Six lessons from producing electronic music

The power of mutants, complexity-conservation, and art as perishable

I’ve been producing electronic music for roughly two years now. Having written and recorded metal in my teenage years, I thought I would have a good sense of the process. I was very much wrong, partly because of my naive approach to composing back then, partially because my current genre has its own distinct set of challenges, and partially because thinking of beauty in terms of entropy has enabled a new mode of analysis that I find fruitful. So here’s a little dispatch from my pattern distillery.

Also, I was inspired to write this by Stray thoughts on composition by Ulkar Aghayeva — check it out, it’s great.

1. Breed mutants whenever you can

When I started producing and composing, I did not use repetition enough. That was a mistake. Of course, repeat too much, and a piece gets boring; your brain has solved the piece, so it has no incentive to pay attention. But if you repeat with moderate changes, it is quite satisfying. Your brain gets to repeatedly confirm that it has correctly identified a pattern, while at the same time being told that there is more to learn; it is worth paying attention, because the pattern keeps changing a bit, so maybe there is an even grander pattern to be learned. A standard technique is to vary an existing melody, but in electronic music there is a very good alternative: vary the timbre and sound design instead of the melody itself. Morph the timbre over time, either within a single repetition or between different repetitions. When you reach a blank spot in a track where you don’t know what to do, try to mutate something that already exists in your track. Perhaps take a theme that was earlier in the foreground and hide it in the background with a more shy timbre. Your listener might not even consciously realize it is being repeated, yet they get the feeling of being “at home”. The track feels satisfyingly familiar and fresh at the same time. It helps preserve the feeling of a single unified piece, rather than different sections of music strung together.

2. Conservation of complexity

All good music is complex in some dimension.¹ But the bounded nature of the human ear and cognition sets an upper limit on the amount of complexity you can fit in a piece. At some point, if you increase the complexity in one dimension, you must reduce it somewhere else. In the subgenres of electronic music I like, a lot of complexity is found in the textures of the instruments. As a consequence, we must dampen the melodic, harmonic, and rhythmic complexity. Otherwise the whole thing becomes an unsatisfying mess — your sensory and computational resources are spread too thin to detect and verify patterns reliably. You cannot appreciate the subtle sound design if the melodies are too busy and varied, while busy sound design will wash out the crispness of the notes that is needed to follow a busy melody. I was not quite aware of how strong this trade-off was before I ran into it. When I started, I wanted to perform some of Bach’s pieces for organ or piano (harpsichord) in the instrumentation used in the genre of psytrance, but all my attempts were disasters with no redeeming qualities whatsoever. The timbres of the genre do not allow the harmonic and melodic complexity of Bach.

3. A strong core idea is a nugget of gold

When you stumble across a strong core idea, you’ve hit the jackpot. For me they are rare, but when you do find one, the track almost writes itself. With a good core idea, you can write your whole piece around it. It establishes a strong characteristic pattern that forms the spine of your piece. It allows itself to be varied, teased, and progressed almost effortlessly.

So what is a strong core idea? It depends on genre and taste, but for me it typically has these properties: it is a short segment of music that instinctually makes me move. It is a palette of sounds and rhythms that click unusually well together. Strong core ideas seem to be highly fine-tuned. I’ve found these only a handful of times in my journey so far, and I’ve always come across them through luck and trial and error — never deliberate search and planning. I find that I just need to produce music without having a core idea, and then I suddenly stumble across one. At that point, I may just get rid of everything I’ve written so far and start over with only the core idea. Here are a couple of snippets of musical textures that, for me, made it easy to write the rest of the piece:

4. During creation your art is highly perishable

The moment you start producing a track, the clock starts ticking. As you listen to your own piece over and over, you lose the ability to empathize with someone who listens for the first time. You might overload your track with complexity, because you have effectively memorized the various elements when you hear them in isolation. This is a big problem, because the final wall of sound might not be digestible for someone who has never heard the individual components repeatedly. It is imperative to try to avoid listening to your track beyond what is strictly necessary, so that you can emulate the experience of a new listener. Also, the first listen after a night’s sleep is when you are closest to being a fresh listener. Don’t waste it.

5. Transitions are extra important in electronic music

I believe the properties of electronic music make transitions unusually important. In many genres, like classical music, jazz, or rock, the instruments are pretty much fixed. The ear more or less knows what timbres to expect, which helps the brain do pattern recognition more easily. The outcome space is narrowed in advance, reducing the room for surprise. Your brain knows what bundles of frequencies go together to make up the individual voices. In electronic music, on the other hand, especially in the sound-design oriented genres that I like, the listener never really knows what sounds to expect — or not to the same extent. This means that it is easy to accidentally produce too large spikes of unpredictability (i.e., entropy) in transitions between musical sections. So you need to spend effort preparing the ground. First, you need to signify that a transition is coming, so the brain can shift from a state of relative relaxation to readiness to detect new patterns. White noise rising in pitch is a classic trick for this. Second, you can tease the timbres that are about to come to make the task of pattern recognition in the beginning of the new section easier. A classic trick here is to take the reverb tail of one of the instruments that is about to hit and play it in reverse. This way, you can fade the new timbre in while your brain is cruising in the easy mode of the already established section of music.

If you make dance music, which I do, transitions become even more important, because you want the manipulation of energy to be predictable. You don’t want the dance floor to wonder whether, in the next section, they should catch a break or go all in. Transitions create confusion when poorly managed, but when done well, they are a golden opportunity to inject controlled tension before release. The transition into a drop is as important as the preparation for a perfect cadence using a dominant chord in classical music.

6. Reproducing tracks is ultimate exercise

Take someone else’s track and reproduce it from scratch as faithfully as you can. It is without a doubt in my mind the best exercise you can do. Usually, in music production, there is no right answer, and so there is no tight feedback loop. But feedback is crucial for effective learning. When you reproduce a track, you have the right answer in front of you. It is a merciless exercise that constantly pushes you out of your comfort zone. It forces you to listen extremely closely, picking out every little sound and trying to recreate it. It requires you to transcribe melody, harmony, rhythm, and dynamics by ear, and it demands that you reproduce their sound design from scratch. You can’t fool yourself, and it is always painfully obvious how far short you are falling. But it is just ridiculous how much you learn if you stick at it. You discover a bunch of tricks that your favorite artists use, but which you don’t notice when listening casually. You get intimately familiar with all the small choices that are integral to the piece.

Recently I did a complete reproduction of the track Only Solutions by Infected Mushroom, as a learning exercise. Of course, there is still a significant gap. But I learned an insane amount doing it. For illustration, I’ll leave you with my attempt next to the original:

My version:

Original:

1. Different people surely have different capabilities for parsing complexity (in addition to differences, for a given individual, in capability across different types of sounds). So I believe the best music for a given individual is music that pushes up to this complexity frontier, but no further. I wouldn’t be surprised if people like genres where their complexity capability is the highest.