ABLETON 12 SOUND DESIGN ATLAS

Dubstep Forum Production Digest — The Scene's Own Technique Notes

Advice from Dubstepforum's production threads (2007–2015), from the Production Bible index to the DSF Q&A sessions, turned into Live 12 stock-device recipes. Covers sub and sine writing, kick and sub, half-step and 2-step drums, snares, skanks, dub delays, resampling, gain staging, sound-system translation, dubplate prep and DJ-friendly arrangement.

Difficulty
Intermediate
Live edition
Standard and up
Max for Live
Not needed

From 2006 into the early 2010s, Dubstepforum's "Production, Hardware & Technical" board was where the scene compared notes in public. The forum collected its best threads in an index, The Dubstep Production Bible (started 2007; most of its links are dead now), and later replaced it with a second guide (2010). It also ran hosted Q&A sessions where accounts presented as working producers answered members' questions. This page boils that advice down into Live 12 recipes built from stock devices.

Read it the way the forum meant it. Forum advice is technique, not documented history. A post says what one person did or recommended on a given day, with the software of that year (FL Studio 3xOsc, Albino, Massive, Live 7's Simple Delay). It is never proof of how a particular record was made. Each technique below is summarised in my own words and linked to its thread. A username appears only where the post is public. Each one carries a confidence note: consensus (several posters agree), single poster, or disputed. The forum identified Q&A guests by account name and the host's introduction, which I could not verify independently, so those answers are marked "attributed". Where the 2007 tool is obsolete, the Live 12 equivalent is given. Where a modern edited version of the same idea exists, a "see also" link points to Attack Magazine's technique section or to The VICE Oral History of Dubstep (Lauren Martin, 2015). Everything here assumes 140 BPM (1/32 = 54 ms, 1/16 = 107 ms, 1/8 = 214 ms, 1/4 = 429 ms, 1 bar = 1714 ms) and Live's octave names (C3 = middle C, so the "E1 = 41.2 Hz" of the posts is Live's E0).

Already covered elsewhere on the site, so linked rather than repeated: sine writing and glide (Deep dubstep bass), sub fundamentals (Sub bass), wobble (Wobble and growl bass), reese (Reese bass), the genre frame (Deep dubstep), drums (Deep dubstep drums), a whole-track walkthrough (Deep dubstep track), rig-aware mixing (Sound system mixing) and record analyses (Famous deep dubstep tracks II).

Threads used

TopicThread (opened)
IndexThe Dubstep Production Bible (2007, 13 pages) · Huge Guide to Producing Dubstep! (2010)
Deepwat makes it deep? (2010)
Sub[production bible 2] Sub-bass (2008) · your technique for deep, clean sub bass (2009) · deep sub bass with movement. (2010) · sub bass click (2012) · Dubstep in the key of? (2014)
Kick and bassKick & Bass Frequencies...? (2008) · [production bible 2] Random bass tips (2008) · [Technique] How to make a kick stand out (2008) · sub bass fighting 808 boom kick (2010)
Wobble, reeseWOBBLE BASS GUIDE (2008) · [Production Bible 2] The Reese Bass (2009)
Drums[Production Bible 2] Drums (2008) · Dubstep Swing/Groove? (2008) · How much swing to add to dubstep drums... (2008) · Bad idea to put your main kick on triplets? (2009) · kick on 1, snare on 3 is dead (2010)
Snares, kicksA Good Snare Source? (2010) · Weighty Snares? (2015) · Burial Cavernous Kick Drums (2011)
Skanks, chords, dub FXReggae skank question!? (2010) · Deep techy dub chords? (2008) · Dub (feedback) delay in ableton live (2009) · root's dub like delay thingy (2009) · Share your delay techniques or secrets! (2011) · Reversed Reverb & Reversed Delay (2009)
ResamplingFor those who want to get into resampling: (2009) · TUTORIAL: 'Single wavecycle' resampling (2010) · What's been your most unconventional production technique? (2008)
Mixdowngain structure and mixing aka THE MONEYSHOT THREAD (2009) · Random Production Tips Thread (2007)
Sound systemsproduction tips for club systems (2008)
Mastering, dubplatesPreparing a track for dubplate (2007) · Preparing your tracks for professional mastering... (2011)
Arrangement[Production Bible 2] Song structures found in Dubstep (2008) · Techniques for Keeping Arrangements Interesting (2010)
DSF Q&A sessions (attributed)Breakage (2009) · Ramadanman (2009) · Jack Sparrow (2009) · Distance (2009) · Pinch (2009) · Cyrus (2009) · Phaeleh (2011)
On the Q&A sessions: in each one a forum host introduced the guest with a bio and links, and an account carrying the artist's name answered in the first person. That is the forum's attribution. It is not independent verification, so the answers below are labelled "attributed". Some posters' signatures list labels or a mastering business; this page mentions them only where a signature says so, and does not treat a signature as proof of identity either. Handles that happen to match a producer or company name (for example "ruckspin" or "jtransition") are treated as plain usernames.

What "deep" meant on the forum

Space as the instrument: rests, sparse parts, subs over mids

Arrangement · Level: essential · Genres: deep dubstep, DMZ-era, Medi-style · Confidence: consensus

Quick start

Load
Operator, Gate
Set
  • Sub clip draw the notes, then delete whole half-bars
  • Operator Release 40–80 ms so each rest is silent
  • Foreground one element at a time: when the sub moves, the chord stays quiet
  • Noise snare: Gate Threshold just under the hit, Hold 60–90 ms, Release 40 ms
  • Space one dub throw into the gap, then silence
You should hear
A sparse groove where the gaps are as loud as the notes and each part gets the foreground in turn.
Live edition
Suite

What the forum advised. In wat makes it deep? (2010), the answers agreed with each other. Basic A said: a few grooving parts, atmosphere, sub over heavy mid-bass, drive used sparingly, half-bar and whole-bar rests, delays. grooki said to leave room for the listener and study how old dub reggae uses gaps. JFK said to avoid enormous white-noise snares. paravrais put the depth down to short, sparse melodies. A secondhand line credited to Loefah, "space, pace, bass", gets repeated. See also: in the VICE oral history, Loefah himself describes the gaps in half-step as an instrument.

  1. Write the rests first. In the sub clip, draw notes and then delete whole half-bars. Keep the instrument's release short (Operator Release 40–80 ms) so each rest is genuinely silent.
  2. Give one element at a time the foreground. When the sub moves, the chord stays quiet; when a chord or siren throw speaks, the sub holds or rests.
  3. Make the space audible with sends, not with more parts: one dub throw into the gap, then silence.
To keep a noise snare in a deep tune, do what alex_l described in the same thread: gate or truncate it so the attack survives and the tail goes, and mix it lower than a brostep snare. In Live, put a Gate (Threshold just under the hit, Hold 60–90 ms, Release 40 ms) or Simpler's amp envelope (Decay 120 ms, Sustain 0) on it.

Sub bass and sine writing

A low-passed square under the sine, so small systems hear the pitch

Instrument: Operator · Level: essential · Genres: deep dubstep, all 140 · Confidence: consensus (2007)

Quick start

Load
Operator, EQ Eight, Utility
Set
  • Operator additive algorithm; Osc A sine 0 dB; Osc B square, same ratio, −24 dB
  • Operator filter LP 24, 220 Hz, Key 50%
  • Osc B level raise until a phone speaker shows the pitch, then back off 3 dB
  • EQ Eight HP 25 Hz 48 dB
  • Utility Bass Mono, Width 0
You should hear
A sub whose pitch reads on a phone, yet on a big system feels the same with Osc B on or off.
Live edition
Suite

What the forum advised. On pages 2 and 3 of the Bible (2007), victorxray explained that rooms and ported cabinets are least predictable around 40–60 Hz, where a note can vanish or boom. His fix was to blend a little low-passed square, or lightly distort the sine, so the odd harmonics carry the pitch. osk agreed. In another 2009 thread, dj vision, Brisance and magma settled that EQ on a lone sine does nothing that a level change wouldn't: there are no harmonics to shape. Dated: the posts used FL 3xOsc and Albino. Live 12 does the same job inside Operator with a parallel chain.

  1. Operator: choose the algorithm with all four oscillators side by side (additive). Osc A sine at 0 dB. Osc B square wave, the same Coarse ratio, around −24 dB.
  2. Turn on Operator's filter (it affects both oscillators): low-pass 24, cutoff around 220 Hz, Key Tracking near 50% so higher notes stay similar.
  3. Raise Osc B until a phone speaker shows the pitch, then back it off 3 dB. On a sub it should feel no different with B on or off.
  4. The alternative from the same posts is light drive on a parallel band. The site's version of that is tape-saturated sub harmonics.

