Industrial, EBM, Noise & Power Electronics

EBM sequenced 16th bass: Analog or Drift with a hard gate

The defining sound: a mono saw or pulse bass playing every 16th, root–root–octave–root, with a short gate so each note is a separate punch. It is the hi-hat, the bass and the harmony at once. The trick is that it should feel like a machine — fixed velocity, no slides, envelope doing all the articulation.

1. Analog, Mono voice mode. Osc 1 Saw, Osc 2 Rectangle with pulse width ~40%, Octave −1, level ~70% of Osc 1. Detune Osc 2 by 3–5 cents only — EBM bass is tight, not fat-chorused.
2. Filter 1 LP24 at 400–800 Hz, resonance 20–30%, filter envelope amount +30–50%, Attack 0, Decay 120–180 ms, Sustain 0, Release 30 ms. The decay length at 120 BPM is the whole character: shorter = "tick", longer = "bark".
3. Amp envelope: Attack 0–1 ms, Decay 150 ms, Sustain 0.4, Release 20 ms. Add the Note Length MIDI effect set to a fixed 1/32 (or 60% of the original length) so every note is gated identically regardless of how it's drawn.
4. Drift alternative: Osc 1 Saw, Osc 2 Rectangle −12 st, MS2 filter (12 dB, grittier), Drift amount 0–10% (low — the machine shouldn't wander), Mono mode with Legato off.
5. Saturator in Analog Clip at 4–8 dB drive, then a Gate keyed from the kick with the Floor at −10 to −15 dB: a hard, rhythmic duck that's more "gated" than "pumped". Or sidechain with Compressor at fast release (40 ms) if you prefer the pump.
6. Velocity is fixed. Accents come from the octave notes (higher pitch reads as louder) and from a second clip with a filter envelope automation every 2 bars.

Variations: double the line one octave up with Wavetable (Basics>Square, very short amp decay) at −12 dB for the electro-industrial "two-machine" sound; put Auto Filter (12.2) in band-pass on the double and automate cutoff over 16 bars.

Body-music drums: stiff kick, gated snare, march toms

Kick on every beat or the "1 and 3 plus pickup" march pattern, a huge snare on 2 and 4 with gated reverb, sparse hats (the bass already does the 16ths), and clanks instead of claps.

1. Kick: Drum Sampler on a short 909/analog kick, Punch effect, decay 200–300 ms. Pitch the kick to the root (A at 55 Hz for the bass above) or just below.
2. Snare: layer a tight snare with a white-noise burst (Operator noise osc, decay 150 ms) and a metal hit (next technique). Send it heavily to a Hybrid Reverb return (Dark Hall, 1.5–2.5 s, pre-delay 0).
3. Gate the reverb return, keyed from the snare (Gate sidechain), Hold 150–250 ms, Release 10 ms, Floor −inf: the 80s gated-reverb slam that body music inherited.
4. Drum Buss on the group: Drive 25–40%, Crunch 15–25%, Boom off (the bass owns the low end), Transients +15–25%.
5. Glue Compressor 4:1, Attack 1 ms, Release 0.1 s, 3–5 dB GR, with Soft clip on. The stiffness is the point.

Metal percussion: foley hits into Drum Rack, extended with Corpus

Hit things. Sheet metal, brake drums, pipes, springs, chains, a cutlery drawer. Record each hit, chop it to a pad, tune it, and use Corpus to make a small hit sound like a large resonant object.

1. Record 20–50 hits of each object at different force and positions. Mono, close-miked (10–30 cm) for detail; one room take for the big-hall version.
2. Drag the long take to a Drum Rack pad, right-click > Slice to Drum Rack by transient — one hit per pad instantly. Or drag the take into Simpler in Slice mode and Convert to Drum Rack.
3. Tune each pad (Simpler Transpose) so the dominant ring of the metal sits on a chord tone of the bass key. Metal has inharmonic partials; pick the loudest one with a Spectrum device and tune that.
4. Add Corpus to individual pads: Plate or Beam types for sheet metal, Pipe/Tube for pipes. Low Decay keeps it a hit; high Decay plus Inharmonics turns a spoon into a gong. Corpus's MIDI frequency sidechain lets the resonance follow the note you trigger.
5. Choke groups: put open-ring and muted-hit versions of the same object in one choke group, like an open/closed hat.
6. Velocity → Simpler filter cutoff and sample start (small offset) so hard hits are bright and attacky, soft hits dull.

