ABLETON 12 SOUND DESIGN ATLAS

Famous Dub Techno Records — How They Were Made

Twelve dub techno records from Burial Mix’s first 10" (1996) to Porter Ricks’ return on Tresor (2017), chosen because something public says how they were made. For each one: what the sources say, what is only heard, and how to rebuild the method in Live 12.

Difficulty
Advanced
Live edition
Suite
Max for Live
Not needed

The first deep and dub techno page covered the canon: Phylyps Trak, Quadrant Dub, M4, Mango Drive, Cyan, Biokinetics, Königsforst, The Coldest Season and Voices From The Lake. This page picks up the rest of the dub lane: the Burial Mix vocal side of Basic Channel, the Chain Reaction roster after its first year, Pole’s defective filter, and the second and third generations who took the sound to Manchester, Reykjavík, Detroit, Amsterdam and Los Angeles. The records were chosen for their documentation, not their fame, and in this genre documentation is scarce. These artists rarely list their gear, and several deliberately hide behind a label or a process. Where gear is missing, the page teaches the process facts that are public: mix the dub live to stereo, record the singer from the first take, use a defective filter as a drum machine, hide a recorder in the street at night, edit someone else’s records into one long mix. Every claim is labelled Documented (linked inline to the interview, label text, sleeve credit or feature it comes from) or By ear (listening analysis or scene knowledge, which is not fact). The Live 12 recipes rebuild a method, never a record. No melodies are transcribed: every piano roll and step grid is a generic shape of at most two bars.

Documented
Stated by the artist, the label’s own release text or catalogue, sleeve credits (as listed on Discogs or MusicBrainz), or a named feature or review, linked inline. Where the source is a review, the page says so: a reviewer’s description is published listening, not a studio fact.
By ear
Listening analysis or scene consensus. Treat it as a hypothesis and check it against your own copy.
Undocumented
No public source found. Gear, BPM and key are left undocumented rather than guessed.
None of the BPMs or keys on this page is documented. Before building anything, drop the record into Live: auto-warp detects the tempo, and the Tuner on a soloed, low-passed copy of the sub gives you the root. Most of the genre sits between about 115 and 128 BPM. At 120 BPM one bar is 2000 ms, a 1/4 is 500 ms and a 3/16 is 375 ms. At 125 BPM a bar is 1920 ms, a 1/4 is 480 ms and a 3/16 is 360 ms. Use BPM to ms for the rest. The basic chord, delay, kick, sub, hiss and pad recipes are already on the dub techno genre page, so this page links to them instead of repeating them.

Overview: the twelve records

RecordYearArtistLabel / cat.Signature method
Never Tell You1996Rhythm & Sound w/ Paul St. HilaireBurial Mix BM-01Vocal side first, version second, mixed live to stereo
Lantau1997MonolakeChain Reaction CR-15Hong Kong field recordings worked into a sampler-and-step-sequencer track
11998PoleKiff SM KIFF012CDA defective Waldorf 4-Pole filter as the percussion section
Huone1999Vladislav DelayChain Reaction CR-2622 minutes of live reprocessing in a dense reverb
Vibrant Forms1999FluxionChain Reaction CRD-07Repetition with heavy live effects processing, source sound secondary
Arrange and Process Basic Channel Tracks2002ScionTresor 200Someone else’s catalogue re-edited into one DJ-mix-shaped album
New World Observer2005Deadbeat~scape sc27cdMicrosamples, with the echo as the lead voice, made entirely in software
Rigning2009YagyaSending Orbs SO 011One long track in ten movements: a city asleep in rain
Liumin2010Deepchord presents EchospaceModern Love LOVE064Tokyo material, and a second disc distilled from the first
Hash-Bar Loops2011DeepchordSoma SOMACD091Hidden-recorder field recordings and synths used only as triggers for effects chains
Topdown Dialectic2013Topdown DialecticTailings; reissued Peak Oil PEAK15Layered laptop textures, quiet rhythms, a process instead of a person
Anguilla Electrica2017Porter RicksTresor 295Self-written software to generate, mix and master in performance

The records

Rhythm & Sound w/ Paul St. Hilaire — “Never Tell You” (1996): the vocal first, then the version

1996 · Burial Mix BM-01, 10" (Documented: Basic Channel catalogue, MusicBrainz) · BPM undocumented (By ear: a slow steppers pulse, check with warp) · key undocumented · Tools: undocumented for this record · Level: intermediate