Single-cycle subs: one looped wave in Simpler

Instrument: Simpler / Wavetable · Level: intermediate · Genres: deep dubstep, 140 · Confidence: consensus

Quick start

Load
Simpler, EQ Eight, Utility
Set
  • Sample one cycle from the steady middle of an 808, zero crossing to zero crossing
  • Simpler Classic, Warp off, Loop on, loop = 1 cycle, Loop Fade 0–1%
  • Voices and Glide Voices 1, Glide 60 ms
  • EQ Eight HP 25 Hz, LP 120 Hz
  • Tuning Transpose until a C plays as C, checked on the Tuner
You should hear
A sub with the 808's waveshape and none of its decay, sliding between notes without retriggering.
Live edition
Standard and up

What the forum advised. In the 2008 sub-bass chapter, jobbanaught looped a single cycle from the middle of an 808 in a sampler. That gives a sub with the 808's waveshape and none of its decay. futures_untold's 2010 single-wavecycle tutorial set out the method: choose one cycle, cut at zero crossings, add tiny fades, loop it in a sampler, map it across the keys. If it sounds buzzy, use a longer cut or play lower. Mushroom Buttons loaded the cycles into Ableton's Simpler with loop and glide on. In his 2011 Q&A, Phaeleh (attributed) described rendering a sine at three octaves and looping it in a sampler.

  1. In the 808's sample editor, find a stretch in the middle where the wave is steady. Select one cycle from zero crossing to zero crossing (zoom right in). Consolidate it, or drag it out as its own sample.
  2. Simpler: Classic mode, Warp off, Loop on. Set Loop Start and Length so exactly that cycle repeats; Loop Fade 0–1%. Map the root by setting Simpler's Transpose until a C plays as C (check with the Tuner).
  3. Voices 1, Glide on for slides without retriggering, which was Mushroom Buttons' point.
  4. The Live 12 shortcut: drag the same cycle onto a Wavetable oscillator to make a user wavetable. Wavetable treats it as a playable, tunable oscillator, with its sub and filter on hand.
A cut that is not exactly one cycle plays slightly out of tune. contakt321 suggested checking the tuning with an analyser. Use Live's Tuner device, or Spectrum on a held C.

Shaping the sine: roll-offs at both ends, drive on a send

Effects: EQ Eight, Saturator, Roar · Level: intermediate · Genres: deep dubstep, Tectonic-style · Confidence: attributed Q&A answer, plus single posters

Quick start

Load
EQ Eight, Roar, Utility
Set
  • Sub: EQ Eight Low Cut 48 dB at 35–40 Hz, Bell +2 dB at 55 Hz (Q 1.2), High Cut 12 dB at 600 Hz–1 kHz
  • Return A: Roar Single, Soft Sine, Drive 30%, Tone −20%
  • Return A: EQ Eight HP 90 Hz, LP 1.5 kHz
  • Return A: Utility Mono, −12 dB
  • Send A pre-fader, so the harmonics stay put while you ride the sub
You should hear
A rounded sine sub with harmonics added only above the fundamental, so the low note never shifts.
Live edition
Suite

What the forum advised. In his 2009 session, Cyrus (account "random trio", attributed) described his sub. It starts as a generated sine shaped with envelope and filter. He rolls off below about 35–40 Hz, lifts around 50–60 Hz, rolls off the top somewhere between 500 Hz and 2 kHz, and puts distortion on a send rather than on the channel. He doesn't always compress it. In the 2007 Random Production Tips Thread, narcossist gave the same idea in a different form: put the sub on a mono group and get presence from high-passed sends rather than insert effects.

  1. On the sub track: EQ Eight, band 1 Low Cut 48 dB at 35–40 Hz, band 2 Bell +2 dB at 55 Hz (Q 1.2), band 8 High Cut 12 dB at 600 Hz–1 kHz.
  2. Make a return just for the sub. Roar in Single mode with a soft curve (or Saturator Analog Clip, Drive 8 dB), then EQ Eight high-passing at 90 Hz. The return adds harmonics only above the fundamental, so the fundamental can't change phase.
  3. Set the sub's Send A pre-fader if you want the harmonic level to stay put while you ride the sub fader. Set Utility on the return to Mono.
  4. Leave the compressor off unless the line has real dynamic problems. A sine's level is already steady.
The site's split-rack approach (sub + mid split rack) is the in-rack version of the same idea. The send version lets several bass tracks share one drive return.

Movement in the sub: modulate level, not cutoff

Modulators: LFO (M4L), Auto Pan · Level: intermediate · Genres: deep dubstep, halfstep · Confidence: consensus

Quick start

Load
Operator, Auto Pan, EQ Eight, Utility
Set
  • Auto Pan Amount 20%, Phase 0° (tremolo), Rate 1/2 bar sync, Shape sine
  • Depth a dip of about 2–3 dB
  • EQ Eight HP 25 Hz
  • Utility Bass Mono
  • If you move cutoff shallow, rate below 0.5 Hz
You should hear
A sub that swells gently in level while keeping its full weight, not thinning like a cutoff wobble.
Live edition
Suite

What the forum advised. In deep sub bass with movement. (2010), the original poster said a filter LFO made his sub thin. JemGrover, Erebus-7 and grooki answered: move the volume instead, keep it subtle, or lay a low-passed reese or noisy patch over a steady sub. Several posters (amphibian, Basic A, Sharmaji) said this kind of moving sub is often a low-passed reese with cutoff and level nudged only slightly, crediting Breakage. That is a secondhand attribution and not evidence. district posted a long chain that ended with a very slow, shallow cutoff LFO (about 0.4 Hz), and he warned that cutoff modulation loses low end.

  1. Auto Pan as a tremolo: set Phase to 0° so both channels move together, Amount 15–25%, Rate synced to 1/2 or 1 bar. A dip of about 2–3 dB is enough on a rig.
  2. For the low-passed-reese route, use the existing reese with a sine under it. Low-pass the reese at 150–200 Hz and let only its beating show through.
  3. If you do move cutoff on a harmonic-rich sub, keep the depth shallow and the rate below 0.5 Hz, and check on a big system that the bottom is still there.
The site's detuned breathing sub gets the same swell from two oscillators inside one Operator. That is mono-safe, unlike the forum's habit of detuning two separate tracks.

Killing the click at the start and end of sub notes

Instrument: any sub synth · Level: beginner · Genres: all · Confidence: consensus

Quick start

Load
Operator, EQ Eight
Set
  • Operator Osc A envelope Attack 4–8 ms, Release 15–30 ms
  • Voices 1, legato with Glide
  • Retrig Osc A phase retrigger on
  • EQ Eight High Cut at 150–200 Hz
You should hear
Sub notes that start and stop cleanly, with no click and no level jumps between notes.
Live edition
Suite

What the forum advised. In sub bass click (2012), the answers converged. The click comes from the wave being cut away from a zero crossing. Add a few milliseconds of attack and release, keep the patch monophonic, and low-pass the sub because the click is high-frequency content. VirtualMark warned that a long release makes the sub "flabby". Sparxy pointed out that editing the bounced WAV only treats the symptom.

  1. Operator Osc A envelope: Attack 4–8 ms, Release 15–30 ms. Voices 1, legato with Glide for overlapping notes, so slides don't re-attack.
  2. Turn on Operator's oscillator phase retrigger (Osc A, Retrig) so every note starts at the same phase, then add the few ms of attack. This avoids both clicks and level differences between notes.
  3. EQ Eight High Cut at 150–200 Hz on a pure sine sub removes any click energy that remains.

Choosing the key for the sub

Writing · Level: essential · Genres: dubstep, deep · Confidence: disputed

Quick start

Load
Operator
Set
  • Resting note F0 / G0 (43.7 / 49.0) as the common sweet spot
  • Lowest D0 / E0 (36.7 / 41.2), depending on the melody
  • Avoid C0 (32.7) as a resting note for club rigs
  • Moving lines can dip below the tonic, e.g. G0 down to D0 in passing
  • Fifth layer (Operator) Osc B ratio 1.5, Fine a little below, −8 dB, low-passed
You should hear
A sub root that carries on most systems, with deeper notes only in passing.
Live edition
Suite

What the forum advised. In 2009 Sharmaji said to pick notes, not frequencies: low D, E and F rumble well. Five years later, Dubstep in the key of? (2014–15) gave several answers. _Agu_ said a C that low is too deep for most systems. He also noted that tuning kick, snare and sub all to C forces the snare to an awkward 130 or 260 Hz. Banesy used F and G. Pentaguatti picked the kick first and tuned the roots to it. outbound, who described himself as a mastering engineer, said a line that sits on one note wants a higher key, while a line that moves can dip below the tonic. test_recordings and smalltock said harmonics matter more than the key. NinjaEdit's trick: two sines a fifth apart, the top one slightly flat.

