ABLETON 12 SOUND DESIGN ATLAS

Famous Downtempo & Chillout Records — How They Were Made

Thirteen downtempo, chillout and nu-jazz records from 1997 to 2013: Groove Armada, Thievery Corporation, Air, Nightmares on Wax, Mr Scruff, Moby, The Cinematic Orchestra, Röyksopp, Zero 7, Lemon Jelly, Emancipator, Tycho and Bonobo. Each was chosen because credits, sleeve notes or interviews say how it was made. The page separates what is documented from what is only heard, and rebuilds each technique in Live 12.

Difficulty
Intermediate
Live edition
Suite
Max for Live
Not needed

Trip-hop has two pages of its own (part one and part two). This page covers the gentler side of slow electronic music: chillout, lounge and nu-jazz records whose making is on the record. A trombone recorded through a rewired hi-fi speaker. A gospel vocal left with its surface noise. A whole album made in a living room with one sampler and a reverb taken from a radio station. A kalimba bought and recorded on the same afternoon. The records run from 1997 to 2013. Each claim is labelled Documented (sleeve credits, Discogs or MusicBrainz catalogue data, an interview or a reputable feature, linked inline) or By ear (listening analysis or scene knowledge, which is not fact). Where a source gives an album-level or general method rather than the record’s own, the page says so. The Live 12 recipes rebuild the technique, not the record. The piano rolls are generic shapes of two bars or less, never transcriptions.

Before you build. No source below gives a tempo, so none is printed. One key is documented (the Moby vocal sample, A minor, per Sound On Sound) and one mode (Zero 7’s “Destiny”, C Phrygian, per Wikipedia). For the rest, drop the record into Live, let auto-warp find the tempo, and put Tuner on a soloed, low-passed copy of the bass to find the root. Most of this music sits somewhere between 80 and 110 BPM. At 90 BPM a bar lasts 2667 ms and a 1/16 note 167 ms; at 100 BPM a bar is 2400 ms and a 1/16 note 150 ms (see BPM to ms). The basics of the style (the lazy break, the felt-not-heard sub, noir Rhodes, the vinyl bed, dub sends) are on the genre page and in the downtempo track walkthrough. This page links to them rather than repeating them. Samples are named because they are credited facts, not as an invitation to use them uncleared.

Overview: the thirteen records

TrackYearArtistLabel · catalogueSignature technique
At the River1997Groove ArmadaTummy Touch · TUCH 011A 50p record, a ten-second sampler, a trombone through a hi-fi speaker
Lebanese Blonde1997 (single 1998)Thievery Corporation4AD · 7243 8 95519 2 1 (French CD)Sitar, horns and percussion players over a lounge-dub production
Kelly Watch the Stars1998AirSource · 7243 8 44978 2 8 (album)Vocoder, Solina, MS-20, Moog and Casiotone around a live drummer
Les Nuits1999Nightmares on WaxWarp · WAP118CDSamples replaced by an orchestra that was written from them
Get a Move On!1999Mr ScruffNinja Tune · ZEN 1280Moondog sample, a blues vocal, a live bassist
Why Does My Heart Feel So Bad?1999MobyMute · CDMUTE230Gospel sample in an S3200, piano chords under it, a sped-up break on a 909
Motion1999The Cinematic OrchestraNinja Tune · ZEN CD45 (album)Sample sketches learned by a jazz band, then re-edited on computer
Eple2001RöyksoppWall of Sound · WALLD071 (not confirmed here)Living-room rig: one Akai sampler, a simple mixer, a radio-station reverb
Destiny2001Zero 7Ultimate Dilemma · UDRCDS043An instrumental bed that waited for the right singer
Space Walk2002Lemon JellyImpotent Fury / XL · IFXLS 150 CDA 1965 NASA spacewalk recording as the lead voice
Soon It Will Be Cold Enough2006EmancipatorSelf-released CDr (no catalogue)Loop software, a violinist, and a student’s home rig
Dive2011TychoGhostly International · GI-145 (album)Soft synth start, analogue chain, own recordings chopped like samples
Cirrus2013BonoboNinja Tune · ZENDNLS357 (digital single)Kalimba notes placed on the transients of a sampled break

The records

Groove Armada — “At the River” (1997): a bargain-bin vocal and a speaker for a microphone

1997 · Tummy Touch (7" TUCH 011, 500 copies, Discogs 318609) · reissued 26 July 1999 on Pepper · tempo and key undocumented · Tools: a sampler with a ten-second limit, synth bass, programmed drums, trombone (Documented) · Genres: chillout, downtempo

Quick start

Load
Simpler, EQ Eight, Saturator
Set
  • Simpler one vocal phrase, Classic mode, Loop on, no sample over ten seconds
  • Drums bounce a normal-tempo break, then lower the tempo with the clip in Re-Pitch mode
  • EQ Eight on the trombone: low cut at 150 Hz, broad bump near 800 Hz, high cut at 5 kHz
  • Saturator Analog Clip for the overloaded edge
You should hear
One vocal phrase looping nearly unchanged over slow, soft drums, with a dark, boxy, band-limited trombone.
Live edition
Standard and up

Why it matters. The record that defined late-90s chillout for many listeners: one old vocal phrase, slow drums, a trombone solo and a lot of air. It was made with almost nothing, and the limitations are audible in the best way.

Documented tools and method. The 1997 7" was limited to 500 copies. Its notes say it “contains a re-recorded sample” of Patti Page’s “Old Cape Cod”, with writing credited to A. Cato (Discogs release 318609). Wikipedia, summarising the Guardian’s “How we made” feature, says the track was written in a cottage in Ambleside in the Lake District. The Page recording came from a 1950s compilation bought for 50p in a local bargain bin. Their sampler could hold only ten seconds at a time. Synth bass and slowed-down programmed drums went under the multi-tracked close-harmony vocal. Andy Cato’s blues trombone solo was inspired by the hymn “Shall We Gather at the River?”. With no microphone in the cottage, a speaker from an Alba hi-fi was rewired to record the trombone. Cato called the result “a real vintage character” (Wikipedia — At the River). The Guardian original could not be read for this page, so these details are second-hand.