Quick start

Load
Echo, Hybrid Reverb
Set
  • Riddim 8-bar loop in Session view; vocal side finished and printed first
  • Echo send, 3/16, FB 55%, HP 250 Hz
  • Hybrid Reverb send, Spring, 2 s
  • Version vocal muted except single words thrown to the echo; instruments in and out as mutes
  • Mixdown one live pass to a Resampling track, riding sends, mutes and Echo feedback by hand
You should hear
A vocal song and its dub from the same riddim: parts dropping out, single words thrown into echo, all mixed in one pass.
Live edition
Suite

Why it matters. Documented: it is the first release on Burial Mix, the imprint Moritz von Oswald and Mark Ernestus used for Rhythm & Sound’s vocal records, and it pairs a vocal side, “Never Tell You”, with “Never Tell You Version” (Basic Channel catalogue). Paul St. Hilaire also sang on three of the Main Street “Round” 12"s, which pair club vocals with instrumentals and dubs in the same way (Main Street catalogue). The Quietus ties the later Rhythm & Sound and Burial Mix work to Basic Channel’s habit of refracting one track into many versions (The Quietus, 2018). This record is the model for vocal dub techno: a song and its dub on two sides of one disc.

Documented method (general, not specific to this record): in his Red Bull Music Academy lecture, von Oswald describes how a dub usually starts from a vocal track and moves to an instrumental, which is then developed into several versions. Mixes were often done live, with the instruments dubbed to stereo tape in the classic way, so they can’t be recalled afterwards. He says “Mango Drive” was mixed entirely live. For vocal sessions he recommends recording from the first moment and letting singers work over a running track (RBMA lecture: Moritz von Oswald). His advice to producers: “Don’t interfere until the person is ready.” Both men also founded the Dubplates & Mastering cutting room to control how their records were mastered (Wikipedia: Rhythm & Sound).

Signature moves → Live 12

  1. Record the singer over a running riddim, from the first second (Documented advice). Loop an 8-bar riddim in Session view and arm an audio track with Record Mode set to record into the Arrangement. Leave it rolling through every pass: in Live 12 each pass lands in its own take lane, so nothing is lost and the singer never waits for you. Comp afterwards from the take lanes.
  2. Build the vocal side first (Documented order). Finish the vocal mix, riddim and all, before touching the dub. Print it.
  3. Make the version from the same stems (Documented practice). Duplicate the tracks into a second arrangement section. Mute the vocal except for single words that you throw to the Dub Delay Send Rack, and take instruments in and out as mutes rather than edits. The method for throws is on Dub FX & Delays.
  4. Mix it live, to stereo, once (Documented). Create an audio track with input set to Resampling and record a full pass while you ride sends, mutes and Echo feedback by hand, using Push or a MIDI controller. Keep the take. Don’t fix it with automation, because the unrecallable pass is the point. The desk method is on Dub FX & Delays: desk approach.
  5. Soft skank and round sub (By ear). Use the genre page’s dub chord, played as an offbeat skank, and a sine sub that sits under the vocal without moving much.
If you record a guest vocalist, record and keep everything, but agree the use of the takes in writing. A dub version built from someone’s voice is still their performance.

What to steal: write the song, then perform the dub. Two sides from one session make a better version than a dub that has to invent its own material.

Monolake — “Lantau” (1997): a field trip worked into the loop

1997 · Chain Reaction CR-15, 12" “Lantau / Macao”; also on the CD Hongkong, CRD-04 (Documented: Chain Reaction catalogue, roberthenke.com) · BPM and key undocumented · Tools: the 1997 Monolake studio list (Documented) · Level: intermediate

Quick start

Load
Audio Effect Rack, Auto Filter, Redux, Hybrid Reverb, Delay
Set
  • Field recording long take warped in Texture mode, Grain Size around 60, Flux 20, a 4-bar window looped
  • Filterbank Audio Effect Rack with four parallel Auto Filter band-passes: 180 Hz, 450 Hz, 1.1 kHz and 2.8 kHz, Q around 4
  • Redux 12-bit, no downsampling
  • Hybrid Reverb Algorithm only, Dark Hall, 2–3 s, Diffusion high
  • Delay send, 1/4 + 3/16, FB 40%
You should hear
A street recording looped and filtered into tonal bands, with the grain of an early 1990s rack reverb.
Live edition
Suite

Why it matters. It shows the second Chain Reaction idea: dub techno as sound design rather than a dub desk. Documented: in 1996 Gerhard Behles and Robert Henke went to Hong Kong for the International Computer Music Conference, and Henke used the trip to make field recordings in and around the city and in Guangzhou. For the 1997 CD, which collected their first vinyl releases, they refined the original tracks with material from those recordings and called the album Hongkong. The record was produced by Behles and Henke between 1995 and 1997 (Monolake: Hongkong release notes). Which recordings ended up in “Lantau” specifically is not documented.

