same sheet music or else they'd sound the same, right? Maybe they're starting with, say, the guitar tablature, because that's what's available, and using that as a basis for reconstructing the rest of the song.

Or maybe it's easier than I think to reconstruct a song entirely by ear. If I was a bored tween in the mid-90s with access to midi authoring software, and I wanted to make my mark on the world by being the first one make a midi out of the new Beck single, and I had unlimited summer vacation time, maybe I would've spent hours painstakingly recreating Devil's Haircut as faithfully as I could, pausing and rewinding the tape over and over as I trial-and-errored my way through each second. Bored tweens have spent their time on more trivial and time-consuming projects, for sure.

In this scenario, I'd describe it as a cover, which raises the question of whether a cover is a performance or a transformative work. It's kind of both, but also kind of neither.

It just goes to show that I should save these kinds of thoughts for the blog, instead of throwing a half-cocked idea out on a social media platform. I didn't even fully know how I felt until I was 500 words in, and even with cybrespace's relatively generous 1024 character limit, that's not nearly enough room to really explore a topic I don't know much about and figure out how I feel about it.

It drives home how scary it is that people on twitter are forming impassioned belief systems about important subjects they're not knowledgeable about based on 240-character hot takes that they have to double down on when people make fun of them. What an absolute disaster of a platform for exchanging ideas. What a nightmare for human discourse.

Anyway, "transformative work" won with 8 votes, "both" got 5, "performance" has 4, and nobody thought it's "neither, which I'd agree with. When I made the poll, I was leaning towards "performance", but now that I've thought it through a bit, I think I'd have to land on "both". The idea I had of merely translating a song from one notation (sheet music) to another (midi) is probably not realistic, and I figure some amount of transformation is inevitable.

This has gone far afield of my original thought which was "coding can be a performance, and that's pretty cool", not coding strictly in terms of programming, but using data to create entertainment. I made the analogy to a piano roll, and maya let me know that there is actually some history of composers recording piano rolls in the early 20th century (i.e. "classical" composers, not just ragtime or saloon music, which is the context in which I usually think of piano rolls being used)—and music-knowers don't think of them any differently than a traditionally recorded performance. Neat! Yet another reason I should've waited to blog about it instead of firing off a post, because I might've thought to look that sort of thing up first. Then again, I might not've. Don't know what I don't know, and all that.

I had more to say today, but I'm currently spending hours tearing out my hair, trying to figure out why the update manager is broken. It says my APT configuration is corrupt, and I've googled all the errors and done everything they said to do and it's still broken. Probably going to have to re-install. Cool and good 🦝