By ear: the vocal phrase loops nearly unchanged; the drums are slow and soft with a lot of room between hits; the trombone is dark and band-limited, with little top end.

Signature moves → Live 12

  1. Work inside ten seconds (Documented constraint). Give yourself one Simpler per sound and no sample over ten seconds. Pick the single best vocal phrase from the record, set Simpler to Classic mode with Loop on, and build the track around that phrase alone.
  2. Slow the programmed drums down (Documented approach). Program a normal-tempo break, bounce it, then drop it into an audio clip in Re-Pitch mode and lower the tempo. The drums lose pitch and gain weight together, which is the point. Complex Pro would keep them bright.
  3. A speaker as a microphone (Documented approach, recipe By ear). A loudspeaker used as a mic has a narrow, boxy response. Recreate it with EQ Eight: low cut at 150 Hz, a broad bump near 800 Hz, high cut at 5 kHz. Then Saturator in Analog Clip for the overloaded edge. Or try it for real: a passive speaker wired into a mic preamp works, at low gain.
  4. Hymn as a starting point for a solo (Documented approach). Learn a public-domain hymn tune and improvise a blues phrase around it, as Cato did. Record several passes into Take Lanes and comp the best phrases.
Clearance. The 1997 notes say the sample was re-recorded, which is one way producers get around master rights. You still need the publisher’s permission for the song. A 50p record is not a cleared record.

What to steal: one phrase, one solo, slow drums. Constraint makes the arrangement for you. More slow-drum ideas are in the lazy break recipe.

Thievery Corporation — “Lebanese Blonde” (1997): players over a lounge-dub bed

On Sounds from the Thievery Hi-Fi (ESL Music, 1996/97) · French CD maxi 1998, 4AD (7243 8 95519 2 1, Discogs 91762) · recorded at ESL Studios, Washington DC (Documented) · tempo and key undocumented · Genres: lounge, downtempo, dub

Quick start

Load
Electric, Corpus, Auto Filter, Echo
Set
  • Pluck a plucked Electric or guitar sample, short phrases bent into notes with pitch automation
  • Corpus String, Decay 70%, Tune to the root, Width 60%
  • Auto Filter Band-pass 1.2 kHz, Res 20%
  • Echo 1/8D, Fb 30%, Wobble 15%
  • Sends percussion and horns to the Echo return; the drums stay dry
You should hear
A twangy, sympathetically ringing sitar-like pluck answering the vocal, floating in delay over a tight, dry lounge-dub groove.
Live edition
Suite

Why it matters. The track that made the Eighteenth Street Lounge sound international. It mixes a dub production with Indian, Brazilian and jazz colours, played by session musicians rather than sampled.

Documented tools and method. The 1998 French CD credits Pam Bricker on vocals, Rob Myers on sitar, Rick Harris on horns and Roberto Berimbao on percussion. Writing is credited to Eric Hilton and Rob Garza, production to Thievery Corporation, and the notes say it was recorded at ESL Studios in Washington DC (Discogs release 91762). Wikipedia says the duo met in 1995 at the Eighteenth Street Lounge. They bonded over dub, bossa nova, reggae and jazz records and set out to combine those styles in a studio. Both are credited with bass, guitar and keyboards (Wikipedia — Thievery Corporation). The same players appear on the track when it was included on The Mirror Conspiracy (Wikipedia — The Mirror Conspiracy). No source found here describes the programming, samples or gear.

By ear: a steady, dubby bassline and a laid-back beat; the sitar answers the vocal rather than doubling it; horns enter as short phrases; a lot of delay on the percussion.

Signature moves → Live 12

  1. Programme the bed, hire the colours (Documented approach). Build drums, bass and keys yourself, then record a sitar, horns and hand percussion as audio over it. Each player gets a few takes and a free hand. Comp the takes in Take Lanes.
  2. Sitar without a sitarist (By ear). Collision with Mallet off and a Membrane or String resonator gives a twangy pluck. A closer route: a plucked Electric or guitar sample into Corpus (String, Decay high, Tune to the root) for sympathetic ringing. Keep the phrases short and bend into notes with pitch automation.
  3. Dub sends on the colours, not the bed (By ear). Percussion and horns go to an Echo return, the drums stay dry. This keeps the groove tight while the players float. The send recipe is on the dub space section.
  4. Lounge harmony from bossa records (Documented influence, recipe By ear). Use minor-seventh and major-ninth voicings on Electric with a slow Auto Pan-Tremolo in tremolo mode for the vintage lounge wobble.
Do not over-quantise the players. Session parts recorded against a loop should keep their own timing. Quantising sitar or hand percussion to the grid removes the reason you hired them.

What to steal: a programmed bed is a stage. Let real players add the colour. For more on guitar-family sounds, see Guitar & Band Instruments.

Air — “Kelly Watch the Stars” (1998): a keyboard museum around a live drummer

Moon Safari, January 1998 · Source / Virgin (CD 7243 8 44978 2 8, Discogs 41970) · French CD single 7243 8 95071 2 6, May 1998 (Wikipedia; not confirmed here) · tempo and key undocumented · Tools: Rhodes, vocoder, Solina, Moog, Korg MS-20, Casiotone (Documented credits) · Genres: French downtempo, space pop

Quick start

Load
Vocoder, Analog, Drift, Chorus-Ensemble, Auto Filter
Set
  • Vocoder on the vocal, Carrier External from a sustained Analog saw chord, Bands 20, Formant 0, Depth 100%, Release around 150 ms
  • Hook sung on one or two notes; the carrier plays the harmony
  • Drift one saw, filter wide open, Attack 150 ms, Release 600 ms
  • Chorus-Ensemble Ensemble mode on the Drift strings
  • Auto Filter MS2 circuit after an Analog square lead, Resonance high, Frequency ridden by hand
You should hear
A vocoder hook playing chords from a one-note vocal, a string-machine pad behind it and a squelchy lead.
Live edition
Standard and up

Why it matters. Air’s “La femme d’argent” is on the second trip-hop page. This is the album’s pop side: a robot-voiced hook, a disco-ish pulse and a pile of 1970s keyboards, all played rather than sampled.