Root (Live name)HzForum view
C032.7Too low as a resting note for most club rigs (the posts call it C1)
D0 / E036.7 / 41.2The lowest posters would go, depending on the melody
F0 / G043.7 / 49.0The common sweet spot ("F for the balls, G for the chest", Banesy, said as a joke)
A0–B055.0–61.7Safe on small systems, less weight

The site covers key choice per rig in sound system tuning and Note to Hz.

Kick and sub

Per-band ducking: the sub keyed by the kick's lows, the mid-bass by its click

Effects: Compressor (sidechain EQ) · Level: intermediate · Genres: dubstep, 140 · Confidence: single poster, widely quoted

Quick start

Load
Compressor, Utility, EQ Eight
Set
  • Sub: Compressor sidechain from Kick, EQ Low Pass 120 Hz, Ratio 4:1, Attack 0.5 ms, Release 80 ms, 2–4 dB GR
  • Sub: Utility Bass Mono
  • Mid-bass: Compressor sidechain from Kick, EQ Band Pass 2 kHz, Q 0.7, Ratio 3:1, Attack 1 ms, Release 40 ms, 1–2 dB GR
  • Mid-bass: EQ Eight HP 110 Hz
  • Listen each Compressor's sidechain Listen button
You should hear
The sub ducks under the kick's body and the mid-bass ducks briefly under its click, each only where they clash.
Live edition
Standard and up

What the forum advised. In Kick & Bass Frequencies...? (2008), quoted again in Random bass tips, shonky bounced the kick twice. A copy low-passed at 100–140 Hz keyed the sub's ducker. A copy high-passed at 300 Hz–1 kHz keyed the mid-bass, so the beater's click cleared a space in the mids. The point was to duck only the band that clashes. Dated: this needed bounced copies in 2008. Live 12's Compressor filters its own sidechain, so it takes two devices and no audio copies.

  1. On the sub: Compressor, open the sidechain panel, Audio From the kick track (Post FX). Turn on the sidechain EQ, Low Pass, about 120 Hz. Only the kick's body triggers the ducking.
  2. On the mid-bass: a second Compressor keyed by the same kick, sidechain EQ Band Pass around 1.5–3 kHz. Only the click triggers it, and the duck is short.
  3. Listen with the Compressor's sidechain Listen button to confirm what each one is hearing.
The same thread had plenty of no-sidechain voices (Sharmaji balanced kick and sub by volume; docwra said good samples need no cuts), and the site's own advice is to arrange the gaps first. See sidechain vs automation vs note gaps. Use this when two parts genuinely overlap in time.

No reverb on the lows: high-pass the kick before its reverb

Effects: EQ Eight, Hybrid Reverb · Level: essential · Genres: dubstep, deep · Confidence: consensus

Quick start

Load
EQ Eight, Hybrid Reverb, Drum Rack, Simpler
Set
  • Return A: EQ Eight Low Cut 48 dB at 200 Hz, High Cut 7 kHz, before the reverb
  • Hybrid Reverb Algorithm, Dark Hall or Room, Decay 1.2–2 s, Pre-delay 15–25 ms
  • Body chain Simpler subby kick, Decay 300 ms; EQ Eight HP 40 Hz; Send A off
  • Top chain Simpler knock kick; EQ Eight HP 300 Hz; Send A −12 dB; −8 dB
  • Claps return high-passed at 400–600 Hz
You should hear
A kick with space around its knock while the low end stays dry and tight.
Live edition
Suite

What the forum advised. miss_molinari (2008) said reverb and delay don't belong on low frequencies, and suggested high-passing the kick's reverb send at roughly 180–210 Hz. breakbait used a subby kick for the body and a second, higher kick, high-passed at 300 Hz, to carry the reverb. Breakage (attributed, 2009) said to high-pass reverb returns, especially on reverbed claps, so they stay snappy. In 2011, Phaeleh (attributed) said his reverb send stays off the kick and sub altogether.

  1. Return A: EQ Eight first (Low Cut 48 dB at 200 Hz, High Cut at 7 kHz), then Hybrid Reverb (Algorithm, Dark Hall or Room, Decay 1.2–2 s, Pre-delay 15–25 ms).
  2. Send the kick's top layer, not its body. In a Drum Rack, each chain has its own sends.
  3. For reverbed claps, follow Breakage's point: high-pass the return at 400–600 Hz so the clap stays a crack, not a wash.

808 against a sine sub: never both at once

Instrument: Simpler / Drum Rack · Level: intermediate · Genres: dubstep, halfstep · Confidence: consensus on not overlapping, sidechain disputed

Quick start

Load
Simpler, Tuner, Drum Rack
Set
  • Tuner on the 808 track; read the tail, not the click
  • Simpler Transpose until the 808 matches the root
  • Tail pad copy to a second pad; Start about 30–50 ms in, past the click; 5 ms fade-in; Decay to the gap length
  • Choke kick, tail pad and sub in one Choke group, or notes that never overlap
You should hear
808 and sub that take turns instead of beating against each other, with the tails filling the gaps.
Live edition
Intro and up

What the forum advised. In sub bass fighting 808 boom kick (2010), hasezwei pointed out that an 808 is mostly sub itself, so it shouldn't play over the sub line. Basic A said to tune the 808 to the key, because two sine tails at different pitches beat and cancel. Echoi left the kick alone and used 808 tails with their attack removed as sub filler between kicks, never overlapping. briskisgoodforu pitched a snappy kick down so its click sat lower, which let the sub roll off below 50 Hz without losing punch.

  1. Tune the 808: in Simpler, set Transpose so its pitch matches the root (Tuner device on the 808 track, read the tail, not the click).
  2. Make a "tail" pad: copy the 808 to a second Drum Rack pad and move Simpler's Start past the click (about 30–50 ms in). Add a 5 ms fade-in and set Decay to the gap length.
  3. Put the kick, the tail pad and the sub in one Choke group if they live in one rack. Otherwise draw the notes so they never overlap.

Octave copies and tempo-timed compression for a kick that cuts

Effects: Shifter, Compressor · Level: intermediate · Genres: dubstep · Confidence: single poster

Quick start

Load
Drum Rack, Simpler, EQ Eight, Compressor
Set
  • Drum Rack pad three Simpler chains with the same kick: Transpose 0, +12 and −7
  • +12 chain EQ Eight HP about 1 kHz
  • −7 chain EQ Eight LP 120 Hz
  • Compressor (on the pad) Attack around 50 ms, Release around 400 ms
  • Ratio 4:1, 3–4 dB GR
You should hear
A kick with extra slap on top and boom underneath, recovering in time for the next beat.
Live edition
Standard and up

What the forum advised. wub's How to make a kick stand out (2008) had two tips. First, duplicate the kick, pitch the copy up an octave for slap and the original down an octave or a fifth for weight, and blend them. Second, compress with an attack of about a 1/32 note and a release of about a 1/4 note, timed to the tempo. His post says 13 semitones for an octave; it is 12. macc and others replied that the ear, not the formula, decides.

  1. On a Drum Rack pad, use three chains with the same kick in Simpler: Transpose 0, +12 and −7 (or −12). Pitching by playback speed makes the up-copy shorter and the down-copy longer, which is the slap-and-boom wub described.
  2. High-pass the +12 chain at about 1 kHz and low-pass the −7 chain at 120 Hz, so each adds only its band.
  3. Compressor on the pad: Attack around 50 ms lets the transient through, Release around 400 ms recovers by the next beat. At 140 a 1/32 is 54 ms and a 1/4 is 429 ms. Ratio 4:1, 3–4 dB of gain reduction.

See also Kick on this site.

Wobble and reese: the forum's twists

Two layers, opposite LFO phase: the wobble that fills its own gaps

Instrument Rack, Auto Filter · Level: intermediate · Genres: dubstep, 2008-era wobble · Confidence: consensus on basics, sub wobble disputed

Quick start

Load
Instrument Rack, Operator, Auto Filter, EQ Eight
Set
  • Chain A Operator saw-ish FM; Auto Filter LP 24, LFO 1/8, Phase 0°; −3 dB
  • Chain B same patch; Auto Filter LP 24, LFO 1/8, Offset 180°; −3 dB
  • Sub chain Operator sine −12 st; EQ Eight LP 100 Hz; 0 dB
  • Rate macro both LFO rates, 1/16 → 1/2
  • Depth macro both LFO amounts, 0 → 80%
You should hear
A wobble where one layer opens as the other closes, filling its own gaps over a steady sub.
Live edition
Suite

What the forum advised. rendr's WOBBLE BASS GUIDE (2008) added a few ideas to the usual LFO-on-a-low-pass. Clone the layer and run its LFO in the opposite phase, so one layer opens as the other closes. Split low, mid and high bands with different LFO rates. Let a second LFO modulate the first one's rate. When you resample, trim off the half-length first cycle so the rest stays on the grid. For a sub wobble he used a sine with an LFO on amplitude; chewie preferred a triangle with an LFO on cutoff. deadly_habit argued that tremolo and pitch LFOs aren't wobble at all. The basics are on the site's LFO wobble recipe.