Documented tools (the 1997 studio, not a per-track list): SCI Prophet VS, Yamaha SY77 and TG-77, Ensoniq ASR-10 sampler, Alesis Quadraverb, Lexicon PCM 80, a Speck filterbank, Roland Juno-6, Waldorf Microwave, Boss RSD-10 delay, Notator on an Atari 1040, Max on a Macintosh, and PX18, a step sequencer Henke wrote in Max (roberthenke.com). On the next record, Gobi (1999), they granulated loops from Hongkong in Max and Reaktor and captured a two-hour jam to DAT (Monolake: Gobi release notes).

Signature moves → Live 12

  1. Record the place, then decide (Documented). Bring back long recordings, not one-shots. In Live, warp them in Texture mode (Grain Size around 60, Flux 20) so you can loop a 4-bar window of a street or market without the transients jumping.
  2. A step sequencer with its own parameters per step (Documented: PX18; recipe By ear). Program the percussion in a MIDI clip and give each step its own value by drawing stepped clip envelopes for Simpler’s Filter Freq and Sample Start. Set note Chance to 70–90% on the off-steps. That moves the sound per step the way a custom sequencer does.
  3. Filterbank colour (Documented: Speck filterbank; recipe By ear). Run a field-recording loop into an Audio Effect Rack with four parallel Auto Filter band-passes (for example 180 Hz, 450 Hz, 1.1 kHz and 2.8 kHz, Q around 4), one per chain, with a macro on each chain volume. Live 12’s Spectral Resonator in Chord mode is the one-device alternative.
  4. Hardware reverbs as character (Documented: Quadraverb, PCM 80; recipe By ear). Use Hybrid Reverb with the Algorithm side only, Shimmer or Dark Hall, 2–3 s, and Diffusion high, after a Redux at 12-bit with no downsampling. That approximates the grain of an early 1990s rack unit without pretending to be one.
  5. Sub and chord from the genre page. Keep the sub and chord plain: in this method the identity is in the recordings and the filtering.

What to steal: travel with a recorder and come back with a palette. Then let a filter, not an EQ, decide which part of the place belongs in the track.

Pole — “1” (1998): the defective filter as drum machine

27 July 1998 · Kiff SM, KIFF012CD (Documented: Discogs release 5680) · BPM and key undocumented · Tools: a defective Waldorf 4-Pole filter (Documented); everything else undocumented · Level: intermediate

Quick start

Load
Operator, Auto Filter, LFO, Gate, Hybrid Reverb
Set
  • Operator Osc A Noise, amp envelope A 0, D 6 ms, S 0, triggered by 1/16 notes at random velocities
  • Auto Filter Band-pass, Resonance 85–95%, Drive on
  • LFO S&H at 1/16 on Frequency; a second S&H at 0.7 Hz on Resonance
  • Gate threshold just above the noise level
  • Hybrid Reverb send, Dark Hall around 5 s, low cut at 500 Hz
You should hear
Clicks and pops ringing at a different pitch each time where hats would be, some dying, some ringing out into long reverb.
Live edition
Suite

Why it matters. It moved dub techno off the dance floor and into the headphones. The crackle that other producers hid became the rhythm section. Documented: Stefan Betke took the name Pole from a Waldorf 4-Pole resonant filter he had broken. Its glitching analog output was the basis of his first tracks, with dub basslines underneath and the filter supplying the percussion (Wikipedia: Pole). The German Wikipedia describes the fault as a characteristic crackle (Wikipedia (de): Pole). His own biography calls it the “now-legendary, defective Waldorf 4-Pole filter” and places the records in a dub tradition where reverbs and basslines mattered more than melody. Betke also worked as a mastering and cutting engineer at Dubplates & Mastering until 2000 (pole-music.com: biography). The credits list him as sole writer and producer (Discogs).

