Neurofunk

The neuro drum kit: short, dry, engineered

Neuro drums are the DnB skeleton with every hit shortened, transient-shaped and clipped. The kick is a 100 ms click-and-thump; the snare is three or four layers; the break is almost gone — replaced by tuned metallic ghosts and tight 16th percussion. Nothing rings.

1. Kick: Drum Sampler (12.1) with a short clicky kick, Decay 120 ms, Punch 30% for the attack, Pitch envelope −6 st over 40 ms. High-pass at 50 Hz — the sub owns everything below. Tune the fundamental to 65–80 Hz regardless of key (neuro kicks are not pitched to the sub).
2. Snare: four pads in one choke group on D1 — body (200 Hz, 150 ms), crack (clap or white-noise burst high-passed 3 kHz), rim (1 kHz click), and a ‘metal’ layer (a short Collision or Corpus hit, see next technique). Per layer: Drum Buss Transients +20, then Saturator Hard Curve 6 dB. The composite is clipped, not compressed.
3. Ghost snare: the same snare pad with velocity 40–55, plus a Velocity MIDI effect so ghosts also lose 2 kHz of brightness (map velocity → Drum Sampler filter cutoff).
4. Hats: a closed hat on 16ths at velocity 60 with accents (velocity 110) on the ‘e’ of each beat, not the downbeat — the off-grid accents make the pattern feel faster. Ride on 8ths at −10 dB.
5. Drum bus: Drum Buss Transients +25, Drive 15%, Damp at 8 kHz to tame the clipped fizz; Glue at 4:1 with 1 ms attack and 3 dB GR; Saturator Hard Curve with Soft Clip on as the final limiter. Keep the bus at −6 dBFS peak.

Tuned metallic percussion with Collision and Corpus

The ‘ting’, ‘clank’ and ‘pipe’ hits that fill neuro drum patterns are physical-model resonators excited by short noise bursts — easy to make, easy to tune to the track, and impossible to get from sample packs without them sounding like someone else’s record.

1. Collision: Mallet excitator with Stiffness 70%, Noise 30%, Volume short. Resonator 1 Tube, Decay 200 ms, Material 0.3, tuned to the 5th of the key (G2 for a C-ish track — the 5th avoids clashing with the bass root). Resonator 2 Plate, Decay 400 ms, mixed at 30%, Pitch +7 st. Both resonators in parallel.
2. Record 8–16 single hits at different Material and Pitch settings into a Drum Rack (Resampling input → Drum Sampler pads). Now they are samples: no CPU, consistent, and you can pitch-envelope them in Drum Sampler.
3. Alternative: any short sample (a rim, a snap) through Corpus in Tube or Membrane mode, Decay 150 ms, with Corpus’s MIDI sidechain so the tuning follows the bass notes.
4. Gate with 80 ms hold and a fast release after Drum Buss so every hit is a 100 ms event.

Variations: run the TD-3 or a modular voice (External Instrument) through Corpus with Erosion before it for ‘processed hardware’ perc; or use Drum Sampler’s Ring effect (ring mod inside the device) at a non-harmonic frequency on a snare sample for an instant metallic layer.

The static sub and the rule of separation

Neuro subs are the least interesting sound in the track on purpose: a sine on long notes following the root, so the mid-range layers can be as violent as you like without the low end ever moving. The sub and the mid basses are separate tracks, always.

1. Operator, oscillator A sine, Osc B–D off. Mono voice, Glide 20 ms. Amp env A 4 ms, D 0, S 1, R 50 ms.
2. Write roots in F0–G0 (44–49 Hz) as half and whole notes. Where the mid bass has a rhythmic phrase, the sub can follow the rhythm as 8th notes but never faster — at 174 an 8th is 172 ms, about 8 cycles of 46 Hz, the minimum to read as pitch.
3. 24 dB/oct low-pass at 110 Hz in EQ Eight so the sub stays out of the kick's 60–100 Hz region in level terms — the sub is a layer under the kick, not around it.
4. Sidechain compressor keyed from the kick: 4 dB GR, 50 ms release. The snare does not duck the sub.