Documented tools and method. The album credits list, per track, who played what. On this track Jean-Benoît Dunckel is credited with Rhodes, vocoder, piano, Solina String Ensemble, Moog Solo, Korg MS-20, Minimoog, glockenspiel, Syrinx, clavinet and handclaps. Nicolas Godin is credited with vocoder, Moog bass, Casiotone, handclaps and glockenspiel, and drums are by Marlon B. The album was recorded at Studio Around The Golf and Studio Gang in Paris, with strings at Abbey Road, and mixed at Studio Plus XXX (Discogs release 41970). Wikipedia gives the single’s catalogue and release date (Wikipedia — Kelly Watch the Stars). No session account specific to this track was found. In a track-by-track interview on the album, the band describe the vocoder and talkbox work on another track, “Remember” (Stereogum, 2024).

By ear: two vocoders trade the hook; a string-machine pad sits behind; the drums are real and slightly loose; the bass is a round analogue mono line.

Signature moves → Live 12

  1. Vocoder hook (Documented instrument, recipe By ear). Put Vocoder on the vocal track, Carrier set to External, and route a sustained Analog saw chord in as the carrier. Bands 20, Formant 0, Depth 100%, Release around 150 ms. Sing the hook on one or two notes and let the carrier play the harmony.
  2. Solina-style strings (Documented instrument, recipe By ear). A string machine is a simple divide-down saw organ through an ensemble chorus. In Drift: one saw, filter wide open, Attack 150 ms, Release 600 ms. Follow it with Chorus-Ensemble in Ensemble mode. The Ensemble mode is the sound.
  3. MS-20 squelch (Documented instrument, recipe By ear). Auto Filter has an MS2 circuit modelled on that filter type. Put it after a simple Analog square lead, Resonance high, and ride Frequency by hand.
  4. A cheap keyboard for charm (Documented instrument). A Casiotone-type preset tone: Operator with a single square-ish wave, then Redux at 10 bits. Use it for a doubling line, not a lead.
Credits are not a session report. The liner notes say who played what, not how each sound was set up. The Live recipes here are reconstructions of the instrument types named, not of Air’s patches.

What to steal: play the keyboards; do not sample them. Real performance on cheap instruments is the whole flavour. More on vintage keys: Keys & Mallets and Analog & Drift.

Nightmares on Wax — “Les Nuits” (1999): write the orchestra from the sample

Single 19 April 1999 · Warp (CD WAP118CD, MusicBrainz) · album Carboot Soul, 12 April 1999 · tempo and key undocumented · Tools: undocumented; orchestral and live session (Documented) · Genres: downtempo, soul

Quick start

Load
Tension, Glue Compressor, Auto Filter, Hybrid Reverb, EQ Eight
Set
  • Sections separate tracks for first and second violins, violas, cellos, Tension with slow attacks, each panned
  • EQ Eight on the string group, Low cut 90 Hz, −2 dB 3 kHz
  • Glue Compressor 2:1, Attack 30 ms, −2 dB GR
  • Auto Filter LP 9 kHz
  • Hybrid Reverb shared send, Hall, Predelay 20 ms, 2.6 s
You should hear
Warm strings in one room, glued and filtered so they sit behind a slow, swung beat and breathe in long phrases.
Live edition
Suite

Why it matters. George Evelyn had already released versions of this idea twice, as interludes on his earlier albums. For Carboot Soul he stopped sampling and paid for the sound instead. It is a clear documented case of a sample-based track rebuilt with players.

Documented tools and method. Wikipedia says “Les Nuits” is a revised version of “Nights Interlude” and “Nights Introlude” from his two previous albums. It samples the Quincy Jones version of “Summer in the City”. Instead of the sampled and synthesised tracks he had made before, it used a full orchestral and live recording session inspired by the samples, which Evelyn then mixed (Wikipedia — Carboot Soul). The single’s UK CD is WAP118CD, released 19 April 1999 (MusicBrainz). No source found here names the arranger, the studio or the gear.

By ear: warm strings and a soft electric piano; a slow, swung hip-hop beat; the track breathes in long phrases rather than loops.

Signature moves → Live 12

  1. Return to your own old idea (Documented approach). Keep a folder of interludes and sketches. A short idea that worked twice can become a full track the third time. Open the old project, export its MIDI, and rewrite it at full length.
  2. Sample sketch, then score (Documented approach). Build the first version from samples. When it works, transcribe your own sketch (not the record) into MIDI parts for strings, keys and bass. Hand those parts to players, or render them with Suite strings as a mock-up before a session.
  3. Mock up the orchestra (By ear). Split the parts across separate tracks: first and second violins, violas, cellos. Use Sampler string patches or Tension with slow attacks. Give each section its own pan position and a shared Hybrid Reverb send, so they sit in one room.
  4. Mix the session yourself (Documented approach). Evelyn mixed the live recordings. Treat the orchestra like a sample: group it, glue it with Glue Compressor and filter it so it sits behind the beat.
Replaying is not clearing. If your score reproduces the sampled song’s melody or harmony closely, you still owe the publisher. Replaying removes the master licence, not the composition.

What to steal: use samples to find the track, then pay players to make it yours. For string programming, see Strings, Brass & Physical Models.

Mr Scruff — “Get a Move On!” (1999): Moondog, a blues shout and a real bassist

May 1999 · Ninja Tune (12" ZEN 1280, Discogs 11366) · album Keep It Unreal · tempo and key undocumented · Tools: undocumented; bass by Sneaky, mixed and engineered by Phil Kirby (Documented credits) · Genres: downtempo, nu-jazz, breaks

Quick start

Load
Drum Rack, Simpler, Compressor, EQ Eight, Utility
Set
  • Drum Rack one blues-shout phrase per pad, triggered only at the end of four- and eight-bar phrases
  • Shouts Simpler One-Shot, Fade Out 20 ms
  • Compressor on the bass DI, 3:1, Attack 15 ms, Release 120 ms
  • EQ Eight Low cut 35 Hz, +2 dB 700 Hz
  • Utility Mono
You should hear
A looping outsider-record hook, short vocal shouts as punctuation and a played, walking-ish bass over a light breakbeat.
Live edition
Standard and up

Why it matters. The downtempo side of Mr Scruff: a crate-digger’s track that puts an avant-garde street musician next to jump blues, with a played bassline holding them together. It became one of Ninja Tune’s best-known records through its use in advertising.

