ABLETON 12 SOUND DESIGN ATLAS

Layering Rules

How to stack sounds in Live 12 without mud or cancellation — frequency band ownership, sub/mid/top splits, polarity and Align Delay checks, mono vs stereo per layer, envelope agreement, tuning, Drum Rack and Instrument Rack zone layering, and sample-plus-synth hybrids.

layeringlayersfrequencybandphasepolarityalign delayutilitymonostereoenvelopetuningdrum rackinstrument rackkey zonevelocity zonesamplesynthspectrum

Layering is adding a second sound so the sum does something one sound could not: a sub that holds the floor plus a mid that carries the note, a click that gives a kick its position plus a body that gives it weight. It fails in exactly four ways: two layers claim the same band (mud), two layers in the same band are out of phase (cancellation), two layers have different envelopes (smear) or two layers are not in tune (beating). Everything on this page is a procedure to catch one of those four before it reaches the mix. Live 12's Drum Rack chains, Instrument Rack key and velocity zones, Utility, Align Delay and Spectrum are the whole toolkit.

Concept

A layer should own a region of the spectrum and a region of time. Ownership means that in that band, that layer is at least 6 dB louder than any other layer, and that outside it the layer is filtered down so it stops competing. Two sounds that both have energy at 150 Hz and both peak at 3 ms after the note will sum unpredictably: at best +6 dB, at worst −∞, depending on their relative phase at 150 Hz, which depends on their start offset, their filter settings and the polarity of the sample. The fix is never "turn one down". It is: decide who owns 150 Hz, high-pass the other above it, then check polarity in the crossover region.

Time ownership works the same way: the click owns 0–15 ms, the body owns 15–300 ms, the tail owns the rest. When envelopes agree (every layer decays in the same order and nobody is still ringing when the next hit arrives) the ear fuses the layers into one instrument. When they disagree you hear two instruments, which is sometimes what you want (a pad under a pluck) but never what you want in a kick. All numbers below are starting points; BPM 128 where it matters.

Techniques

Frequency band ownership: the sub/mid/top split

Device: EQ Eight, Utility, Audio Effect Rack · Level: essential · Genres: all

Split every layered sound into three roles and give each role one layer. Sub (30–90 Hz) is one mono oscillator. Mid (90–600 Hz) is the layer with the character. Top (above 2 kHz) is noise, click or a bright synth. Filter each layer so it stops at its border with a 24 dB/oct slope. The borders are where the next two techniques happen.

  1. Make an Instrument Rack with one chain per role. On each chain, EQ Eight with 24 dB/oct cuts at the borders. Sub: LP ~100–120 Hz. Mid: HP ~80–90 Hz, LP ~2–3 kHz. Top: HP ~2 kHz.
  2. Set the overlaps deliberately: 10–30 Hz wide at the sub/mid border, up to half an octave at the mid/top border. Too narrow a crossover sounds like three instruments; too wide reintroduces the phase problem.
  3. Solo each chain and sweep the rack's note range; confirm each chain sounds incomplete on its own. If one chain sounds like a finished bass, it has too much band.
  4. Level: sub 0 dB, mid −3 to −6, top −10 to −15 as starting points. Check on a Spectrum (Max hold on) that the summed slope falls roughly 6–9 dB per octave above 100 Hz.
A loudspeaker and a limiter both respond to the sum. Two layers sharing 100 Hz do not double the bass, they double the peak at 100 Hz, which costs headroom. One owner per band gives maximum level per dB of peak.

Polarity check with Utility phase invert

Device: Utility, Spectrum · Level: essential · Genres: all

Before any other adjustment, flip the polarity of one layer and listen to the crossover band. Whichever setting is louder and fuller in that band is correct. Utility's Ø L / Ø R buttons do this per channel; use both for a mono-compatible flip.

  1. Loop a single hit or a held note with both layers playing. Put Spectrum on the group or on the master, Max hold on, and watch the overlap region (for kick/sub: 50–120 Hz; for mid/top: 1–3 kHz).
  2. Toggle Ø L and Ø R together on the second layer. One state gives a higher trace in the overlap; keep that one.
  3. If both states look the same, the layers are ~90° apart: neither cancels nor adds. Move to Align Delay (next technique) to rotate them into alignment.
  4. Flip both channels or neither. Flipping only L on a stereo layer makes it wide and hollow in mono, a different tool (see Mid/side & stereo).
Polarity inversion is a 180° phase shift at every frequency at once, zero latency. For two sources with similar waveforms (two kick samples, a sub and a body from the same synth) it is the single most likely fix. It cannot fix a frequency-dependent offset, which is what Align Delay and EQ phase are for.
Samples from commercial packs come in both polarities; a kick that "pushes" the speaker cone out has positive polarity. Mixing a positive kick with a negative 808 is the most common source of a thin low end that no EQ can fix.