  1. Instrument Rack with two identical chains. In chain B's Auto Filter, set the LFO Offset (phase) to 180°.
  2. Map both LFO rates to one macro. The rate-modulating-the-rate trick is an LFO (M4L) mapped to that macro at a slow 2-bar rate.
  3. Keep the sub as a separate, unmodulated chain. Both the forum's camps end up there anyway, and a big rig rewards a steady sub.

The reese detune argument: one steady oscillator vs equal and opposite

Instrument: Analog / Wavetable · Level: intermediate · Genres: dubstep, halfstep, dnb · Confidence: disputed

Quick start

Load
Wavetable, Utility
Set
  • Anchor chain Wavetable Osc 1 saw, Detune 0; 0 dB
  • Beating pair chain Wavetable Osc 1 saw +20 ct, Osc 2 saw −20 ct; −4 dB
  • School macro Anchor chain −inf (equal and opposite) → 0 dB (anchored)
You should hear
One knob moves the reese between a symmetric beating pair and one with a steadier, in-tune core.
Live edition
Suite

What the forum advised. moodswing's reese chapter (2009) called for two to four saws with all but one detuned 25–50 cents, a low-pass at 150–200 Hz with high resonance, tube or tape drive, and mono legato. ruckspin's reply was to detune two oscillators by equal and opposite amounts at equal level, which keeps the reese in tune. moodswing objected that this loses the steady bass and suggested keeping the undetuned oscillator louder. Brisance would drop the low-pass and use notches in the range the reese sweeps through. contakt321 built his in Operator with FM.

The full builds are on Wavetable unison reese and notch sweeps. See also Attack's Reese Bass with Virtual CZ.

Drums: half-step, 2-step and swing

The "step": a small hit one 1/16 before the kick or snare

Drum Rack · Level: essential · Genres: dubstep, 2-step-influenced 140 · Confidence: single poster (compiled chapter)

Quick start

Load
Drum Rack, Velocity
Set
  • Step pad a second, smaller snare, or the same one high-passed at 400 Hz
  • Placement one 1/16 before the main kick or snare
  • Velocity step notes at 35–50
  • Velocity MIDI effect Random 10–15 on that pad's chain
  • Variation duplicate the 4-bar clip, start marker one bar in
You should hear
A bouncier groove with a quiet stepping hit leading into each kick or snare, never quite the same twice.
Live edition
Intro and up

What the forum advised. Disco Nutter's drums chapter (2008) said a quieter snare or hat one 1/16 before the main kick or snare gives the "step" feel, and that lower-velocity kicks and hats give bounce. For instant variation he suggested shifting a four-bar loop by one bar. hugh, in the same thread, suggested automating velocity and reverb so the track feels held, then released.

  1. On a Drum Rack, put a second, smaller snare (or the same one high-passed at 400 Hz) on its own pad so its level and reverb send are separate.
  2. Draw the step notes at velocity 35–50. Add the Velocity MIDI effect with Random 10–15 on that pad's chain so no two are identical.
  3. For the bar-shift variation, duplicate the 4-bar clip, then use the clip's Start marker one bar in (or rotate the notes with the Recombine transformation's Rotate mode in Live 12).

See also Deep dubstep drums and Attack's Mystik Dubstep Beat Dissected (2012: 140 BPM, half-time snare, a dub echo on one snare hit).

Half-step against 2-step placements, and keeping the kick mixable

Drum Rack · Level: intermediate · Genres: dubstep, halfstep, 2-step · Confidence: disputed

Quick start

Load
Drum Rack
Set
  • Three clips bare, second-kick, 2-step; switch per 8 bars
  • Intro and outro first 16–32 bars bare or second-kick, kick on the grid
  • Middle 32 triplet kicks and 2-step shuffles
  • Switch-up one kick on the 1/8T grid, then mute every other lane until the next snare
You should hear
Pattern changes that keep the middle moving while the sections a DJ mixes on stay straight.
Live edition
Intro and up

What the forum advised. kick on 1, snare on 3 is dead (2010) is partly an argument about counting. Some posters count half-time beats, others count eighths, so "snare on 3" and "snare on 4" can mean the same thing. 3za said early dubstep needed a second kick before the snare, and that leaving it out now and then sounds enormous. paravrais listed garage-style placements. frogman said one kick and one snare per bar is boring. A year earlier, in Bad idea to put your main kick on triplets? (2009), duskky gave a compromise: keep the kick straight in the sections a DJ mixes on, and save the triplet skips for the middle. step correct's switch-up: put one kick on a triplet, then strip everything out until the snare.

  1. Keep three clips in the Drum Rack track (bare, second-kick, 2-step) and switch per 8 bars. Session view makes the comparison easy.
  2. duskky's rule as an arrangement habit: the first 16–32 bars and the outro use the bare or second-kick pattern with the kick on the grid. Triplet kicks and 2-step shuffles go in the middle 32.
  3. For step correct's switch-up, draw one kick on the 1/8T grid (Ctrl/Cmd+3 toggles triplet grid), then mute every other lane until the next snare.

See also Attack's Create Syncopated 2-Step Grooves Like A Pirate and the site's UK garage page.

Swing the hats, keep the kick and snare on the grid

Groove Pool · Level: essential · Genres: dubstep, deep, garage-leaning 140 · Confidence: consensus on hats, disputed on snares

Quick start

Load
Drum Rack, Simpler
Set
  • Hats and shakers in their own clip on their own track; kick and snare straight
  • Groove Pool Swing 16ths or MPC 16 Swing on the hat clip only
  • Groove amounts Timing 60–80%, Random 0–5%, Velocity 10–20%
  • No-groove route Simpler Start 1–5 ms later on the off hats, shorter decay
You should hear
Hats that lean and roll while the kick and snare stay locked to the grid.
Live edition
Intro and up

What the forum advised. In Dubstep Swing/Groove? (2008), habitualbeatscamp and r swung only the hats and rolling percussion and kept kick, snare and clap quantised. shonky built a Logic groove template from a drum loop and applied it to the other parts. lilt set synth LFOs to 1/12 or 1/6 for groove. In How much swing... (2008), one poster swung one or two of the four snares in a loop. daft cunt liked snare swing for listening but not for DJ mixing, and kwality kept snares tight. kwality added that swing also comes from sample start points and attack/decay. In his 2011 Q&A, Phaeleh (attributed) said he uses swing of about 33% or 66% depending on how garage the tune is, and always varies hat velocity.

  1. Split the hats and shakers into their own Drum Rack pads, or their own clip. Grooves apply per clip, so put them in a separate clip on a separate track if the kick and snare must stay straight.
  2. Groove Pool: drag in one of the Swing 16ths or MPC 16 Swing grooves from the Core Library and apply it to the hat clip only. Start with Timing 60–80%, Random 0–5%, Velocity 10–20%.
  3. shonky's template method in Live: right-click a break or live drum loop → Extract Groove(s), then apply that groove to the hat clip.
  4. Phaeleh's 33% / 66% are Cubase swing values. In Live, A/B a light and a heavy swing groove and pick the one that matches the tune's garage lean.
kwality's other route needs no groove at all: nudge the hat samples' start points (Simpler Start 1–5 ms later on the off hats) and shorten their decay, and the pattern sits back without any note moving.

Starting from a break and rearranging it into half-time

Simpler Slice / Drum Rack · Level: intermediate · Genres: dubstep, halfstep, breaky 140 · Confidence: attributed Q&A answers

Quick start

Load
Simpler, Drum Buss, EQ Eight, Glue Compressor
Set
  • Simpler Slice, by Transient, Sensitivity 60%, on a 1–2 bar break
  • MIDI main kick slice on beat 1, main snare slice on beat 3, ghosts and hats in order between
  • Drum Buss Drive 10%, Crunch, Boom off
  • EQ Eight and Glue HP 35 Hz; Glue Compressor 2 dB GR
  • Every 8 bars add or swap one percussion pad, never more
You should hear
The break's own swing and ghost notes, rearranged into a half-time feel that keeps evolving.
Live edition
Standard and up

What the forum advised. Breakage (attributed, 2009) said he usually starts by loading a break and rearranging its hits into a half-time feel, then spends hours swapping sections. He layers one-shots on top even when he starts from a kit, and keeps minimal beats moving with small additions: new hats, shakers, congas, a ride. Distance (attributed, 2009) chopped live drum loops and breaks and rebuilt loops from the slices. Dated: they used Recycle and Logic's EXS24. Live 12's Simpler Slice mode does the job.