Documented tools and method. The 12" credits bass to Sneaky and mixing and engineering to Phil Kirby. Its notes say the A-side samples Moondog’s “Bird’s Lament” (Discogs release 11366). The Moondog source is “Bird’s Lament (In Memory of Charlie Parker)” from the 1969 album Moondog (Bill Mitchell, 2014). The vocal source is attributed in different ways. The song page names T-Bone Walker’s “Hypin’ Woman Blues” and also lists Rubberlegs Williams (Wikipedia — Get a Move On!). The artist page names Shifty Henry (Wikipedia — Mr Scruff). Treat the vocal source as disputed. No gear or software is documented.

By ear: the Moondog saxophone figure loops as the hook; the vocal shouts come in as short punctuations; a walking-ish bass line keeps it moving; the beat is a light breakbeat.

Signature moves → Live 12

  1. Hook from an outsider record (Documented approach). Look for loops in unusual places: street musicians, library records, early minimalism. Moondog’s music has steady, repeating patterns that loop naturally. Warp in Complex for a full ensemble, or Beats if it is percussive.
  2. Vocal shouts as punctuation (Documented source type, recipe By ear). Cut short phrases from a blues or jump record into a Drum Rack, one phrase per pad. Trigger them at the end of four- and eight-bar phrases only.
  3. A real bassist over the samples (Documented credit). Record a bass player over the loop rather than programming the bass. If you play yourself, track several passes and comp. Keep it in mono and compress lightly.
  4. Check your credits (Documented problem). Three published sources name three different vocal sources. Write down the exact record, side and timing of each sample when you make the track.
Do not trust one sample database. As this record shows, published attributions can disagree. Your own notes at the time of making are the only reliable record.

What to steal: one strange loop, one human voice, one human bassist. For drum-break programming, see Breakbeat Chopping.

Moby — “Why Does My Heart Feel So Bad?” (1999): a gospel sample left dirty

Single 11 October 1999 · Mute (CD CDMUTE230, Discogs 31621) · album Play (CDStumm172, Discogs 2285) · vocal sample in A minor (Documented) · tempo undocumented · Tools: Akai S3200, Cubase on a Mac, Roland TR-909, Emu Proformance, Yamaha SY22/SY85, Roland Juno-106, Soundcraft Spirit, ADAT, DAT (Documented) · Genres: downtempo, electronic gospel

Quick start

Load
Simpler, Tuner, Drum Rack, Drum Buss, Glue Compressor
Set
  • Simpler the gospel phrase, Classic, Warp off; find its key with Tuner (A minor)
  • Break audio clip in Re-Pitch mode, sped up, over a 909 kick and snare in a Drum Rack
  • Drum Buss Drive 10%, Transients +20%
  • Glue Compressor 2:1, −2 dB GR
  • Noise do not de-noise the sample; keep the crackle
You should hear
A dirty gospel phrase with its crackle intact, piano chords following it and a crisp, sped-up break over 909 kick and snare.
Live edition
Standard and up

Why it matters. Play is the best-selling downtempo album of its era, and this track has a full Sound On Sound walkthrough. It shows the whole method: a sample, a few chords, a drum machine, and a mix done at home.

Documented tools and method. Sound On Sound’s walkthrough says the track was recorded and mixed in Moby’s home studio, straight to DAT, with no recall. He made 10 to 15 mixes and picked by ear. Sequencing was in Cubase on a Mac, through a Soundcraft Spirit desk, with an Alesis ADAT. The vocal sample, in A minor, sat in an Akai S3200. The track was built from the vocal first, then the piano chords. The main piano is an Emu Proformance, doubled by an old Yamaha synth. Drums are a Roland TR-909 with a hip-hop break on top, sped up to tempo rather than time-stretched. Strings come from a Yamaha SY22 and SY85, and sub-bass from a Juno-106. A Yamaha SPX900 gave the reverb. He tried a surface-noise filter and dropped it, deciding to “leave the vocal with all the scratches and surface noise” (SOS, February 2000). SOS describes the source as a 1953 gospel choir recording. Wikipedia names a different source and flags its own citation as unreliable (Wikipedia), so the source is uncertain here. Play was made in his Manhattan home studio from 1997 (Wikipedia — Play).

By ear: the piano chords follow the choir; the beat is crisp with a break on top; the sub is felt more than heard; strings rise in the second half.

Signature moves → Live 12

  1. Vocal first, chords second (Documented approach). Load the phrase into Simpler (Classic, Warp off, so it plays at its own pitch). Find its key with Tuner, then write piano chords that follow it. Live 12’s Scale setting (here A minor) keeps the MIDI editor on the right notes.
  2. Speed the break up, do not stretch it (Documented approach). Put the break in an audio clip in Re-Pitch mode, so it speeds up and rises in pitch together. Layer it over a 909 kit in a Drum Rack (Suite’s 909 kit), with the 909 kick and snare as the backbone.
  3. Keep the noise (Documented decision). Do not de-noise the sample. If anything, put the crackle on its own track so you can ride it. The vinyl layer recipe covers adding noise; here, you leave the original’s.
  4. Mix to a stereo file in one pass (Documented approach). Make several full mixes in a row, each a real-time pass with fader moves. Record them to an audio track with Resampling as the input, then pick the best by ear the next day.
Re-Pitch changes the key. A break sped up from 90 to 98 BPM rises about 1.5 semitones. That is fine for drums but not for anything tonal: never put the vocal or a melodic loop in Re-Pitch unless you have worked out where it lands.

What to steal: leave the flaws in. The surface noise is part of what makes the record feel old and human. For the sub, see felt-not-heard bass.

