Dub Techno Drums: The Space Between the Hits

Busy isn't moving

A four-on-the-floor kick, a polite hat and a lovely chord on top can still feel like a corridor. The reflex is to add: a rimshot here, a ghost hit there, shaker on every sixteenth. Now it's busier and exactly as static, because there's nothing left to wait for. A hit only lands as a surprise if the ear has been given time to expect it. In sparse drumming, the gaps do most of the work. (The chord and its echoes are covered in the dub techno piece; this one is about the drums underneath.)

There is no one dub techno drum part

The form is wider than any single pattern. In a 2012 interview with Nowamuzyka (external site, opens in a new tab), Rod Modell says the Basic Channel records he liked most were the beatless sides, "just a hook + FX". MusicRadar's 2022 history of techno hardware (external site, opens in a new tab) notes that kicks, snares and hats in dub techno often took "a back seat" to chords and delayed stabs. Moritz von Oswald, at the 2008 Red Bull Music Academy (external site, opens in a new tab), described music in which "you have something going on rhythmically, but it actually stands still at the same time", and patterns that "weren't really defined, but they work with each other and are constantly changing". Those are descriptions of an effect, not drum programming instructions, and none of the settings below are attributed to any of them.^[1][2][3]

So nothing here assumes a particular kick, tempo, rimshot position or swing amount. The experiment is one way to hear what a gap does, nothing more.

Give the kick and bass separate jobs

Decide the low end before the percussion, because everything else leans on it. The kick keeps the pulse. The bass answers it. If both put their weight into the same few tens of hertz at the same moment, the louder or longer one masks the other and the pulse blurs. The simplest fix is rhythmic: the bass plays where the kick isn't, with short notes that finish before the next kick arrives. If they have to overlap, give the sub to one of them and filter the other, and keep the lowest octaves in mono. That is one workable arrangement. A long, boomy kick with a bass that ducks around it is another, and neither is the correct recipe.

A two-bar experiment

This is an invented pattern, not a transcription of any record, and the numbers are starting points to argue with. Count sixteenth steps 1 to 16 per bar, so step 17 is the first step of bar 2, and set the tempo anywhere around 120 to 130 BPM. Use any one-shots: a kick, a short bass note, a closed hat, a rimshot. Keep levels fixed throughout.

Kick, both bars: steps 1, 5, 9 and 13, same velocity.
Bass, both bars: steps 3 and 11, one sixteenth long. Each sits two steps after a kick and ends before the next one.
Closed hat, bar 1: steps 8 and 16, around velocity 45 of 127. Both are the last sixteenth before a kick.
Closed hat, bar 2: step 8 only, nudged about 10 ms late. At 125 BPM a sixteenth is 120 ms, so that is a small lean towards the kick, not swing.
Rimshot, bar 1 only: step 14, around velocity 100. Bar 2 has no rimshot.
Rimshot send: one quarter-note delay, four steps, filtered to roughly 300 Hz to 4 kHz, with feedback low enough that only two or three repeats are audible.

First, build the busy version as a control: the same rimshot at step 14 in both bars, and a hat on every sixteenth that isn't a kick. Then switch to the list above. Same sounds, same levels, far fewer events.

What the missing hit is doing

In bar 1 the rimshot lands on step 14, the sixteenth immediately after the kick on 13. In bar 2 you wait for it and it doesn't come, so for the length of one bar the pattern has a hole. The delay answers: four steps later the first repeat lands on bar 2 step 2, the next on step 6 and the next on step 10, each one sixteenth after a kick. The rhythm of bar 2 is made of the answer to a hit that isn't there, quieter and in different places from the original, and it comes from one hit you programmed, not four.

The hats work the same way at a smaller scale. Bar 1 has two, bar 2 has one that leans late, so the step-16 hat only plays in bar 1. The kick on step 17 still gets its pickup; the kick at step 33, where the two-bar loop restarts, arrives without one, and the pickup returns a loop later. A pickup that comes back after a gap sounds like a decision; one on every bar is furniture.

One tutorial, Attack Magazine's 2021 Beat Dissected on Basic Channel-style dub techno (external site, opens in a new tab), makes a related point, that movement can come from delay rather than heavily swung drums, and adds that this is "by no means a rule". Treat the pattern above as a separate experiment, not a version of that one.^[4]

The kick and bass are the reason any of this is audible. Bass attacks on 3 and 11 never coincide with a kick attack, though a kick's tail can still run under either bass note, so check the decays for overlap. The offset gives the rimshot repeats clear air to sit in, and the filtered echo stays out of the sub range. Add a fourth rimshot, or let the bass hang over the kick on 13, and the space you were designing is gone. One tail, well placed, can say more than four extra hits, and it costs less in the mix.

Ten minutes tonight

Programme the busy version, loop two bars, and listen for a minute. Then delete every hit from the rimshot and hat lines except the ones listed above and add the one delay throw. Loop it for a minute. Next, move the bar-1 rimshot from step 14 to step 12, and the echoes land on step 16 and then bar 2 steps 4 and 8, hard against the bass and hat. Listen for the crowding. Finally, undo the 10 ms nudge and put it back. You don't need a drum plug-in, a new hat or swing; the exercise is working out which absence makes the next hit matter.