Distortion stacking: Pedal → Roar Serial → Overdrive → Saturator

One distortion at 100% is fizz. Industrial distortion is a stack of moderate stages with filtering between them, each adding a different harmonic signature, so the result is dense but still has a shape.

1. High-pass before driving. Low end that enters a distortion becomes intermodulation mud. HP at 60–120 Hz on the distorted chain; keep the clean low end on a parallel chain or separate track.
2. Pedal in Fuzz for the "broken amp" first stage; Mid +2 to +4 dB so the fuzz has voice. Sub switch off.
3. Roar in Serial routing: Stage 1 Diode or Soft Sine for warm compression, Stage 2 Rectify or Fold for the metallic octave-up grind. Enable the inter-stage filter (band-pass around 1–2 kHz) so stage 2 only sees the mids. Use Roar's modulation matrix: Envelope Follower → Stage 2 drive, so loud hits fold more.
4. Overdrive with its band-pass narrowed (roughly 200 Hz–5 kHz) acts as a "telephone" squeeze and a final compressor; Dynamics low (~20%) for more squash.
5. Saturator Analog Clip as a safety clipper. Then EQ Eight: find the one whistling resonance (sweep a narrow boost) and notch it −4 to −8 dB.
6. Gain-match every stage with its output control so you're hearing tone, not loudness. Live 12.3's A/B compare on the rack is ideal for this.

Self-oscillating noise from Roar Feedback routing

Roar's Feedback routing sends the output back into its input with a delay time and filter in the loop. Pushed high on a simple source (a sine, a voice, even silence plus Roar's own noise modulation), it sings, screams and collapses into tones — a contained version of the no-input mixer.

1. Insert a Limiter at −6 dB ceiling before you touch the feedback amount.
2. Roar: routing Feedback, shaper Fold or Hard, Drive 10–20 dB. Feedback amount 70%, feedback time short (3–20 ms) — at these times the loop resonates as a pitch (1/time: 10 ms ≈ 100 Hz).
3. Raise feedback slowly toward 100%; enable the feedback filter (band-pass) and sweep it — the screaming tone follows the filter. Map feedback time and filter frequency to a Push encoder each: this is the instrument.
4. Modulate feedback time with Roar's LFO at random/noise shape for unstable, stuttering pitch; modulate Drive with the Envelope Follower so the input level controls the scream.
5. Resample the best 30 seconds. Feedback performances never repeat.

Harsh noise wall: no-input feedback, Erosion, Redux and Spectral Time

A HNW piece is a static, dense, full-scale block of noise that changes only microscopically over 10–60 minutes. The art is in the texture: two or three decorrelated noise layers with slowly moving internal grain, and almost no "events". Japanoise-style harsh noise is the opposite — constant violent change — but uses the same sources.

1. Source A — no-input feedback: a mixer's output into its own input (see the dedicated page), recorded into Live. Or internally: an audio track monitoring its own resampled output with heavy distortion and a limiter in the loop (careful).
2. Source B — synthetic: Operator with all four oscillators set to Noise (Noise osc has a colour/tone control) at different filter cutoffs, or Bastl Pizza with both oscillators cross-FM'ed at audio rate through its waveshaper — chaotic, pitched noise.
3. Roar in Multiband: each band a different shaper so the wall has three textures at once. Erosion Wide Noise mode adds stereo hash around its frequency; Redux at low bits and a low rate (2–8 kHz) with jitter turns broadband noise into a granular crackle.
4. Spectral Time's delay with high spray and a pitch shift smears the spectrum so the wall "breathes" without changing level. Automate Tilt or spray over many minutes, nothing faster.
5. For a true wall: route L and R from different sources (two tracks hard-panned) — decorrelated noise is enveloping; identical noise collapses to a centre buzz.
6. Record 20+ minutes, then do the editing as minute-scale crossfades between takes.

Power electronics vocal: distorted megaphone chain

The shouted, blown-out voice delivered through what sounds like a PA in a concrete room. Band-limited to a megaphone range, overdriven to the point where words are half lost, with a short slap that makes it feel physical.

