Sound Design for Dance Music: Bass, Leads & Pads

Nearly everything in a dance track comes from one of two synthesis approaches, and picking the wrong one for the job wastes hours. This sits alongside the wider process covered in the complete guide to AI music production, where sound design is one stage in a chain that ends with a finished master.

Subtractive synthesis basics

Subtractive synthesis starts with a harmonically dense source — a sawtooth, a square, or a supersaw stack — and removes energy with a filter. It is the backbone of house, techno and trance because it maps directly onto how those genres are arranged: a filter envelope opening over 2–4 bars is a build; a filter closing on the last beat of a bar is a stab. Sawtooth waves give you every harmonic, so they respond predictably to a low-pass filter sweep. Square and pulse waves are hollower and work better for basses that need to sit under a kick without fighting for the same 60–100 Hz band.

Additive and FM for extra edge

Additive synthesis builds a sound from individual sine partials rather than carving one down. In practice, few producers build additive patches from scratch — instead they use FM (frequency modulation) as a shortcut to additive-style results: one oscillator (the modulator) shifts the pitch of another (the carrier) fast enough to generate new harmonics rather than vibrato. A modulator set two octaves above the carrier at a low index adds metallic bite to a pluck without needing distortion. This is the fastest way to get the glassy top end on plucks and bells that pure subtractive patches can't reach, because a low-pass filter can only remove harmonics, not add ones the source never had.

Every patch starts with a decision about how many oscillators are doing the fundamental job and how many are decoration. A common mistake is stacking four detuned saws for a bassline that only needed one — the extra layers smear the transient and cause phase cancellation in mono, which is exactly where a club kick and bass need to be rock solid.

• Single oscillator, sub-tuned an octave down: the default for kick-locked bass. Clean, mono-safe, and leaves headroom for the kick's own low end.
• Two oscillators, detuned 3–7 cents: adds width and movement without smearing pitch. Good for leads and pads where you want the ear to perceive one thick note rather than two competing ones.
• Unison stacks of 4–8 voices, spread wide: reserved for pads and trance leads that need to fill the stereo field. Never use heavy unison on sub bass — collapse it to mono and the low end thins out or phases.

Rule of thumb: anything below roughly 150 Hz should be committed to mono at the oscillator stage, not fixed later with an EQ mono-maker. It is easier to design correctly than to repair in the mix — a point covered in more depth in our guide to mixing electronic music.

Shaping the filter envelope

The filter envelope is what turns a static waveform into a percussive event. Attack, decay, sustain and release each answer a specific question: how fast does the filter open, how fast does it settle, where does it settle, and how long does it take to close after the note ends. For a house pluck at 122 BPM, a typical starting point is 0 ms attack, 120–180 ms decay, near-zero sustain, and a short release matched to the note length so the sound doesn't bleed into the next 16th. For a pad, invert nearly everything: 200–800 ms attack, high sustain, and a release of one to two bars so chords overlap smoothly.

Adding movement with LFOs

A static filter cutoff sounds flat under sustained playback, which is why almost every dance patch has at least one LFO doing quiet work. A slow LFO (0.1–0.5 Hz) on filter cutoff gives a pad a slow breathing motion. A tempo-synced LFO at 1/16 or 1/8 on a bass filter is the entire mechanism behind a Reese-style pumping bassline, and the same trick synced to 1/4 notes on a pad's amplitude is a cheap, effective substitute for sidechain compression if you don't have a compressor instance free. Keep LFO depth modest on anything carrying low frequencies — a wide, fast LFO on sub bass causes audible pitch wobble that reads as out of tune rather than "wobbly" in the way dubstep basses intend.

A club kick is rarely one sample. It is usually two or three layers, each doing one job, summed so the total reads as a single tight hit rather than a smear:

1. Sub layer (40–70 Hz): a sine or triangle click, pitched down and given a short amplitude envelope (20–40 ms decay) so it doesn't ring into the next beat at 128 BPM.
2. Body layer (70–200 Hz): the punch. Often a TR-909-style analog kick with a fast pitch envelope sweeping from around 150 Hz down to the fundamental in 30–50 ms.
3. Click/transient layer (2–5 kHz): a very short noise burst or beater transient that cuts through on small speakers and phone playback, where the sub layer is inaudible.
4. Sum and check mono compatibility. All three layers should be dead centre; any stereo width on a kick risks phase cancellation on mono club rigs.
5. Carve space for the bass. If the bassline occupies 40–100 Hz, either duck the bass on every kick hit or notch a narrow dip in the kick body around the bass's fundamental so the two never mask each other.

This layered approach is why hardware-modelled kicks from the TR-909 remain a genre standard — the original circuit's pitch-envelope decay and the way its distortion stage saturates the body layer are hard to replicate with a single synthesized sine, which is why most producers still start from a sampled or modelled 909 body and add their own sub and click layers on top.

The Reese bass — named after Kevin "Reese" Saunderson's 1988 track "Just Want Another Chance" — is two or more detuned sawtooth oscillators run through a low-pass filter, producing a churning, beating texture from the interference between the detuned waves. It underpins drum & bass, dubstep and modern techno bass design.