Align Delay for sub-millisecond offset between layers

Device: Align Delay · Level: intermediate · Genres: techno, dubstep, DnB, trap

When two layers start at the same sample but their fundamentals are not in phase (different synths, different filter slopes, a sample with silence at the start), polarity alone leaves them partly cancelled. Align Delay shifts one layer by samples, milliseconds or metres so you can rotate the phase at the crossover frequency into alignment.

CrossoverHalf-cycleSweep rangeMode
60 Hz (kick/sub)8.3 ms0–8 msms
120 Hz (sub/mid bass)4.2 ms0–4 msms
500 Hz (body/presence)1.0 ms0–1 mssamples (0–48)
2 kHz (click/body)0.25 ms0–12 samplessamples
  1. Put Align Delay on the layer without the transient (usually the sub). Delaying the transient layer moves the hit off the grid.
  2. Loop the hit. Sweep the delay across one half-cycle of the crossover frequency while watching the peak meter on the group. The highest peak is phase alignment; the lowest is full cancellation.
  3. Combine with polarity: if the best position is near the end of the sweep, flip Ø and sweep again from 0; the correct alignment is usually within a quarter cycle.
  4. Write the final value into the track name or a macro. Never leave Align Delay mapped to a live macro: it changes timing.
A delay of t seconds is a phase shift of 360·f·t degrees at frequency f. It aligns one frequency exactly and nearby frequencies approximately, which is why you align at the crossover and not at the fundamental of each layer. Sub-millisecond precision matters above 500 Hz; the samples mode exists for that.

Mono vs stereo per layer

Device: Utility, Chorus-Ensemble, Wavetable Unison · Level: essential · Genres: all

Width is a per-layer decision, not a per-sound decision. The rule: sub mono always; mid narrow (Width 40–70%) or mono below a crossover; top as wide as you like. A wide sub cancels in mono and wastes vinyl/club system headroom; a mono top sounds small. Utility's Bass Mono with its frequency control makes the rule automatic on any layer.

LayerWidthSource of widthCheck
Sub (<90 Hz)Mono (Utility Mono or Width 0)NoneCorrelation meter at +1
Mid bass / body (90–600 Hz)Mono below 120–180 Hz, 40–70% aboveUnison, detune, Chorus-EnsembleMono-sum: level drop under 1 dB
Presence (600 Hz–3 kHz)60–100%Short delays, Haas under 15 ms, Phaser-FlangerMono-sum: no hollowness
Top (>3 kHz)100–130%Wavetable Unison Shimmer/Noise, Wide Noise in Erosion, Utility WidthOnly that it does not vanish in mono
Audition the whole stack through Utility Mono on the master (keep one on the master chain off by default; see Mix templates). If the mono version loses more than 1.5 dB of the bass note, a mid layer is too wide or the Bass Mono frequency is too low.
Stereo width on a sub means L and R differ; on a mono club system they add, and whatever differs cancels. Width above 2 kHz is almost free because the wavelengths are short and the room scatters them anyway.

Envelope agreement across layers

Device: Simpler, Operator, Drift, Envelope Follower · Level: essential · Genres: all

Layers fuse when their amplitude envelopes line up: same onset time, decays in a sensible order (top shortest, sub longest, or all equal for a tight stab), and no layer still sounding when the next hit lands. The practical rule: the sub decays last and the click decays first, and every layer's attack is within 2 ms of the others unless you want a deliberate pre-echo.