The Cinematic Orchestra — Motion (1999): sample sketches, a jazz band, then the edit

27 September 1999 · Ninja Tune (CD ZEN CD45, MusicBrainz; LP ZEN 45) · tempo and key undocumented · Tools: sample collections, live jazz players, computer editing (Documented, album level) · Genres: nu-jazz, downtempo

Quick start

Load
Simpler
Set
  • Sketch a two- to four-bar loop of drum, bass and melodic samples
  • Players a straight pass, then several free passes, all kept in Take Lanes
  • Edit mute the sketch, cut and loop the best bars, Consolidate (Ctrl/Cmd+J)
  • Simpler Slice, by Region for phrases
  • Fades 2–10 ms at every edit; no quantising inside a looped bar
You should hear
Played jazz parts looped more exactly than a band would repeat them, with the human feel kept inside each loop.
Live edition
Intro and up

Why it matters. The clearest documented nu-jazz method: write with samples, have a band learn and improvise on the sketch, then cut the band’s playing back together like samples. It sits between a DJ’s record and a jazz record.

Documented tools and method. Wikipedia says Jason Swinscoe assembled collections of samples (drum patterns, basslines, melodies), then had musicians learn them and improvise around them. He then re-mixed the resulting drafts on computer to make the final album. Credited players include Tom Chant (saxophones, piano, electric piano), Jamie Coleman (trumpet, flugelhorn), Phil France (acoustic and electric bass), T. Daniel Howard (drums), Alex James (piano, electric piano) and Saidi Kanba (percussion) (Wikipedia — Motion). The page cites a 2002 Exclaim interview, which could not be read here. This is an album-level method; no track-level session account was found.

By ear: loops that sound sampled but are played; long modal vamps; drums with a hip-hop weight under jazz playing; edits that repeat a phrase more exactly than a band would.

Signature moves → Live 12

  1. Sketch with samples (Documented approach). Build a two- to four-bar loop from drum, bass and melodic samples in Session View. This is a score, not the record: it tells the players the groove and the harmony.
  2. Players learn the sketch, then stretch it (Documented approach). Record each player over the loop: first a pass that plays it straight, then several free passes. Use Take Lanes so every pass is kept.
  3. Edit the band like a sampler (Documented approach). Mute the sketch. Cut the best bars from each take, loop them, and rearrange them in Arrangement View. Consolidate edited sections (Ctrl/Cmd+J) so you can treat them as new clips.
  4. Keep human timing inside machine repetition (By ear). Loop a played bar exactly, but do not quantise inside it. The loop repeats; the feel inside it stays human.
Edit points need room tone. Cuts in an acoustic recording click or jump when the room changes. Use short fades (2–10 ms) at every edit, and crossfade over cymbal tails instead of cutting them.

What to steal: treat your own sessions as source material. More on comping and resampling: Resampling Workflow.

Röyksopp — “Eple” (2001): one sampler, one mixer, one borrowed reverb

Single 23 July 2001, Wall of Sound (CD WALLD071, per Wikipedia; not confirmed here) · album Melody A.M., 8 October 2001 (CD WALLCD027, Discogs 18950) · tempo and key undocumented · Tools: Akai sampler, simple mixer, radio-station reverb (Documented, album level) · Genres: chillout, electronic pop

Quick start

Load
Simpler, Redux, EQ Eight, Hybrid Reverb
Set
  • Simpler a two- or four-bar keyboard backdrop in Slice mode
  • Second record a disco element matched with clip Transpose and Detune
  • Group Redux Bits 12, Rate around 30 kHz, EQ Eight high cut at 14 kHz
  • Hybrid Reverb the only reverb: one return, Plate IR, Decay 1.6 s, Predelay 15 ms; everything sent to it
You should hear
A bright, chiming keyboard figure over a gentle disco pulse, everything sharing one sampler character and one room.
Live edition
Suite

Why it matters. Melody A.M. made Bergen a byword for melodic chillout. Its gear list is so short that it doubles as a lesson in finishing records with very little.

Documented tools and method. Stereogum’s 20th-anniversary feature says the album was made in the duo’s living room in Bergen. Svein Berge and Torbjørn Brundtland describe the core rig: “an Akai sampler, a really simple mixer and a reverb” taken from a radio station. The same piece says “Eple” samples the keyboard backdrop of Bob James’s “You’re As Right As Rain” (from Two, 1975) and elements of Skyy’s “Disco Dancing” (Stereogum, 2021). Wikipedia gives the single’s catalogue and date and the Bob James source (Wikipedia — Eple). No synth or sequencer specific to this track is documented.

By ear: a bright, chiming keyboard figure over a gentle disco pulse; the reverb is the same on everything, which makes the track sound like one room.

Signature moves → Live 12

  1. One reverb for the whole track (Documented approach). Put a single Hybrid Reverb (a plate or chamber IR) on one return and send everything to it at different levels. Do not add reverb on individual tracks. One space glues a sample-based track together.
  2. Take the backdrop, not the lead (Documented approach). “Eple” uses a keyboard backdrop rather than a melody. Look for accompaniment passages in records: comping chords, arpeggios, pads. Chop a two- or four-bar section into Simpler in Slice mode.
  3. Combine with a second, unrelated record (Documented approach). Add elements from a different genre (here a disco record) for rhythm. Match keys with clip Transpose and Detune before you start arranging.
  4. Sound like one Akai (By ear). Group all sample tracks, then put Redux (Bits 12, Rate around 30 kHz) and a gentle EQ Eight high cut at 14 kHz on the group, so everything shares one character.
Two samples in two keys. Transposing a full-mix sample by more than a semitone or two in Complex Pro starts to sound processed. Pick sources already close in key, then use Detune (cents) for the last step.

What to steal: limit yourself to one sampler sound and one reverb, and finish the track. For chopping keyboard parts, see Sampler-Era Sound Design.