1. Layer two sawtooths, one tuned 8–15 cents sharp and one 8–15 cents flat relative to the root note. Wider detuning gives a more aggressive, chorused growl; narrower detuning stays smoother and more "liquid".
2. Add a third oscillator an octave down, undetuned, to anchor the sub weight so the bass still reads as a defined pitch rather than pure texture.
3. Low-pass filter around 300–800 Hz at rest, with a gentle resonance boost (2–4 dB) to emphasise the beating frequencies without turning it into a whistle.
4. Modulate the cutoff with a slow LFO (0.3–1 Hz) or an envelope follower keyed to the kick, so the texture swells and breathes across the bar rather than sitting static.
5. Split the range: send everything below ~90 Hz through a mono, unfiltered sub layer, and keep the detuned Reese texture confined to the midrange (150 Hz–2 kHz) so the low end stays solid while the growl does the work higher up.

Reese basses are heavy CPU and heavy on stereo width, so always A/B in mono. If the growl disappears in mono, the detuning is too wide relative to how much of the sound you've placed off-centre.

House and tech house plucks share a recognisable envelope shape regardless of the source waveform: near-instant attack, a decay of 80–200 ms, zero sustain, and a filter envelope decay set slightly shorter than the amplitude decay so the tone closes down just before the volume fades, giving the classic "thock" character. A few practical points:

• Start from a saw or a saw+square blend for brightness, then low-pass filter from around 4–6 kHz down to 800 Hz–1.2 kHz over the decay phase.
• Add a short FM bell layer on the attack transient only (envelope decay under 40 ms) for the metallic click that cuts through a busy mix without raising overall loudness.
• Automate filter resonance subtly across a chord progression — a 2–3 dB resonance bump on the more tense chord in a progression draws the ear without needing volume automation.
• Keep plucks mono-to-narrow stereo (under 30% width) if they're also carrying the bassline's root note; save wide stereo for pads and atmosphere layers.

Acid lines follow a related but distinct recipe built specifically around the TB-303: a single sawtooth or square oscillator, a steep 18–24 dB/octave low-pass filter with high resonance (70–90%), and an envelope with accent and slide programmed per-step rather than per-note. The instantly recognisable "squelch" comes from resonance pushed close to self-oscillation combined with the filter envelope amount set high enough that the cutoff sweeps a full octave or more per accented step.

Pads carry harmonic and emotional information without competing for attention, which means their job is almost entirely about restraint. The Juno-106 remains the reference point for this because its chorus circuit — two slightly detuned, phase-modulated copies of the signal — creates width and movement from a single voice, without the harshness of a heavily detuned unison stack.

1. Use two or three oscillators, lightly detuned (2–5 cents), rather than a large unison stack — this keeps the pad from dominating the stereo field.
2. Apply a chorus or ensemble effect post-oscillator rather than relying purely on detuning; a Juno-style chorus with two LFO rates around 0.5 Hz and 0.8 Hz gives the classic shimmer.
3. Set a slow attack (300–600 ms) and a long release (1–2 bars) so chord changes overlap and breathe rather than cutting abruptly.
4. Low-pass filter above 8 kHz at rest, then automate it open gradually across a build section for tension, closing back down once the drop lands so the pad steps back and leaves room for the lead.
5. High-pass everything below 200 Hz — pads rarely need low end, and leaving it in only fights the bass and kick for space.

When it's time to sit the pad correctly against the rest of the arrangement, the placement and EQ decisions covered in mixing electronic music matter more than any further synthesis tweak — a pad that's beautifully designed in solo but sits in the same 800 Hz–2 kHz range as the lead will always lose the fight.

Modelled and sampled versions of these three machines cover most of the sonic vocabulary of house, techno and acid. Knowing what each one actually contributes helps you choose the right emulation rather than reaching for a generic preset labelled "909 kick".

Faithful emulation matters most for the 303's filter behaviour and the 909's kick saturation — both are non-linear and hard to fake with a clean digital filter. For the Juno's chorus, most modern chorus effects get close enough that the exact circuit modelling matters less than getting the LFO rates and depth right. Treat these machines as starting points, not endpoints: layering a modelled 909 kick with your own sub sine, or running a 303 patch through extra saturation, is standard practice rather than a compromise.

A repeatable patch checklist

Sound design goes wrong most often through accumulation, not bad ideas — a patch gains a chorus, a saturator, a reverb send and three unison voices over an hour, and nobody can say which decision made it worse. Working through a fixed checklist each time keeps patches lean and mix-ready:

1. Choose the role first (bass, lead, pad, texture) before touching an oscillator.
2. Pick the minimum number of oscillators that role needs — usually one for bass, two for leads, two or three for pads.
3. Set the filter envelope to match the genre's rhythmic feel before adding any effects.
4. Add movement (LFO or envelope modulation) only if the static version already sounds usable.
5. Check mono compatibility before adding stereo width effects.
6. Add saturation or distortion for harmonic interest before reverb, so the reverb doesn't smear the added harmonics.
7. Automate one parameter across the arrangement (filter cutoff, resonance, or an FX send) so the sound develops rather than looping identically for 32 bars.

Building a signature sound across a project

A recognisable signature sound usually comes from two or three deliberate, repeated choices rather than one secret plugin: a consistent kick-layering recipe used on every track, a specific detuning amount you always reach for on leads, or a fixed FM bell texture used as a transient layer across a whole EP. Save these as a small custom preset bank rather than rebuilding from scratch each session — consistency across a release is what makes a set of tracks read as a coherent body of work rather than four unrelated experiments.

Once your patches are in place, finishing the track properly — arrangement, automation, and knowing when to stop tweaking — has more impact on the final record than any further synthesis refinement. And when the mix is done, run it through the checks in our guide to AI mastering before final delivery, since even a perfectly designed patch can sound thin without correct loudness and tonal balance at the mastering stage.