  1. Drop the break onto a MIDI track to get Simpler in Slice mode (by Transient). Right-click → Slice to Drum Rack if you want per-slice processing.
  2. Rewrite the MIDI so the main kick slice lands on beat 1 and the main snare slice on beat 3. Keep the ghost and hat slices in their original order between them; that keeps the break's swing.
  3. Layer a clean one-shot kick and snare on separate pads, playing only the backbones. Breakage's "subtle changes": every 8 bars, add or swap one percussion pad (shaker, conga, ride), never more than one.

Snare, kick and percussion design

The three-layer snare: reverbed clap, dry body, long noise

Drum Rack · Level: intermediate · Genres: dubstep, deep, dark 140 · Confidence: single poster; layering itself is consensus

Quick start

Load
Drum Rack, Simpler, EQ Eight, Hybrid Reverb
Set
  • Clap chain (D1) Simpler clap, Decay 250 ms; EQ Eight HP 600 Hz; Send A 0 dB
  • Body chain Simpler snare, Attack 6 ms, Transpose −3; EQ Eight HP 150 Hz, cut 300 Hz −3 dB; −4 dB
  • Noise chain Simpler white noise, Decay 450 ms; EQ Eight HP 3 kHz; −14 dB
  • Return: Hybrid Reverb Plate, 1.5–2.5 s, return high-passed at 500 Hz
  • Parallel Compressor Dry/Wet 30%, Ratio 8:1, Release 300 ms
You should hear
A weighty snare that can sit loud without a big peak, its attack from a reverbed clap.
Live edition
Suite

What the forum advised. In Weighty Snares? (2015), Eskimo described three layers. A clean clap with a long reverb tail supplies the attack. A dry snare underneath gives the body and needs no strong transient of its own. High-passed white noise with a long release fills it out. Because only one layer carries the transient, the snare can be loud without a big peak. Hashkey kept darker snares around 1–2 kHz, used clip distortion and parallel compression, and kept 200–500 Hz clear for the bass. NinjaEdit boosted 200–250 Hz instead, so the two disagree. jaydot transposed the sample down a few semitones. Older chapters added more: two oh one (2007) ran the snare into a plate or spring and then heavy compression with a long release. In A Good Snare Source? (2010), therapist bounced the layered snare and deleted the sources.

  1. Load three Simplers on one Drum Rack pad (Ctrl/Cmd-drag samples onto the same pad, or group chains in a nested rack).
  2. Clap chain: send it to the reverb return (Hybrid Reverb Plate, 1.5–2.5 s, return high-passed at 500 Hz). Body and noise stay dry.
  3. On the pad's output, add Saturator in Hard Curve or Roar with a little drive for Hashkey's clip, and a parallel Compressor (Dry/Wet 30%, Ratio 8:1, Release 300 ms) for two oh one's long-release squash.
  4. Once it works, Freeze and Flatten or Bounce to New Track (12.2), and keep the single sample, as therapist did.

See also Attack's Layering Claps and Snares and the site's Snare and clap.

Distant, cavernous kicks: wet bus, low-cut, dry turned down

Effects: Hybrid Reverb, Auto Pan, Echo · Level: intermediate · Genres: deep, Burial-influenced, ambient 140 · Confidence: single posters

Quick start

Load
EQ Eight, Hybrid Reverb, Auto Pan, Echo
Set
  • Kick fader down 8–12 dB, Send A pre-fader at 0 dB
  • Return A: EQ Eight HP 120 Hz
  • Hybrid Reverb Convolution large space, Decay 3 s, Pre-delay 30 ms
  • Auto Pan tremolo, 1/16, Amount 25%
  • Echo Character Noise 10%, Wobble 15%, LP 3 kHz
You should hear
A distant, cavernous kick that is mostly room, flickering slightly in its tail.
Live edition
Suite

What the forum advised. Burial Cavernous Kick Drums (2011) is a set of guesses, not a method from the artist. Treat it as technique only. Useful ideas from it: a pre-fader reverb bus with a low-cut, plus a slight tremolo or sidechain on the reverb bus (extremesociety); reverb with the dry signal turned right down (JBE); layering game or film effects in place of a drum kick, then adding slight reverb, tape drive and a filter so they stay far away (kiene).

  1. Turn the kick's own fader down 8–12 dB and set its Send A to pre-fader at 0 dB. Most of what you hear is now the room.
  2. Return A: EQ Eight low cut at 100–150 Hz, then Hybrid Reverb in Convolution mode with a big space. A short Auto Pan tremolo (phase 0°) on the tail gives it the flicker extremesociety described.
  3. For kiene's distance, finish with Echo's Character section (Noise and Wobble) and its low-pass at 2–4 kHz, or Roar's tape-style curve with the Tone turned down.

Skanks, chords and dub FX

The skank: a muted offbeat chord into tape echo and spring

Instrument: Electric / Operator / any chord source · Level: essential · Genres: deep dubstep, dub, roots 140 · Confidence: consensus on the chain, details vary

Quick start

Load
Electric, Auto Filter, Phaser-Flanger, Echo
Set
  • MIDI triad on the offbeats, 1/32 long, high velocity
  • Electric amp Decay 70–110 ms, Sustain 0
  • Auto Filter HP 300 Hz, Res 10%; Phaser-Flanger Phaser, Amount 25%, Rate 0.3 Hz
  • Return A: Echo 1/8 triplet, Feedback 60%, Reverb 30%, Wobble 20%
  • Send A modest, ridden up on phrase ends
You should hear
A muted offbeat strike that blooms into tape-like triplet echoes and spring, clear of the bass.
Live edition
Suite

What the forum advised. In Reggae skank question!? (2010), several people answered. FSTZ: chords on the upbeat into a space echo. nitz: a plain triad on the offbeats with a triplet delay and high feedback. nutrician: a Rhodes or guitar slightly detuned, with spring reverb, tape delay and a phaser. egoless: the guitarist's muted, damped strike, played hard, then delay, then reverb. Everyone used the echo as a send. Dated: the posts used RE-201 emulations and Dubstation. Live 12's Echo has a tape character and its own spring-style reverb.

  1. Electric (or Operator with a short FM tine): amp Decay 70–110 ms, Sustain 0. Short notes (1/32) give egoless's muted strike; velocity high.
  2. Auto Filter high-pass at about 300 Hz so the skank never touches the bass.
  3. Return A: Echo synced to 1/8 triplet (nitz) or 3/16, Feedback 50–70%, Reverb amount 20–35% for the spring feel, Wobble 15–25% for tape movement. Set the skank's Send A to a modest level and ride it up on phrase ends.

The site's version is skanking offbeat chords. See also Dub FX and delays.

Dub chords: minor 7ths and 9ths, effects above 300 Hz

Instrument: Analog / Drift / Operator · Level: intermediate · Genres: deep dubstep, dub techno-leaning 140 · Confidence: consensus

Quick start

Load
Drift, EQ Eight, Echo, Hybrid Reverb
Set
  • Drift saw plus square, low-pass at 1–2 kHz
  • Amp envelope attack 5–15 ms, Decay 400 ms, Sustain 30%
  • Return: EQ Eight high-pass 300 Hz first
  • Return: Echo Stereo, L/R 3/16 and 1/4, then Hybrid Reverb
  • Automation send, Echo filter and reverb size per 8 bars
You should hear
Short minor 7th and 9th chords whose delay and reverb do the sustaining, all above 300 Hz.
Live edition
Suite

What the forum advised. In Deep techy dub chords? (2008), the replies had a lot in common. kwality and fiziks suggested starting from minor 7 and minor 9 chords. fiziks said reverb and tape delay are essential and suggested heavily filtered field recordings for atmosphere. corpsey layered sine, saw and square, each with its own filter and reverb. osk said to take care reverbing the lows, to keep stereo delay above about 300 Hz, and to automate reverb, delay and cutoff over time. distal pointed out that minimal arrangements expose every detail, so choose samples carefully.

  1. Drift or Analog: two oscillators (saw plus square), low-pass at 1–2 kHz, a slow amp attack of 5–15 ms, Decay 400 ms, Sustain 30%.
  2. Return: EQ Eight first (high-pass 300 Hz), then Echo with Ping Pong-style stereo (Channel mode Stereo, L/R times 3/16 and 1/4), then Hybrid Reverb. Automate the send, the Echo filter and the reverb size per 8 bars.
  3. fiziks' field-recording bed: a long ambience on its own track, Auto Filter low-pass at 800 Hz, Utility −18 dB, sent to the same return.

See also Attack's Dub Techno Synth Chords and Chords and stabs.