  1. Match onsets. Samples: trim leading silence in the clip, or use Simpler's Start point; synths: attack 0–2 ms on every layer of a drum, identical attack on every layer of a chord.
  2. Order the decays: click < top < body < sub for drums. For chords/stabs, identical decay on all layers; otherwise the chord's timbre changes as it fades.
  3. Cap the longest decay at the hit interval: at 128 BPM a 4-on-the-floor kick repeats every 469 ms, so sub decay under ~350 ms keeps it clean.
  4. For sustained sounds (pads), agree on release instead: all layers 200–600 ms, or the pad "unlayers" on every note-off.
  5. Use one Envelope Follower (M4L) from the body layer to drive the top layer's Utility gain if the top is a loop or a noise bed that cannot be enveloped at the source.
The ear groups sounds that start together (common onset) and that share a modulation pattern. Matching envelopes exploits that grouping; mismatching them deliberately (a pad swelling under a pluck) is how you make two instruments from one MIDI part.

Tuning layers: fundamentals, harmonics and inharmonic samples

Device: Tuner, Simpler, Operator, Utility · Level: intermediate · Genres: techno, trap, dubstep, DnB

Two layers at 55 Hz and 57 Hz beat at 2 Hz: a slow wobble in the sub that sounds like a bad mix. Tune every pitched layer to the same note, then decide whether the harmonic content agrees. A kick sample has a pitch (usually its tail) and an 808 is a pitched sine; a snare has a resonant mode around 180–250 Hz; a noise layer has a band centre. All of them can be tuned to the key of the track.

  1. Put Tuner after each pitched layer, loop the hit, read the note of the sustained part (the first 20 ms are inharmonic, ignore them).
  2. Transpose in Simpler (semitones + Detune in cents) or Operator (Coarse/Fine) until all layers show the same note. For drums, tune to the root or the fifth of the track's key; for basses, obviously the note.
  3. Kick and sub: tune the kick tail to the same note as the sub's lowest note, or an octave above. A kick at F and a sub at E will beat on every bar in E.
  4. Inharmonic layers (metallic hits, Collision, Corpus): tune the loudest partial, not the perceived pitch, and keep them in the presence band where beating is less audible.
  5. With Live 12 tuning systems: if the set uses a non-12-TET scale, synth layers follow it automatically; sample layers in Simpler only follow if the tuning system is applied to the track. Check that layered samples are not sitting 10–20 cents off the synth layers.
Two tones within about 15 Hz of each other produce audible beating; above that they are heard as a rough interval. In the sub region, every interval is inside that window, so the only acceptable relationships are unison and octave.
A 1–3 cent detune between a sub and a mid layer gives slow movement without beating; keep it under 5 cents and keep it on the mid layer, never on the sub.

Drum Rack layering via chains on one pad

Device: Drum Rack · Level: essential · Genres: all

A Drum Rack pad can hold any number of chains on the same note. Each chain gets its own devices, so per-layer EQ, Align Delay and Utility live inside the pad, and the pad itself has one volume, one Choke group and one output. This is the cleanest way to build layered drums that stay portable and performable from Push.

  1. Drop the first sample on the pad. Open the pad's chain list (show chains), then drag the second sample onto the same pad holding Cmd/Ctrl so it adds a chain instead of replacing.
  2. Per chain: EQ Eight for band ownership, Align Delay if the layer has no transient, Utility for Ø and width. Velocity zones per chain if a layer should only appear on hard hits.
  3. Set the pad's Choke group so all chains stop together; leave the chain Choke empty.
  4. Add rack macros for the layer balance; on Push these show up on the pad's macro page. Use Macro Variations to store a "soft" and a "hard" version of the layer balance.
  5. Bounce in place (12.2) or Freeze/Flatten when the layer is final, and keep the rack as a preset for the next track.
Drum Sampler (12.1) already includes Sub, Punch and Noise layers inside one device. For simple layering, a single Drum Sampler is often enough; use chains when the layers need their own EQ and timing.

Instrument Rack key and velocity zones

Device: Instrument Rack · Level: intermediate · Genres: all

Layering across the keyboard instead of across the spectrum: Instrument Rack Key zones hand different registers to different chains (sub synth under C2, a saw above), Velocity zones switch layers in by playing force (a click layer only above velocity 100) and the Chain selector crossfades layers from a macro. Fades at the zone borders prevent a hard jump.

  1. Open the Chain list, enable Key and Vel zone editors. Drag each chain's range; drag the fade handles at the ends to overlap by 2–5 semitones (key) or 10–20 velocity steps.
  2. Chain selector: enable the Chain editor, give the Air chain a 64–127 range, map a macro to the Chain selector. The macro now fades the layer in.
  3. Velocity-switched layers need envelope agreement too: a click that only appears above vel 96 should still start within 2 ms of the body.
Zones are layering by condition rather than by band. They are the only way to make a layered instrument that plays differently depending on how it is hit, and they cost nothing to automate because the condition is in the MIDI.