Generation 1: FM growl in Operator (ratio FM + feedback)

The first generation of almost every neuro bass is an Operator patch with non-integer FM ratios, an envelope on the modulator level, and the built-in filter with its own envelope. It sounds thin on its own — that is correct. It gets fat in generations 2 and 3.

1. Operator, algorithm with D→C→B→A fully serial (the leftmost algorithm). Osc A: sine, level 0 dB. Osc B: sine, Coarse 2, level −12 dB, envelope A 0, D 300 ms, S 0.3 — this decaying modulator is the ‘yoi’ vowel. Osc C: Coarse 3, Fine 520 (ratio 3.52 — non-integer gives the metallic inharmonics), level −20 dB. Osc D: Coarse 0, Fine 500 (0.5 ratio, sub-harmonic grit), level −24 dB.
2. Operator Filter: low-pass 12 dB, cutoff 600 Hz, Resonance 20%, filter envelope amount +40% with D 250 ms. Pitch envelope: −12 st, 60 ms decay for a knock on each note. Mono, Glide 40 ms.
3. LFO (Operator’s own) at 1/8 synced, triangle, mapped to Osc B level (via LFO destination) at 30% for a rhythmic wobble inside each note.
4. Saturator Medium Curve 8 dB after; high-pass 90 Hz. Record 8 bars of a one-note 8th-note riff with a few octave jumps and legato slides. This is generation 1.

Generation 2: Roar feedback, Shifter and the formant stage

The second generation is where the sound becomes a neuro bass: the resampled gen-1 audio goes through Roar in Feedback routing with the feedback pitch tuned to a harmonic, a Shifter for inharmonic sidebands, and a modulated formant/band-pass stage for vowels. Modulation should be rhythmic and synced so the result chops cleanly later.

1. Roar 1 in Feedback routing: stage shaper Fold, Drive 10 dB, Tone slightly dark. Feedback section: amount 40%, Pitch +7 st (the 5th — tuning the feedback to the note is what makes it sing rather than howl), feedback filter band-pass around 800 Hz. In Roar’s modulation matrix, map Envelope Follower → Feedback amount at −30% so loud notes feed back less (self-regulating).
2. Shifter in Frequency mode: shift +23 Hz (not a musical interval — frequency shifting creates inharmonic sidebands that beat), Dry/Wet 35%. Alternative: Shifter in Ring mode at 1/16 sync for a rhythmic ring-mod stutter.
3. Auto Filter (12.2 redesign) band-pass, resonance 40%, LFO synced 1/8, amount 40%, cutoff centred 900 Hz. This is the vowel sweep; sync it so every 8th note has the same formant shape — that is what makes the chop in generation 3 consistent.
4. Roar 2 in Multiband: low band Soft Sine 3 dB (keeps the 120–300 Hz body solid), mid band Hard 8 dB, high band Bit Crush 6 dB for fizz. Post-filter low-pass 9 kHz.
5. Resample 8 bars (Resampling input on a new audio track, or Bounce in place 12.2). Normalise. This is generation 2 — usually already usable as a ‘wall’ layer.

Variations: swap Shifter for Auto Shift (12.1) with a slow LFO on pitch for a drifting ‘detuned machine’; put Spectral Resonator (Chord mode, few partials, Decay 200 ms) before Roar for a ‘tuned’ ring; use Pitch Hack with a +12 st shift and reverse at 20% for the glitchy high octave.

Generation 3: chop, re-pitch, rearrange (Simpler Slice + Sampler)

Neuro bass writing is editing. The gen-2 recording is sliced at 8th or 16th notes, loaded into a Drum Rack or Simpler Slice, and played back as a new riff: slices out of order, re-pitched ±12, reversed, with different slices answering each other. Then the result is resampled again.