1. Band-limit hard first: EQ Eight HP 250–350 Hz, LP 3.5–4.5 kHz, a broad +4 to +8 dB peak at 1.2–2 kHz. That is the megaphone/horn response.
2. Overdrive with its band-pass narrowed so only the mids distort; Dynamics low = maximum squash.
3. Amp in Lead or Heavy with Cabinet 1x12 for the "small speaker pushed too far" box resonance. Stereo off (mono is more brutal).
4. Roar Single Hard as the last snarl, and Echo with a 60–90 ms slap and small reverb for the concrete room. Gate/Duck in Echo so the slap opens between phrases.
5. Compressor 8:1 fast at the end so whispers and screams sit at the same level — power electronics vocals are flattened, not dynamic.
6. Live performance: run the same chain on a return track and feed it from your mic input track; map Amp Gain and Echo feedback to a Push encoder for scream swells.

Sample-based industrial rock: chopped riffs, dialogue and drum loops (NIN/Ministry-era style)

The late-80s/90s method: guitars recorded, then sampled and re-triggered from MIDI like a drum machine; drums as layered loops; film-and-TV dialogue as texture; everything distorted together on a bus. The result is mechanical rock — riffs that are too tight to be played.

1. Record or DI a guitar playing single palm-muted chugs on low D, an open D5 power chord, and a few scrapes/feedback squeals. Slice to Drum Rack — each articulation is a pad.
2. Sequence the riff as MIDI at 100% quantise; the riff below shows the root chug against the fifth stab. Live 12 Recombine can shuffle the pattern for variations.
3. Run the whole Drum Rack through Amp + Cabinet + Roar on the rack's output, so all samples distort together and glue.
4. Drums: a loop layered under programmed kicks/snares (Simpler Slice mode, Beats warp), Drum Buss Crunch 30–50%, plus a gated snare (see body-music drums).
5. Dialogue/speech: short phrases (from your own recordings or properly licensed sources) through the megaphone chain, Grain Delay or Beat Repeat. Treat them as hooks.

Rhythmic noise: noise gated into a beat with Gate and Beat Repeat

Rhythmic noise (power noise) is a distorted kick plus noise that has been chopped into rhythm. The noise is not a drum sample — it's a continuous wall keyed by a pattern, so every hit has a slightly different texture.

1. Create a key track: a Drum Rack (or simple MIDI clip into a noise-burst pad) with the rhythm you want, routed to nowhere (Master off / volume −inf). Use it as the Gate's sidechain source.
2. On the noise track, Gate with sidechain from the key track, Attack 0.1 ms, Hold 20–50 ms, Release 20–80 ms, Floor −inf. Longer hold = "wall chunks", shorter = "machine-gun".
3. Beat Repeat after the Gate in Gate mode with chance 20–30% for bursts of 1/32 repeats; Variation to scramble grid sizes.
4. Roar Digital Clip over all of it, then sidechain-duck from a heavily distorted 909 kick (Drum Sampler, Punch + Bitcrush effect).

Tape loops of machines: field recordings into closed loops

The Throbbing Gristle / Neubauten tradition: record a factory, a printing press, a washing machine, an escalator, then loop a fragment until the machine becomes a rhythm section. Loop length is the composition.

1. Listen through the recording for a fragment with an internal rhythm (a press stroke, a belt slap). Set the clip loop around exactly one cycle with Warp off — let it be its own tempo.
2. Or: Warp on in Beats mode, set warp markers on the machine's strokes, and the factory now plays at your tempo, in time with your EBM bass.
3. Load the loop in Simpler Classic with loop mode on and a 30–60 ms crossfade; play it chromatically — slowed-down machines turn into drones, sped-up into insects.
4. Tape character: Vinyl Distortion light, Echo with Noise and Wobble raised and feedback 0 (used purely as a tape-flavour stage), or see tape & lo-fi.
5. Run three loops of different lengths simultaneously (e.g. 1.0 s, 1.37 s, 2.1 s). They phase against each other — the rhythm evolves without any edits.

Contact mic and foley: amplifying the inside of objects

A piezo contact mic hears vibration through solids, not air: springs, sheet metal, a cymbal scraped with a file, a table. It's the classic harsh-noise performance tool — a box of junk with a piezo glued in, played with hands and run through pedals.