Sample plus synth: hybrid layering

Device: Simpler, Drum Sampler, Operator, Drift · Level: intermediate · Genres: techno, trap, DnB, dubstep

The sample brings character and transient; the synth brings a clean, tunable fundamental with a perfect envelope. Typical pairs: sampled kick + Operator sine sub; sampled Reese + Drift sub; sampled snare + Operator noise crack; vocal chop + Wavetable pad doubling the pitch. The synth owns the sub and tuning, the sample owns everything above.

DecisionSampleSynth
OwnsTransient and 80 Hz–10 kHzSub below 80–100 Hz; pitch
FilterHP 24 dB at the crossoverLP 24 dB at the crossover, or pure sine (no filter needed)
EnvelopeAs recorded; shorten with Simpler decayMatch or exceed the sample's decay by 20–50 ms
PhaseFixed (it is the reference)Align Delay here; sine phase is predictable
TuningTune the tail to the synth (Simpler Detune)Set by MIDI note; use Pitch env for the kick thump
Operator's Pitch envelope with a −12 to −24 semitone initial drop over 20–40 ms gives the synth sub its own "knock" that fuses with the sample's transient. Set the sample's Simpler Start to skip its own sub cycle if the two still fight.

Verifying a layer stack on Spectrum and in mono

Device: Spectrum, Utility, EQ Eight · Level: essential · Genres: all

A procedure, not a sound. Run it on every layered instrument before it goes into the arrangement.

  1. Owner check. Solo each layer, Spectrum Max hold on. Each should cover its band and fall at least 12 dB outside it.
  2. Sum check. All layers on; the summed trace should be smooth with no notch at the crossovers. A notch at the crossover = phase problem: polarity, then Align Delay.
  3. Mono check. Utility Mono on the rack output. Level drop below 1.5 dB and no change in the sub = pass.
  4. Headroom check. Peak meter on the rack output: compare the peak of the sum with the loudest individual layer. If the sum peaks more than 4 dB above the loudest layer, two layers share a peak band; move a crossover.
  5. Interval check. Loop the hit at the track's fastest repetition (16ths at tempo). Nothing should still be audible when the next hit begins, except deliberately.

Settings / targets

Crossover points, widths and level offsets by sound type. Levels are relative to the loudest layer; crossovers are 24 dB/oct unless noted.

SoundLayersCrossoversWidth per layerLevel offsetsDecay order
Techno kickSub sine · body sample · click90 Hz · 1.5 kHzMono · Mono · 60%0 · −2 · −9 dBClick 10 ms, body 150 ms, sub 300 ms
Trap 808 + kick808 (sub+mid) · kick topKick HP 100 Hz, 12 dBMono · 50%0 · −4 dBKick 80 ms, 808 free (sidechain the tail)
Dubstep bassSub sine · growl mid · noise top100 Hz · 3 kHzMono · Bass Mono 150 Hz, 70% · 120%0 · −3 · −12 dBAll equal; gate together
Reese (DnB)Sub sine · detuned saws · optional top80 Hz · 4 kHzMono · 50–80% · 100%0 · −2 · −15 dBSustained; identical release
Snare909/body · crack · room3 kHz LP on body, 1.5 kHz HP on crackMono · 80% · 120%0 · −7 · −14 dBCrack 20 ms, body 120 ms, room 200 ms
Dub chord stabAnalog chord · Wavetable air2.5 kHzBass Mono 200 Hz, 60% · 100%0 · −10 dBIdentical attack 20 ms and release
PadDrift body · Granulator III texture · noise air600 Hz · 5 kHz50% · 100% · 130%0 · −6 · −16 dBRelease 400–800 ms on all
LeadMono saw core · octave-up double · transient pluckNone (same band, tuned) · 1 kHz HP on pluckMono · 80% · 60%0 · −8 · −10 dBPluck 60 ms, others sustained

Width rule of thumb: Bass Mono frequency = the lowest frequency a layer is allowed to be stereo at; 120–180 Hz for club systems, 250 Hz for vinyl, 80 Hz for headphones-only material. Level offsets are the point where the layer is missed when muted but not identifiable when soloed-out.

Common mistakes

Related