1. Simpler, Slice mode, slicing by Beat at 1/8 (16 slices per bar). Playback Mono, Trigger. Slice to new Drum Rack so each 8th becomes a pad.
2. Write a 2-bar riff using 5–8 pads: a low slice on the downbeat, a bright slice on the ‘and’, a reversed slice (Simpler Reverse) leading into beat 3. Transpose pads ±5 or ±7 st for call-and-response. Keep a 16th of silence before the snare.
3. Pads in the same key range on a Sampler instead: load the whole gen-2 file into Sampler, set Sample Start per zone, and use the Pitch/Osc section’s FM (Sampler’s internal oscillator, ratio 2, 20%) for yet another modulation layer, plus LFO 2 → Sample Start for per-note position scanning.
4. Use Live 12’s Recombine (Reverse and Rotate at 1/8) and Time Warp on a duplicate of the MIDI clip to generate fill variations. Keep what works; neuro riffs are 2 bars that change every 8.
5. Resample (gen 3), then go back to step 1 if it needs to be denser. Three generations is normal, five is not unusual. Each generation: EQ HP 120 Hz, mono, normalise.

The Reese grind layer (Analog + Roar Mid-Side)

Under the chopped layers, modern neuro usually keeps a sustained, slowly-moving Reese for the 100–400 Hz body — the ‘grind’ — so the drop has a floor when the chops leave gaps. It is dark, distorted and mono-low, wide-high.

1. Analog: Osc 1 saw, Osc 2 saw detuned −14 cents and one octave down (Osc 2 Octave −1). Both to Filter 1, LP 24 dB at 350 Hz, resonance 15%, filter env +20% with 400 ms decay. Mono, Glide 50 ms.
2. Roar Mid-Side: Mid channel Soft Sine 6 dB (solid centre), Side channel Hard 10 dB (grind on the edges). Post low-pass 3 kHz in EQ Eight.
3. Utility: Mono below 150 Hz, width 110%. Auto Pan-Tremolo (12.3) in tremolo mode at 1/4 synced, 15% depth, so the layer breathes with the kick instead of being a flat wall.
4. Play whole notes or the phrase’s root changes only; this layer is harmonic support. Sidechain from the kick (3 dB) and from the snare (2 dB, 30 ms release).

Variations: Wavetable Vintage category table with Unison Phase Sync 4 voices instead of Analog for a more modern, phasier grind; or the Reese rack from the Reese rack page with the Grind macro up.

Atonal stabs, sci-fi pads and the ‘wrong note’ hook

Harmony in neuro is sparse: a dissonant stab on the off-beat, a thin pad on a tritone or minor 2nd against the root, a filtered noise texture. These are deliberately small in the mix — they exist to make the bass sound bigger by contrast.

1. Wavetable: a Collection or Distortion category table, position modulated slowly by LFO 1 (0.1 Hz). Unison Random Note, 3 voices, amount 20% for pitch instability. Amp env A 2 ms, D 180 ms, S 0 for stabs; A 800 ms, S 0.8 for pads.
2. Stabs: a two-note chord of root + tritone (F + B) or root + minor 2nd, on the ‘and’ of 2 every other bar. Note Length at 1/16. Meld with the FM or Reed engine, Scrambler (12.2) at 20%, works equally well for stabs.
3. Spectral Resonator in Freq mode at 30% so the stab rings in an inharmonic cloud for a beat; Echo 3/16 ping-pong with Duck so the tail disappears under the bass.
4. Pads: high-pass 400 Hz, sidechain 6 dB from the drum bus — pads in neuro are almost subliminal and only reveal themselves in breakdowns.

Sci-fi FX: risers, downlifters and the drop-out

Transitions in neuro are engineered as carefully as the bass: pitch-rising noise, reversed snare tails, a granular freeze of the bass, and 2 beats of total silence before the drop.

1. Riser: Operator with Osc A Noise (white), amp env A 1.5 s. Automate the Auto Filter high-pass cutoff from 100 Hz to 8 kHz over the 4 bars before the drop, and Shifter pitch from 0 to +12 st over the last 2 bars.
2. Bass freeze: route 1 bar of the gen-3 bass into Granulator III, Grain Size 60 ms, Spray 10%, Scan stopped on a bright slice; automate Pitch +12 st over the last bar of the build. Hybrid Reverb Shimmer with Freeze on for the last 2 beats, then cut everything.
3. Drop-out: the last 2 beats before the drop are silent except the reverb tail (automate the master send or use a group with a Gate). The first kick of the drop carries no reverb.
4. Downlifter after the drop: a reversed, pitched-down (−12 st, Simpler Warp Complex Pro) copy of the riser, 1 bar, −12 dB.

More on this on FX, risers and impacts.