The send-to-itself feedback loop on a Live return

Routing: Return tracks, Echo, Shifter · Level: intermediate · Genres: dub, deep dubstep, roots · Confidence: consensus (Live 8 era)

Quick start

Load
Echo, EQ Eight, Shifter, Saturator, Utility
Set
  • Return B right-click its own Send B, Enable Send; start at −inf, up to about −6 dB
  • Echo 3/16, Feedback 0%, LP 2.5 kHz
  • EQ Eight HP 250 Hz, LP 4 kHz
  • Shifter and Saturator Pitch −0.3 st; Soft Sine 3 dB
  • Utility Gain −1 dB, Limiter last while you learn it
You should hear
Dub repeats that darken and sink a little in pitch on each pass, building without running away.
Live edition
Suite

What the forum advised. In Dub (feedback) delay in ableton live (2009), hurlingdervish put the delay on a return and enabled that return's send to itself (right-click the send). Raising it makes the return feed itself, and the return's level controls how much builds up. q23 added sending one return into another, for example a delay on the snare's reverb only, and compared it to classic 1970s desk routing (his comparison, not a documented history). In root's dub like delay thingy (2009), Sharmaji said much of the old dub delay sound is output fed back in. He suggested a pitch shifter inside the loop with the send on a knob. r preferred changing the delay time, because a speed change sounds more natural than a pitch shifter. Dated: hurlingdervish thought Simple Delay had no real feedback, and baydestrian corrected him. Live 12's Delay and Echo both have feedback, a filter and Freeze. The return loop is still worth having, because it lets you put anything inside the loop.

  1. On Return B, right-click its own Send B knob → Enable Send. Leave it at −inf to start.
  2. Inside Return B, put the delay, then EQ Eight with high- and low-cuts. These darken each pass and stop the loop from running away in the lows.
  3. Add Shifter set a fraction of a semitone down (Sharmaji's pitch shifter in the loop): each repeat sinks a little. Add Saturator for tape-style compression of the build-up.
  4. Bring Return B's Send B up slowly. The loop gains with each pass, so keep a Limiter at the end of the return while you learn it.
  5. r's alternative needs no loop: in Live 12's Delay, set the mode to Repitch and automate the delay time. The repeats bend in pitch like a tape echo.
A self-feeding return can run up to 0 dBFS in a second. Keep Utility at −1 dB inside the loop and a Limiter last. Automate the send rather than leaving it high.

See also Attack's Creating Dub Delays With Standard Plugins, the site's dub delay throws and Delay design.

Five delay tricks from the 2011 delay thread

Effects: Delay, Echo, Grain Delay · Level: intermediate · Genres: dubstep, deep, 140 · Confidence: single posters each

Quick start

Load
Delay
Set
  • Delay Sync off, L 449 ms, R 663 ms (about 20 ms off the grid at 140)
  • Feedback 40%
  • Filter on, BP 1.2 kHz, width 3
  • Transient smear 3 ms, Feedback 0, Dry/Wet 30% on the snare
You should hear
Repeats that sit just behind the grid without phasing against the hits, filtered to one band.
Live edition
Intro and up

What the forum advised. In Share your delay techniques or secrets! (2011), each trick came from one poster. amphibian used filter delays automated to choose which band repeats, and grain delays for pads. O Captain put a straight delay on triplet melodies, or a triplet delay on straight ones. psychedelicatessen added about 3 ms of delay to a snare to change its transient. efence set delays about 20 ms off the grid so the repeats don't phase with the hits. Bresson bounced a long delay tail, reversed it and led it into the sound. djake chained four delays on four sends, each with its own time and filter, all automated.

TrickLive 12 setting
Off-grid repeats (efence)Delay, Sync off, 1/4 = 429 ms, so 449 ms is about 20 ms late
Cross-feel (O Captain)Delay synced to 1/8 on a triplet melody, or 1/8 triplet (143 ms) on a straight one
Transient smear (psychedelicatessen)Delay at 3 ms, Feedback 0, Dry/Wet 30% on the snare, or Drum Rack chain Delay in samples
Grain pads (amphibian)Grain Delay, Spray 40 ms, Pitch +12, Feedback 50%, as a send
Reverse lead-in (Bresson)See reverse reverb, using the delay return

Reverse reverb and reverse delay into a drop or a vocal

Workflow: Bounce to New Track, clip Reverse · Level: beginner · Genres: dubstep, deep · Confidence: consensus

Quick start

Load
Hybrid Reverb
Set
  • Duplicate the clip Sample → Rev
  • Hybrid Reverb 100% wet, Decay 3 s
  • Bounce to New Track then Rev again on the bounced clip
  • Place trimmed to half a bar before the hit, with a fade-in
You should hear
A swell that rises into the word or hit and ends exactly where it starts.
Live edition
Suite

What the forum advised. In Reversed Reverb & Reversed Delay (2009), hurlingdervish and deadly_habit described the classic method: reverse the sound, add reverb or delay, print it, reverse it again. The tail now swells into the word or hit. deadly_habit used it on vocal lead-ins and about half a bar before a drop. He also layered the reversed and forward versions with automated envelopes. Dated: in 2009 this meant exporting from one program and reversing in another. In Live 12.2 it is two clicks.

  1. Duplicate the clip, open Sample → Rev. Add Hybrid Reverb (or Echo) at 100% wet on that track.
  2. Right-click → Bounce to New Track. On the bounced clip press Rev again. The swell now ends where the sound starts.
  3. Trim it to half a bar before the hit and draw a fade-in. Delete the helper track.

One plate for the whole tune, returns high-passed hard

Effects: Hybrid Reverb, EQ Eight, Auto Pan · Level: essential · Genres: deep, atmospheric 140 · Confidence: attributed Q&A answer; agrees with forum consensus

Quick start

Load
Hybrid Reverb, EQ Eight, Utility
Set
  • Return A: Hybrid Reverb Algorithm Plate, Decay 2.2 s, Pre-delay 20 ms
  • EQ Eight (after) HP 300 Hz 48 dB, LP 8 kHz
  • Utility Width 100%
  • Sends kick and sub off; pads −6 dB; tight sounds and vocals −15 to −20 dB
You should hear
One shared plate giving the whole tune the same space, with kick and sub left dry.
Live edition
Suite

What the forum advised. In his Q&A (2011), Phaeleh (attributed) described his reverb setup. He uses one plate reverb send with the same settings on every tune, and a little of it on almost everything except kick and sub. Returns are high-passed hard (he put the cut somewhere around 100–500 Hz) with the top rolled off, and his template puts the EQ after the effect. He builds atmosphere from filtered delays and reverb-into-delay chains, puts an LFO on the pan of FX, and pans sounds that share a range away from each other instead of carving them with EQ. Older posts agree in spirit. daft cunt (2007): "Dubstep loves reverb." ThinKing (Bible p1) used two drum reverb buses, a room and a short dry one, at different send levels. Dam13n (2010) counselled restraint.

  1. Save a Live template with Return A (plate, EQ after) and Return B (filtered delay), so every tune starts with the same space.
  2. Turn the Kick and Sub sends off. Pads get Send A at −6 dB, tight sounds and vocals at −15 to −20 dB.
  3. For the FX pan movement, put Auto Pan on the FX return (Rate 2 bars, Amount 40%).

See also Attack's Adding Atmosphere With Delay And Reverb and the site's Reverb and space.

Resampling

Crush the highs in parallel, bounce, repeat

Audio Effect Rack, Redux, Roar · Level: intermediate · Genres: dubstep, mid-bass · Confidence: single poster

Quick start

Load
Audio Effect Rack, EQ Eight, Redux, Chorus-Ensemble, Roar
Set
  • Dry chain Utility 0 dB
  • Crushed highs chain EQ Eight HP 1.5 kHz 48 dB; Redux Downsample 4, Bit Depth 10
  • Then Chorus-Ensemble Chorus, Amount 30%; Roar Single, Drive 40%; EQ Eight HP 1.5 kHz, LP 9 kHz
  • Chain level −10 dB; Amount macro until the crush reads as texture
  • Bounce to New Track repeat with different settings, two or three generations
You should hear
A dirty bass whose high mids and treble carry crushed, chorused texture while its body stays clean.
Live edition
Suite

What the forum advised. dc_wolftown's resampling primer (2009) went like this. Make a dirty bass. Send it to a second channel with bit reduction, chorus, distortion and EQ, plus a crossover so only the high mids and treble are processed. Blend that with the dry channel, bounce, reload and process again; or split the sound into four bands. He suggested testing subtle downsampling early to see whether a sound takes degradation well. 86. preferred heavier crushing with racks of several crushers. Dated: CMT, Camel Crusher, FL's Edison. In Live 12 it's one rack and Bounce to New Track.

  1. Build the rack above on the bass track. Blend the Amount macro until the crush reads as texture, not noise.
  2. Bounce to New Track (12.2). On the new audio track, load the same rack again (or Drum Buss and Echo), set it differently, and bounce again. Two or three generations is the forum's typical depth.
  3. Chop the best bars into Simpler Slice mode and play them as new phrases. The full workflow is on Resampling workflow.