Zero 7 — “Destiny” (2001): a bed built for a voice

Single 6 August 2001 · Ultimate Dilemma (CD1 UDRCDS043, MusicBrainz) · album Simple Things, 23 April 2001 · C Phrygian (Documented, Wikipedia citing Hooktheory) · tempo undocumented · Tools: Rhodes or piano to start (Documented, general method) · Genres: chillout, downtempo soul

Quick start

Load
Electric, Chorus-Ensemble, EQ Eight
Set
  • Electric soft Mallet, Pickup near the tines
  • Chorus-Ensemble Chorus, Amount 25%
  • Auto Pan-Tremolo Tremolo, 1/8D, Amount 20%
  • EQ Eight Low cut 100 Hz, −3 dB 1.5 kHz
  • Scale C Phrygian; lean on the flat second (D♭ over C)
You should hear
A soft, wobbling electric piano bed in Phrygian colour, leaving space in the middle of the spectrum for a voice.
Live edition
Standard and up

Why it matters. The record that introduced Sia to a wide audience, and a clean example of an instrumental written as a bed for a singer who had not yet been found.

Documented tools and method. Wikipedia says the song began as an instrumental, and that Henry Binns felt it was missing something until Sia came in. Binns and Hardaker wrote and produced it, with Sophie Barker and Sia as co-writers. It gives the mode as C Phrygian (Wikipedia — Destiny). In a 2019 interview Binns says he usually starts on a Rhodes or piano and builds a melody over it. Sometimes they start from a sample. Their tracks are rarely beat-driven. Early on they made elaborate tracks and then stripped them back to make a bed for a singer. On “Destiny”, many singers tried it before Sia recorded it late, quickly. Binns: “you either got it in one take or you’ll never get it!” (Songwriting Magazine, 2019). The pair had worked as engineers at RAK Studios in London (Wikipedia — Zero 7).

By ear: soft electric piano and strings; a gentle, unobtrusive beat; two female voices in close harmony; a lot of space in the middle of the spectrum for them.

Signature moves → Live 12

  1. Start at the keys (Documented approach). Electric (Live’s electric piano) with a soft Mallet and the Pickup near the tines, through Chorus-Ensemble and a gentle Auto Pan-Tremolo. Write the harmony first and leave the beat for last.
  2. Use the Phrygian colour (Documented mode). Set the clip’s Scale to C Phrygian in Live 12. The flat second (D♭ over C) is the mode’s signature. Lean on it in a pad or bass move, then let the singer find the melody.
  3. Strip back to a bed (Documented approach). When the track is “finished”, mute half of it. Remove anything in the 300 Hz–3 kHz range that moves while the vocal would be singing. A bed is a track with a hole the shape of a voice.
  4. Audition singers on the same bed (Documented approach). Export the bed and send it to several singers. Keep their takes on separate tracks and compare. Do not rearrange for each singer until one is chosen.
Too full a bed loses singers. If every bar has a moving keys part, a strong melody has nowhere to go. Binns’ own account is that the early, elaborate versions were hard for singers to fit.

What to steal: write for a voice you do not have yet. For close vocal production, see whisper vocal and Voice & Vocals; for Live 12 scale tools, Tuning, Scales & MPE.

Lemon Jelly — “Space Walk” (2002): a found voice as the lead instrument

Single 7 October 2002 · Impotent Fury / XL (CD IFXLS 150 CD, MusicBrainz) · album Lost Horizons (IFXLCD 160) · tempo and key undocumented · Tools: undocumented · Source: 1965 Gemini 4 spacewalk audio (Documented) · Genres: chillout, downtempo

Quick start

Load
EQ Eight, Roar, Erosion, Echo
Set
  • Voice warped in Complex Pro, phrase starts moved to bar lines
  • EQ Eight Low cut 300 Hz, High cut 3 kHz
  • Roar Single, Drive 5 dB
  • Erosion Wide Noise, Amount 15%
  • Echo send, 1/2, Fb 30%
You should hear
A radio-band voice with static floating over a slow groove, the music swelling in the gaps between transmissions.
Live edition
Suite

Why it matters. Lemon Jelly built whole tracks around spoken word: a voiceover, a list of place names, a children’s song. “Space Walk” uses a NASA recording as its only voice, so the music has to score a real event.

Documented tools and method. Wikipedia says the track uses Ed White’s 1965 spacewalk recording from the Gemini 4 mission, chosen after Nick Franglen and Fred Deakin listened to the album Flight to the Moon (1969). The album’s other tracks show the method: on “Ramblin’ Man” an Errol Flynn sample was dropped because it named only four places, and John Standing recorded around 200 instead. On “Nice Weather for Ducks” the original children’s recording could not be cleared, so Enn Reitel re-recorded the vocal (Wikipedia — Lost Horizons). Franglen told York Vision that voices often start as samples from old records or television. Sometimes they are used as they are, sometimes re-recorded by a chosen performer, with pitch and pace adjusted. He says there is “no real formula to it, it just has to feel right” (York Vision). No gear is documented.

By ear: radio-band voice and static float over a slow, building groove; the music swells in the gaps between transmissions; guitar and keys repeat a gentle figure.

Signature moves → Live 12

  1. Score the speech, not the beat (Documented approach). Lay the spoken recording on its own track first. Mark the gaps between phrases with locators. Write the music so it rises in the gaps and drops under the words.
  2. Fit the voice to the tempo (Documented approach). Warp the speech in Complex Pro and move warp markers so phrases start on bar lines. Keep the speaking speed natural inside each phrase; only the start points move.
  3. Radio-transmission colour (By ear). Band-limit the voice (300 Hz–3 kHz), add Roar or Saturator for grit, and a little Erosion in Wide Noise mode for static. If the source is already band-limited, add nothing but level.
  4. Re-record when you cannot clear (Documented approach). If a voice cannot be cleared, hire a performer to read the same kind of material in the same style. Write a new script rather than copying the original text.
“Public” recordings are not always free. US government recordings such as NASA audio are generally public domain, but a compilation album of them may carry its own rights. Use the original source, not a record that reissued it.

What to steal: a voice with a story gives the arrangement its shape. For more on vocal processing, see Advanced Voice.

