I wrote up a forum post in a “how to synthesize drones” thread which, I think, contains the most coherent thoughts I’ve put together on the subject. Maybe that’s not saying much, but here it is for posterity, expanded a little bit.
I use the word “drone” in a more general sense than some people, but more strictly than others. If I control a sound in terms of level rather than “playing notes”, I generally consider it a drone. It’s not an absolute rule, but drones usually have a (more or less) fixed pitch. There may be rhythmic accents.
I don’t quite understand how a band like Earth is considered “drone” when they’re clearly playing riffs, have melodies and standard chord progressions and so on. That’s far too loose a definition for me. Nor does it have to be an unrelenting, 25-minute long pure sine wave.
When I create drone-based music, this is what I think about:
- Depth, width, power, distance, gentleness vs. forcefulness, cleanliness vs dirtiness, spectral balance, harmonic structure.
- Texture. Micro-structure, granularity, etc. This can come from FM or other (near-) audio rate modulation, the beating of inharmonic frequencies against each other, repeating delays, granular synthesis, timestretching, the content of any samples used, or other sources. It could be a “natural” and inherent part of the means of sound production, or it could be intentionally added modulation. As an example, the sound of the carrier of a dial-up modem is a steady beep, which I would categorize as having little or no texture, but when the actual signals modulate it, we can hear structure in it even if it’s too rapid for us to follow — that’s a kind of texture.
- The balance between stasis and change in the medium term. Perhaps it’s a weakness of mine, but I want some motion to take the place of discrete notes and melodies. That motion could be the result of random or periodic modulation (including rhythm), “natural” feedback processes, or manual (usually improvised) control.
- Form. That is, structural change over time on the “song” scale.
Simply fading in, holding steady for some minutes, and fading out is usually unsatisfying, regardless of any meta-narrative about separating music from time, or a temporal window on an endless vibration. Changes in volume, timbre, adding or removing layers, changes in harmonic structure or spatial cues or background noise add interest even when they are not the defining feature of the piece. Usually I don’t plan form in advance, but set up opportunities for improvisation and then let the form flow naturally as I record. If that’s not effective enough, I will edit the recording to enhance or expand these structural changes, or reject the recording if I feel it just doesn’t say anything.
I almost always set up at least two voices, because relative variations in level, spatial characteristics or timbre can be much more interesting than absolute variations of a single voice, and because they can lead to shifts in texture or the creation of new textures. Sometimes extra voices have their origin in the original voice, and just involve additional or different processing.
Although I’m talking about drones here, this corresponds quite a lot to Curtis Roads’ concept of “multiscale composition.” As I’ve said before, my act of composition is spread out between pre-recording, recording and post-recording phases — but it’s all composition, even if there are no “notes”, some is spontaneous, and some a reaction. Why not use the ears as a tool of imagination, and not just the brain?