See also Attack's Six Bounce To New Track Tricks To Try In Live 12.2.

A drum loop as the reverb impulse, to stamp rhythm onto pads

Effects: Hybrid Reverb (Convolution) · Level: advanced · Genres: deep, experimental 140 · Confidence: single poster

Quick start

Load
Hybrid Reverb, EQ Eight
Set
  • IR a one-bar drum loop, no reverb, dragged onto the convolution IR slot
  • Hybrid Reverb Convolution only, Blend 100% conv, Dry/Wet 100%
  • EQ Eight HP 200 Hz
  • Print Bounce to New Track, then Simpler Slice to re-sequence
You should hear
A sustained pad that pulses in the drums' rhythm and tone.
Live edition
Suite

What the forum advised. In What's been your most unconventional production technique? (2008), shonky described using a convolution reverb at 100% wet with a drum loop as the impulse. That imposes the drum rhythm on pads and voices, and he recorded the results to reuse. In the same thread, thinking sent drum MIDI patterns to synths to get atonal loops locked to the drums, and gravious used the first few milliseconds of pitched-down film samples as hits.

  1. Export a one-bar drum loop (kick, snare and hats, no reverb). Drag it onto Hybrid Reverb's convolution IR slot. Hybrid Reverb accepts user audio as an impulse.
  2. Set Dry/Wet to 100% and the algorithm blend fully to convolution. A sustained pad now pulses in the drums' rhythm and tone.
  3. For thinking's trick, copy the drum MIDI clip onto an Operator track. The synth plays the drum rhythm as pitched blips that lock to the groove.

Mixdown and headroom

The "money shot" gain structure: drums at −9, sub two below

Utility, meters · Level: essential · Genres: all · Confidence: consensus on the method, disputed on playback loudness

Quick start

Load
Utility
Set
  • Every track Utility first at −6 dB; faders at 0 dB
  • Kick and snare peaks near −9 dBFS soloed
  • Pure sine sub about 2 dB under the drums (−11 dBFS)
  • Pads, chords around −18 dBFS; shakers, hats −20 dBFS or lower
  • Full mix −6 to −4 dBFS, no limiter on the mix bus
You should hear
An unlimited mix that peaks around −6 to −4 dBFS with drums as the reference and everything else in proportion.
Live edition
Intro and up

What the forum advised. THE MONEYSHOT THREAD (2009) ran to 62 pages. Its opening pages set out rough peak targets per channel played solo. macc, whose signature links a mastering business, gave: drums around −8 to −10 dBFS, a pure sine sub a dB or two lower, pads around −18, shakers lower. On page 2, Sharmaji and TeReKeTe (quoted) built mixes with kick and snare peaking around −9 that end at about −6 to −4. Depone (quoted) pulled every channel down 6–8 dB before loading anything and turned the speakers up instead. macc said to lower the source rather than the fader, to turn the reference down to your level rather than your mix up, to watch intersample peaks near 0, and that sidechain, multiband and extreme EQ are often plasters over bad levels. serox and james fox pushed back: unlimited mixes sound weak when you share them, so put a limiter on the copy you send to friends.

Element (soloed)Peak targetNotes
Kick, snare−9 dBFS (−8 to −10)The reference everything hangs on
Pure sine sub−11 dBFSAbout 2 dB under the drums; distorted bass judged by ear
Pads, chordsaround −18 dBFSThey sum; many quiet parts
Shakers, hats−20 dBFS or lowerQuieter than you think
Full mix−6 to −4 dBFSNo limiter on the mix bus
  1. Template: put a Utility first on every track at −6 dB (or set each instrument's output 6 dB lower). Faders start at 0 dB.
  2. Solo the kick, then set its Utility so the channel peaks near −9 dBFS. Do the same for the snare. Bring the sub in at about 2 dB below, judged on peaks and then by ear.
  3. Put the reference on a track routed straight to the Main, with Utility pulling it down until it sounds as loud as your mix.
  4. If you need to share a loud copy, put a Limiter on a separate bounce. Keep the mix itself unlimited for mastering or cutting.

The site's headroom maths is on Loudness and limiting. See also Attack's How Gain Staging In Your DAW Can Help Keep Your Mix Clean And Punchy.

Where mono starts: 100, 150 or 500 Hz?

Utility (Bass Mono), EQ Eight M/S · Level: essential · Genres: all sound-system music · Confidence: consensus that sub is mono, cut-off disputed

Quick start

Load
Utility, EQ Eight
Set
  • Sub and bass tracks Utility Width 0%
  • Main Utility Bass Mono on, 120 Hz (150 Hz if a stereo pad has wide low mids)
  • Dubplate check EQ Eight M/S, S, low cut 300 Hz; solo S and listen for bass
  • Mono test Utility Mono on the Main; nothing should disappear
You should hear
A low end that is identical in both channels and a mix that survives a mono rig without losing parts.
Live edition
Standard and up

What the forum advised. In Random bass tips (2008–09), the bass-should-be-mono answers came from three directions. Rendr: bass should be mono and not panned below about 500 Hz. Deadly Habit: the sub must be mono for vinyl. caeraphym: most club rigs are mono from about 30–120 Hz. The Bible's mastering lines and the dubplate thread (thinking) said up to 500 Hz; daft cunt said 100 Hz for vinyl and up to 300 Hz if unsure. One poster cited a mastering engineer who disputed "sub = mono" as an absolute rule. In production tips for club systems, Sharmaji said the mix must work in mono without being bounced mono. See also: in the VICE oral history, Loefah recalls a panned bassline that had to be cut mono.

  1. On the sub and bass tracks: Utility, Width 0% (true mono).
  2. On the Main: Utility Bass Mono on, Frequency 120 Hz. Raise it to 150 Hz if a stereo pad has wide low mids.
  3. For a dubplate, check the side channel. EQ Eight in M/S mode, set to S, with a low cut at 300 Hz; solo the S channel and listen for any bass.
  4. Test with Utility Mono on the Main: if a part disappears or drops a lot, its stereo is phase-based and won't survive a mono rig.

The site's stereo chapter: Mid/side and stereo.

High-pass everything but kick and bass, and split moving basses into bands

EQ Eight, Audio Effect Rack (Multiband split) · Level: essential · Genres: all · Confidence: consensus / single poster

Quick start

Load
EQ Eight, Audio Effect Rack, Roar, Redux, Chorus-Ensemble
Set
  • Default EQ Eight band 1 Low Cut 12 dB at 70 Hz on every pad, chord, hat, FX and vocal
  • Low chain (<120) EQ Eight LP 120 Hz 48 dB; Utility Mono
  • Mid chain (120–2k) EQ Eight BP 120 Hz–2 kHz; Roar Drive 30%; −3 dB
  • High chain (>2k) EQ Eight HP 2 kHz; Redux Downsample 3; Chorus-Ensemble Width 120%; −9 dB
  • Mute test at the end of every session
You should hear
Clean low end owned by kick and bass, and a moving mid-bass distorted and widened only where it can take it.
Live edition
Suite

What the forum advised. In Random bass tips, Jtransition said to high-pass everything except kick and bass, starting around 70 Hz. wang said to roll off rather than hard-cut, and to check every sound, hats included. Deadly Habit split heavily modulated basses into bands on separate buses, processed each and recombined them. In 2007, daft cunt (Random Production Tips) added a ruthless rule: if muting a track doesn't change the tune, delete it.

  1. Save EQ Eight with band 1 at Low Cut 12 dB, 70 Hz as a default preset. Drop it on every pad, chord, hat, FX and vocal.
  2. For a moving mid-bass, use the band-split rack. Distort only the mid band, widen only the high band, and keep the low band clean and mono.
  3. Do the mute test at the end of every session.

Sound-system translation

Judging the sub for a rig you can't hear at home

Monitoring · Level: essential · Genres: dubstep, deep, all sound-system music · Confidence: disputed by experience

Quick start

Load
Spectrum
Set
  • Spectrum on the Main, Block 8192, Average 1, Scale Log
  • Compare the sub band against a trusted reference at matched loudness
  • Systems monitors, headphones, a phone, then the corner of the room
  • Mixdown two hours, then listen fresh the next day
You should hear
A sub level judged against references rather than your monitors, adjusted by a dB or two after a real rig.
Live edition
Standard and up

What the forum advised. production tips for club systems (2008–2012) shows the problem cutting both ways. misk found that a sub which sounded modest on monitors was enormous in a club. Promise One heard his own sub come out too quiet on the Plastic People Funktion-One, and called sub level the hardest thing to get right. Promise One and bassmusic also said big rigs compress, so leads and vocals need to sit louder than you'd expect and snares and hats lower. Sharmaji wanted content down to about 26–30 Hz; daft cunt said cutting below 30 Hz mostly buys headroom. Other habits: Ramadanman (attributed, 2009) judged how the bass would drop by playing it on a big home subwoofer. cryptic (2007) listened from a corner or the hallway. Pinch (attributed, 2009) capped mixdown sessions at about two hours because ears tire and mixes drift bright, and referenced a trusted track first.