By ear: no kick drum in the techno sense. Clicks and pops from the filter sit where hats and percussion would be, a round dub bassline carries the groove, and long reverbs fill the rest. What exactly was wrong with the filter is not documented.

Signature moves → Live 12

  1. Make a crackle source that misbehaves (Documented idea; recipe By ear). On an audio track, run Operator with Osc A set to Noise and a very short amp envelope (A 0, D 6 ms, S 0) into Auto Filter: Band-pass, Resonance 85–95%, Drive on. Trigger it with a MIDI clip of 1/16 notes at random velocities. Add an LFO (Max for Live) in S&H mode at 1/16 on Auto Filter Frequency so every click rings at a different pitch.
  2. Make it unreliable (By ear). Put a Gate after the filter with its threshold just above the noise level, and a second S&H LFO at a slow, unsynced rate (0.7 Hz) on Auto Filter Resonance. Now some clicks ring and some die, and the pattern never quite repeats. That is the “broken” part.
  3. Arrange the crackle, don’t sprinkle it (By ear). Edit the click clip so the loudest pops fall where a hat or a rim would. Set note Chance to 60% on the rest. The listener hears a percussion part, not damage.
  4. Dub bassline under it (Documented: dub basslines). Use Drift with Osc 1 Triangle, the filter at 300 Hz, Glide 80 ms, and a 2-bar line that leaves space after each note. Add Saturator with Soft Sine at 3 dB so the third harmonic carries on small speakers.
  5. Reverb before melody (Documented priority). Send the crackle, not the bass, to Hybrid Reverb with a Dark Hall around 5 s and a low cut at 500 Hz in the reverb EQ. The space comes from the pops ringing out.
Very high resonance with Drive on can produce sudden peaks of +15 dB or more. Put a Limiter at −6 dB ceiling on the click track while you design it, and keep the monitor level down.

What to steal: before you throw away a faulty device or a bad recording, try using it as the drum machine.

Vladislav Delay — “Huone” (1999): 22 minutes of live reprocessing

June 1999 · Chain Reaction CR-26, 12"; collected on Multila, CRD-09, 2000 (Documented: Discogs release 11278, Discogs release 705) · BPM and key undocumented · Tools: undocumented · Level: advanced

Quick start

Load
Grain Delay, Auto Filter, Echo, Hybrid Reverb
Set
  • Group groove, bass and chord into a Resampling track, 20 minutes recorded in one go
  • Effects on a controller Grain Delay Spray 20 ms and Pitch, Auto Filter frequency, Echo feedback 40–85%
  • Bass four clips with Follow Actions: Next, Chance 70%, Other 30%
  • Hybrid Reverb on the chord track, 40–60% wet, Decay 6–8 s, Freeze on a macro
  • Kick low-passed to around 1.5 kHz, reverb sidechained only 2–3 dB
You should hear
A buried 4/4 pulse under a bass that changes shape every few bars, with reverb as part of the timbre rather than a send.
Live edition
Suite

Why it matters. It stretched the Chain Reaction 12" into something closer to a long improvisation. Sasu Ripatti later played drums and percussion in the Moritz von Oswald Trio, from 2009 to 2015 (Wikipedia). Documented: the Discogs credits list mastering and lacquer cutting at Dubplates & Mastering, by Moritz von Oswald (Discogs). Multila compiles the Huone and Ranta 12"s, and the title piece “Huone” runs about 22 minutes (Wikipedia: Multila). His general method is described as a mixture of synthesis, vocal recording and live reprocessing (Wikipedia: Vladislav Delay). A Boomkat review reproduced on his Bandcamp hears heavy reverb, cavernous production, and broken basslines that change every few bars (Bandcamp: Multila, a review).

By ear: the 4/4 is buried rather than missing. The bass changes shape every few bars instead of looping, and the reverb is part of the timbre, not a send that you notice.