Emancipator — Soon It Will Be Cold Enough (2006): loop software and a violin

2006 · self-released CDr, no catalogue number (Discogs 1613410) · reissued 2008 by Hydeout Productions · tempo and key undocumented · Tools: Acid Pro and Reason (Documented, period of the album) · Genres: downtempo, chillhop

Quick start

Load
EQ Eight, Compressor, Utility, Hybrid Reverb, Glue Compressor
Set
  • Violin close mic, two takes panned slightly apart
  • EQ Eight Low cut 180 Hz, −2 dB 2.5 kHz
  • Compressor 2:1, Attack 20 ms; then Utility Width 120%
  • Hybrid Reverb send, Hall, 2.2 s
  • Master Glue Compressor at 2:1, then a gentle high shelf cut
You should hear
Violin over hip-hop drums and loops that change by adding and removing layers, with a soft, warm master.
Live edition
Suite

Why it matters. A student record, burned to CD at home and sold through MySpace, that Nujabes’ label picked up and reissued. It is a model for the bedroom downtempo that followed: loop software, strings and patience.

Documented tools and method. Wikipedia, citing OregonLive, says Doug Appling made and self-released the album in 2006 while studying at the College of William & Mary, burning the CDs himself. He was producing in Acid Pro and Reason at the time and moved to Ableton Live in 2008–09. He began by chopping loops in Acid Pro in high school. He studied violin from age 4 to 12 and later played guitar, drums and bass. Hydeout Productions reissued the album in 2008 (Wikipedia — Emancipator). The CDr credits Cindy Kao on violin and Thao Nguyen on vocals on “When I Go” (Discogs release 1613410). No samples are documented.

By ear: violin and acoustic guitar over hip-hop drums; loops that change by adding and removing layers; a soft, slightly lo-fi master.

Signature moves → Live 12

  1. Acid-style loop work in Session View (Documented tool, recipe By ear). Acid Pro was built for loops that follow tempo and key. Session View does the same: drop loops into slots, warp them, and build sections by launching scenes. Record the scene changes to Arrangement when it works.
  2. A Reason-style rack (Documented tool, recipe By ear). Reason encourages one instrument per channel with its own effects. In Live, make an Instrument Rack per part with its effects inside, so each part can be saved and reused.
  3. Record a string player (Documented credit). Record a violinist over the loops, close-miked, several takes. If you play, record yourself. Layer two takes panned slightly apart for width.
  4. A lo-fi glue (By ear). On the master, Glue Compressor at 2:1, then a gentle high shelf cut. The effect is warmth, not crunch.
Loop libraries carry licences. Commercial loop packs are usually royalty-free for music, but read the terms. Chopped loops from records are not covered by any loop-library licence.

What to steal: start from loops, finish with a player. For Session View workflow, see Performance & Live.

Tycho — “Dive” (2011): a soft synth start, an analogue chain, and own recordings chopped like samples

Album Dive, Ghostly International (CD GI-145, 15 November 2011, MusicBrainz) · title track, 8:19 · tempo and key undocumented · Tools: Minimoog, Roland Space Echo, transformer preamps, Distressors, REAPER (Documented, album level) · Genres: downtempo, chillwave, ambient pop

Quick start

Load
Echo, Saturator, Compressor, Redux, Simpler
Set
  • Echo 1/8D, Fb 25%, Noise 20%, Wobble 25%
  • Saturator Analog Clip, Drive 4 dB
  • Compressor 4:1, Attack 30 ms, Release auto
  • Redux Bits 12, Rate above 30 kHz, Jitter off; then print to audio
  • Simpler a recorded phrase in Slice mode, replayed from MIDI; bass and kick kept clean
You should hear
Soft, saturated synth lines with tape-like wobble and a chopped, sampled-sounding phrase over slightly crushed drums.
Live edition
Suite

Why it matters. The record that set the warm, sun-faded sound of 2010s downtempo. Scott Hansen has explained his process in detail, which makes it unusually easy to rebuild.

Documented tools and method. Wikipedia says the album took about five years, was made mostly with hardware and analogue gear, with many sounds from a Minimoog, and was arranged, EQ’d and compressed in REAPER. It was his first album with mic-recorded acoustic instruments and outboard gear. The title track was the only one begun on digital equipment, starting on a software synth before analogue parts were added (Wikipedia — Dive). In XLR8R’s studio feature Hansen describes his chain as synth, then analogue delay, then preamps, then compression, with more processing in the box. He uses a Roland Space Echo for tape saturation and warmth and Distressors as compressors. He starts with about 16 bars of melody, adds bass early, and arranges drums last over a loop template; Dive uses a lot of sampled drums, and most songs began on guitar. “It’s always been hardware” (XLR8R, 2012). To EARMILK he said that on “Dive” much of what followed the soft synth became analogue, including bass guitar and drums. He records sounds, chops them into notes and phrases on the timeline, and often runs tracks through old 12-bit converters (EARMILK, 2015).

By ear: soft, saturated synth lines with tape-like wobble; a chopped guitar phrase; drums that sound sampled and slightly crushed; a long, slow build.

Signature moves → Live 12

  1. Treat a soft synth like hardware (Documented approach). Hansen ran software synths as if external: out to a console and back in. In Live, put the synth on its own track, route its output to an audio track through a chain of “outboard” devices, and record the result. Then delete the synth and work with the audio.
  2. The analogue chain (Documented order). Synth → delay → preamp → compression. In Live: Echo (Character: Noise and Wobble up for tape), then Saturator (Analog Clip, gentle) as the preamp, then Compressor with a slow attack.
  3. Chop your own recordings (Documented approach). Record a guitar or synth phrase, slice it into single notes with Simpler’s Slice mode, and replay the slices from MIDI. It sounds sampled, but every note is yours.
  4. Twelve-bit warmth (Documented approach). Put Redux on the drum group or the master bus: Bits 12, Rate high enough to keep the top (above 30 kHz), Jitter off. It adds the crisp, slightly gritty edge he describes.
Wobble on everything is seasick. Tape-style pitch wobble works on one or two layers. Put it on the bass and drums too and the tuning goes vague. Keep the bass and kick clean.

What to steal: make sounds, print them, and then chop them as if someone else made them. More analogue character: Tape, Lo-Fi & Analog Character.