1. Piezos are high impedance: use a Hi-Z / instrument input or a DI/buffer, or the low end vanishes and it sounds tinny. Gain high — they're quiet.
2. Piezos have a big mid resonance; EQ Eight cut at 200–500 Hz and a gentle HP. Gate to stop handling noise between gestures.
3. Spectral Resonator with its MIDI sidechain fed from a chord clip turns scrapes into pitched, in-key clangs — great to bridge noise and EBM.
4. Then distort. Contact-mic sources love Fold (Roar) and Pedal Fuzz because their dynamics are extreme — the scrape stays quiet, the hit explodes.
5. Foley for industrial: record hammers on anvils, chains in buckets, a drill through a metal can (all with contact and air mics together, panned L/R for natural decorrelation).

Modular noise via External Audio Effect, Pizza and Data Bender

Use the eurorack as an outboard destruction unit in Live's signal path. External Audio Effect sends a track out of your interface, through the modular, and back, so the modular processing sits inside a Live chain where you can automate dry/wet and put Live devices before and after it.

1. Insert External Audio Effect on a return track. Set Audio To to an interface output pair, Audio From to the modular's return. Use the device's hardware latency setting (measure with a click) so the processed signal lines up with the grid.
2. Clock: CV Tools CV Clock Out (DC-coupled output) into Data Bender's clock input so its Bend/Break buffer cuts lock to Live's tempo. Unclocked, Data Bender is pure chaos — useful for HNW.
3. Data Bender's Corrupt section (decimation, dropouts, glitchy buffer errors) is industrial in a box: send a drum bus through it at 30% wet. Nautilus with feedback and Reversal makes metallic smeared tails; Mojave can sit at the end of the chain as a further texture stage.
4. Pizza as a noise source: cross-modulate both oscillators at audio rate, open the waveshaper, record into Live through an External Instrument (MIDI to CV for pitch so you can still play it from Push).
5. Ikarie: use its envelope follower on the returning signal to drive its own cutoff — dynamics-dependent filtering that tames the loud parts.
6. Live's Dry/Wet on External Audio Effect is the performance control: map it to a macro and ride it.

More on routing in hardware & modular integration.

Dynamics in noise music: silence, cuts and density as structure

Without melody or harmony, the musical parameters are density, loudness, spectrum and silence. Power electronics in particular lives on contrast: a minute of low hum, then a full-scale scream, then a hard cut to nothing.

1. Treat a Utility gain lane as the score. Draw it in minutes, not bars: long slow swells over 2–4 minutes, then vertical cuts.
2. Hard cuts: a Gate keyed by a MIDI clip with long notes. The silences must be truly silent (Floor −inf) — the shock of the cut is the event.
3. Interactive density: Envelope Follower (Max for Live) on the vocal track, mapped to Roar Drive on the noise wall — when the performer screams, the wall gets denser.
4. Spectrum as dynamics: automate a high-pass from 20 Hz to 2 kHz over a minute — "loudness" decreases without any gain change. Reverse it for a body-blow return.

Mastering harsh content safely

Noise and heavy distortion are already crest-factor-crushed, so they read louder than their LUFS number, and their energy is in the ear's most sensitive band. Mastering is about protecting playback systems and ears while keeping the aggression.

1. Remove DC and sub-sonics: Utility's DC filter plus a 20–30 Hz high-pass. Rectifying and folding distortion create DC offset that eats headroom and stresses speakers.
2. Tame the 2.5–5 kHz pile-up with a broad −1 to −2 dB on EQ Eight (or Multiband Dynamics on a mid band) — not more; the harshness is the genre. Tame fizz above 8 kHz with the top band compressing at 2–4:1.
3. Limiter with True Peak enabled and ceiling −1 dB. Bit-crushed and clipped material generates big intersample peaks — without true-peak limiting it will distort further on lossy encoding.
4. Loudness: EBM −9 to −7 LUFS; rhythmic noise −8 to −6; harsh noise/HNW often sits at −6 to −4 short-term already — don't limit it harder, it will only get smaller. Power electronics: master the peak sections to around −8 short-term and let the quiet sections be really quiet.
5. Reference on small speakers and a phone at low volume: if it's painful there at a quiet level, the 3 kHz band is too hot.