  1. Spectrum on the Main, Block 8192, Average 1, Scale Log. Compare the sub band's height against a trusted reference at matched loudness, not against your feeling.
  2. A/B on three systems in one sitting: monitors, headphones, a phone. Then the corner of the room (cryptic). If the sub is inaudible on the phone, use the filtered-square harmonic.
  3. Time-box the mixdown: two hours (Pinch), then stop and listen fresh the next day. Keep the monitoring level low between checks.
  4. If you get to hear it on a real rig, take notes and adjust by a dB or two only. Promise One and misk got opposite surprises, so one night is a data point, not a rule.

The site's full treatment: Sound system mixing.

Mastering and dubplates

Preparing a mix for a cutting engineer

Export, EQ Eight, Utility · Level: essential · Genres: dubstep, deep, any vinyl release · Confidence: consensus (2007 and 2011 threads agree)

Quick start

Load
Utility, EQ Eight
Set
  • Main bypass everything except metering; no Limiter, no Multiband Dynamics
  • Level if peaks pass −3 dBFS, pull down with a Utility on the Main
  • Checks Utility Mono; EQ Eight M/S, solo S, low-cut 300 Hz: no bass in the side
  • Export 24-bit WAV, Dither off, Normalize off
You should hear
A dynamic, unsquashed mix with mono bass that a cutting engineer can work with.
Live edition
Standard and up

What the forum advised. The 2007 thread Preparing a track for dubplate was later copied into the Production Bible. jtransition (a handle; his signature doesn't say who he is) wrote from a cutter's point of view. Peak about −6 dBFS for 24-bit files (−2 to −3 for 16-bit). Avoid sibilance, out-of-phase material, flat-topped mixes, heavy limiting, excessive highs, excessive sub-harmonics and narrow boosts over about 2 dB, and don't use multiband compression. He gave side lengths of about 6–7 minutes for a 10" and about 10 for a 12"; louder cuts take more groove. breaksbaron: one track per side under about 7 minutes cuts loudest. thinking and jamminb wanted mono bass and a cut around 30–40 Hz, which cutting houses often do themselves. Several posters, Ramadanman among them, said to skip mastering plugins and ask the cutting house what it wants. In 2011, safeandsound, whose signature names a mastering studio, gave the same checklist for professional mastering: take the master limiter off, look for squared-off peaks, pull the master down if it clips, send 24-bit WAV or AIFF at the project rate, undithered, with 3–6 dB of headroom. Bounce level was disputed. Ramadanman (attributed) bounced with about 0.3 dB to spare and left the dynamics to the engineer. That only works with nothing squashing the mix.

FormatComfortable length per side (2007 thread)Trade-off
10" dubplateabout 6–7 minLonger means quieter
12" (one track)about 10 minOne track under about 7 min cuts loudest
  1. Bypass every device on the Main except metering. If peaks pass −3 dBFS, pull the mix down with a Utility on the Main (safeandsound's compromise) rather than adding a limiter.
  2. Run a check pass: Utility Mono on the Main, and listen for vanishing parts. Then EQ Eight in M/S, solo S, low-cut 300 Hz: there should be no bass in the side channel. Check for harsh "s" sounds around 6–9 kHz with a narrow sweep and fix them with a de-esser (Multiband Dynamics on the top band) on the vocal track, not on the master.
  3. Export 24-bit WAV, Dither off, Normalize off. Write the side length on the order. Ask the cutting house about their level and format preferences before you send.
  4. Pinch (attributed) recut tracks up to three times after playing them out; Cyrus (attributed) treated cutting dubs as a way to learn. Budget for a recut.
From the other side of the lathe: in the VICE oral history, the Transition cutting engineer Jason Goz describes keeping levels down so the music keeps its dynamics. That matches the forum's advice; it does not tell us who any forum account was.

Arrangement: DJ-friendly structure

16-bar blocks, an intro the DJ can mix on, a drop that lands in time

Arrangement · Level: essential · Genres: dubstep, deep · Confidence: consensus

What the forum advised. In Song structures found in Dubstep (2008), serox answered as a DJ: the first sound and the first drop have to be in time, and an off-time breakdown is the worst thing a record can do. kidlogic: odd-length intros make mixing hard. jobbanaught: 16, 32 or 48-bar intros, never starting mid-bar. manray: a beat or marker at the very start, and multiples of 8. On page 2, Sharmaji (whose signature lists his labels) gave a formula: a 32-bar intro with no sub, a 1-bar break with an FX or sample, then the drop. inertia_dubz gave ranges, and setspeed said a 64-bar drop section is common. The Bible's compiled template ran 16/32 intro, 64 main, 16 break, 48 second main, 16 outro. Breakage (attributed) kept his intros under 32 bars. kidlogic's trick: an odd-length breakdown before the second drop keeps the DJ playing past it, since DJs often mix out once the phrasing returns to normal. See also: in the VICE oral history, Artwork recalls early tracks built as loops with a 16-bar intro before the drop.

  1. Set Live's loop brace to 16 bars and build in those blocks. Use Locators (Create → Add Locator) at 33, 97, 113 and 161 as the section map.
  2. Intro: the first beat on bar 1 beat 1 with a kick or a clear transient (manray's marker). No sub for the first 32 bars (Sharmaji), so the outgoing tune's bass owns the room during the blend.
  3. Bar 32: one bar of space, with only a sample or FX throw. The drop lands on bar 33 beat 1.
  4. Outro: 16 bars of drums mirroring the intro, so the next DJ can mix out of you.

See also the site's Deep dubstep track, whose arrangement is built on the same DJ-friendly intro and outro.

Keeping a long, sparse track moving

Arrangement, automation · Level: intermediate · Genres: deep, minimal 140 · Confidence: consensus on the principle

Quick start

Load
Auto Filter, Echo
Set
  • Plan one change per 8 bars: add, remove, or a processing move; never two at once
  • Ramps 8–16-bar automation: Auto Filter frequency, a send level, Echo feedback
  • Rework Bounce to New Track, slice the mix (Cmd/Ctrl+E), rearrange, copy the moves back
You should hear
A long, sparse track that keeps evolving without blocky jumps.
Live edition
Suite

What the forum advised. In Techniques for Keeping Arrangements Interesting (2010, replies to 2014), didi said to change something every 8 bars. Another poster, quoted there, said every 2. drdeft used long fades over 8–16 bars. erratech said to tease parts in so the arrangement isn't blocky, and to make barely noticed changes to processing or phrasing. mks put it as tension and release: decide when to push an element forward and when to strip it back. jameshk bounced the whole track, chopped and rearranged the segments, then applied the best moves back to the project. Breakage (attributed) and Ramadanman (attributed) both start from a beat that develops. Breakage keeps a minimal beat alive with one small percussion change at a time.

  1. Plan one change per 8 bars on paper: an added part, a removed part, or a processing move (a filter, a send). Never two at once.
  2. Use 8–16-bar automation ramps rather than step changes: Auto Filter frequency, a send level, Echo feedback.
  3. jameshk's trick in Live 12.2: Bounce to New Track for the whole mix, slice it in Arrangement (Cmd/Ctrl+E), rearrange, then copy the moves that work back onto the real tracks.

Dated advice and the Live 12 equivalent

Forum-era tool or claimLive 12 equivalent
FL 3xOsc or Subtractor "filter fully closed" for a sineOperator Osc A sine, or Wavetable with the Basics sine
Recycle, Phatmatik, EXS24 for breaksSimpler Slice mode, Slice to Drum Rack
Bouncing kick copies to key sidechainsCompressor's sidechain EQ (low-pass or band-pass)
Live 7–8 Simple Delay "has no feedback" (it did)Delay (Repitch mode, Freeze, filter) and Echo (tape character, reverb, modulation)
RE-201 and Dubstation plug-ins for skanksEcho with Wobble, Noise and Reverb, as a return
CMT bitcrusher, Camel CrusherRedux, Roar, Drum Buss
Logic groove templatesGroove Pool, Extract Groove(s)
Export, reverse in another app, re-importClip Rev plus Bounce to New Track (12.2)
Hand-made single-cycle samplesSimpler loop, or a user wavetable in Wavetable
"−6 dBFS peaks for 24-bit" with no true-peak meterSame target; Limiter's True Peak option and the Main meter's peak readout catch intersample overs

Common mistakes the forum kept correcting

Related