Bonobo — “Cirrus” (2013): a kalimba placed on a break’s transients

Single 21 January 2013 · Ninja Tune (digital ZENDNLS357, MusicBrainz) · album The North Borders (CD ZENCD195, MusicBrainz) · tempo and key undocumented · Tools: kalimba, sampled drum break, percussion-record chimes, Prophet-5, Logic (Documented) · Genres: downtempo, electronic

Quick start

Load
Sampler, Analog, Hybrid Reverb, Echo
Set
  • Rhythm Convert Drums to New MIDI Track from the break, copy it to the melody clip, re-pitch each note by hand
  • Sampler your own kalimba notes mapped in the Zone tab, kept above about 250 Hz
  • Echo 1/16D, Fb 20%
  • Analog two saws, low-pass at around 300 Hz, cutoff automated upward across the whole build
  • Chimes bell or chime hits through Hybrid Reverb with a long shimmer, high and quiet
You should hear
A kalimba locked to the break like a second percussion part, chimes shimmering above and a bass that opens slowly.
Live edition
Suite

Why it matters. A modern downtempo classic with an unusually clear account of how it was made, from the shop where the instrument was bought to the synth that plays the bass. Its central idea, giving a melodic instrument the rhythm of a drum break, is easy to steal.

Documented tools and method. On the Tape Notes podcast, Simon Green says he had moved from London to New York in 2011, and his first setup was a Mac, a pair of monitors and a Prophet-5. He was working in Logic and moved to Ableton after The North Borders. The track starts with a kalimba bought in a Lower East Side shop and recorded the same day. Its notes were mapped to the transients of a drum break taken from a record; the chime sounds came from a percussion record. The bass is the Prophet-5, with its filter opened up through the build. He made the track in one or two days, and says he was “kind of sad when it was done” (Tape Notes TN:144). Wikipedia gives the album and single dates and catalogue (Wikipedia — The North Borders).

By ear: the kalimba pattern locks to the drums like a second percussion part; chimes shimmer above; the bass opens slowly so the track grows without new parts.

Signature moves → Live 12

  1. Give a melody the break’s rhythm (Documented approach). Put the break in an audio clip, right-click → Convert Drums to New MIDI Track, or Extract Groove. Copy the resulting rhythm into a MIDI clip for your melodic instrument, then change each note’s pitch by hand. Or apply the extracted groove from the Groove Pool to the melody clip.
  2. Record an instrument the day you get it (Documented approach). A cheap acoustic instrument (kalimba, music box, thumb piano) recorded note by note into Sampler gives you a playable instrument. Drag the notes into Sampler’s Zone tab and map each to its key.
  3. A filter opening as the arrangement (Documented approach). Write a simple Prophet-style bass in Analog or Drift: two saws, low-pass at around 300 Hz. Automate the cutoff upward across the whole build, so the track grows with no new parts.
  4. Chimes from a percussion record, or your own (Documented approach). Sample bell or chime hits, put them through Hybrid Reverb with a long shimmer, and keep them high and quiet.
Watch the low end of a re-pitched groove. When a melodic part copies the kick’s rhythm, low notes on kick hits pile up energy below 200 Hz. Keep the instrument’s notes above about 250 Hz, or high-pass it at 200 Hz.

What to steal: borrow rhythm from one place and pitch from another. Groove tools in Live: MIDI Tools; playable instruments from your own recordings: Sampler, Simpler & Drum Rack.

Patterns across the canon

Read these thirteen together and a few habits keep coming back.

  1. Small rigs, finished records. A ten-second sampler in a cottage (Groove Armada), one Akai, a simple mixer and a radio-station reverb (Röyksopp), a home studio mixed straight to DAT (Moby), Acid Pro and Reason in a student room (Emancipator), a Mac, two monitors and a Prophet-5 (Bonobo). All Documented.
  2. Samples first, players later. Nightmares on Wax replacing samples with an orchestra, The Cinematic Orchestra teaching sample sketches to a band, Mr Scruff’s live bassist, Thievery Corporation’s session sitar and horns. All Documented.
  3. Keep the flaw. Moby left the vocal’s surface noise; Groove Armada recorded through a hi-fi speaker; Tycho runs tracks through old 12-bit converters. All Documented.
  4. The voice comes last, or from somewhere else. Zero 7 wrote a bed and auditioned singers; Lemon Jelly scored NASA audio and re-recorded uncleared speech with actors; Air played the hook through vocoders. Documented (Air’s vocoder is a credit, not a session account).
  5. Rhythm borrowed from drums. Bonobo mapping a kalimba to a break’s transients is Documented; the same idea underlies chopping melodic slices to a drum pattern (Röyksopp, Tycho), which is By ear.
  6. Arranging by filter and layer, not by drop. The Prophet-5 filter opening on “Cirrus” is Documented. Long, slow builds across the other records are By ear.
HabitRecordsLive 12 core
Sample sketch, then playersLes Nuits, Motion, Get a Move On!, Lebanese BlondeSession View sketch, Take Lanes, Consolidate
One sampler’s characterEple, At the River, Why Does My Heart Feel So Bad?Simpler, Redux on the group, one reverb return
Speed and pitch togetherAt the River, Why Does My Heart Feel So Bad?Re-Pitch warp mode
Bed for a voiceDestiny, Space Walk, Kelly Watch the StarsEQ Eight carve, Complex Pro, Vocoder
Print through a chainDiveEcho Character, Saturator, Compressor, Redux, Resample
Groove transferCirrusConvert Drums to MIDI, Extract Groove, Sampler zones
Where documentation was thin. Several accounts are second-hand: Groove Armada’s method comes through Wikipedia’s summary of a Guardian feature that could not be read here, and The Cinematic Orchestra’s method through Wikipedia’s summary of a 2002 interview. For Air, Thievery Corporation and Mr Scruff only credits and samples are documented, with no session account. Nightmares on Wax rests on a single Wikipedia passage. Röyksopp’s and Tycho’s gear is album-level. Moby’s sample source is disputed between SOS and Wikipedia, and Mr Scruff’s vocal source is attributed three ways. Single catalogue numbers for “Kelly Watch the Stars” and “Eple” come from Wikipedia only and are marked as not confirmed here. Nicolas Jaar was considered, but no readable source on how his records were made could be reached for this page.

Related