Signature moves → Live 12

  1. Reprocess while recording (Documented method; recipe By ear). Put the groove, bass and chord tracks on a group, then route the group into an audio track set to Resampling and arm it. Record 20 minutes in one go while you play the effects: Grain Delay Spray and Pitch, Auto Filter frequency, and Echo feedback on the group, all on a controller.
  2. A bassline that keeps mutating (Documented: review; recipe By ear). Write four 2-bar variations of a simple bass figure in one MIDI clip, then use clip Follow Actions (Next, Chance 70%, Other 30%) on four clips so the order is never fixed. Live 12’s Shape and Stacks generators make quick variations.
  3. Reverb as timbre (Documented: review). Use Hybrid Reverb on the chord track, not a send, at 40–60% wet, Decay 6–8 s, with Freeze mapped to a macro. Freeze the tail under the next section so one passage turns into the room for the next.
  4. Kick buried behind the reverb (By ear). Use the genre page’s soft kick, then low-pass it further, to around 1.5 kHz, and sidechain the reverb track only lightly (2–3 dB) so the pulse is felt more than heard.

What to steal: treat effects as the instrument you improvise on, and record the whole improvisation. The editing comes afterwards, if at all.

Fluxion — “Vibrant Forms” (1999): repetition, then heavy live effects

1999 · Chain Reaction CRD-07, CD; Vibrant Forms II followed as CRD-11 in 2000 (Documented: Chain Reaction catalogue) · BPM and key undocumented · Tools: undocumented for this record · Level: intermediate

Quick start

Load
Analog, Auto Filter, Chorus-Ensemble, Echo, Hybrid Reverb
Set
  • Source a single Analog saw stab, 1- or 2-bar loop left unchanged
  • Auto Filter LP 24, 700 Hz, env 15%; cutoff macro 200 Hz → 4 kHz
  • Chorus-Ensemble Chorus, 0.3 Hz
  • Echo 1/4 + 3/16, FB 55%, HP 250 Hz; feedback macro 20 → 92%
  • Hybrid Reverb Dark Hall 4 s, 30%
You should hear
A one-bar loop that never changes its notes but keeps moving for minutes through filter, chorus, echo and reverb.
Live edition
Suite

Why it matters. Konstantinos Soublis showed that the Chain Reaction sound could be played as a live effects performance by one person, not only produced in a Berlin mastering room. Documented: his music is described as repetitive forms with heavy live effects processing, telling a spontaneous story, with sounds processed so heavily that the source hardly matters (Wikipedia: Fluxion). In 2016, announcing Vibrant Forms III, he described the three albums as one story, his studio as a mix of analog and digital machines whose interplay matters most, a lot of time spent experimenting without knowing what he would make, improvising as he does live, and mistakes that led to breakthroughs (RA news, 2016, retrospective). He calls each work “an aural experiment”.

Signature moves → Live 12

  1. Source is secondary (Documented). Start from anything: a single Analog saw stab, a sample, a click. Don’t spend time on the source patch. All the design goes into the chain after it.
  2. A performance rack, not a preset (Documented: live effects; recipe By ear). Build an Audio Effect Rack on the chord with Auto Filter, Chorus-Ensemble, Echo and Hybrid Reverb in series, and map eight macros as below. Save Macro Variations for “dry”, “half” and “flooded” so you can jump between states mid-take.
  3. Record improvised passes, keep the mistakes (Documented). Arm a Resampling track and record ten-minute takes. Mark any accident you like with a locator and build the next take around it.
  4. Repetition holds it together (Documented: repetitive forms). Keep the MIDI loop 1 or 2 bars long and unchanged for minutes. The movement comes only from the effects.

What to steal: make the effects chain the instrument and practise it like one. A 1-bar loop is enough material for ten minutes if the hands keep moving.

Scion — “Arrange and Process Basic Channel Tracks” (2002): versioning someone else’s catalogue

1 June 2002 · Tresor, Tresor 200 CD (Documented: MusicBrainz, Wikipedia: Basic Channel) · BPM: follows the source tracks · Tools: undocumented · Level: intermediate

Quick start

Load
EQ Three
Set
  • Sources whole stereo tracks, Beats mode with Preserve Transients for drums, Complex for washes
  • Loops the most stripped-back 8 bars of each, run three or four times their length
  • EQ Three kill the outgoing low band exactly where the incoming one comes in
  • Dub return one shared delay return fed from both sources, ridden only in the transitions
You should hear
Long blends of two sources stripped back to chord, delay and kick, with familiar sections looped into something hypnotic.
Live edition
Intro and up

Why it matters. Scion, the duo FACT names as DJ Pete and René Löwe (better known as Substance and Vainqueur, which is scene knowledge here), made the first Chain Reaction release, “Emerge”, CR-01 in 1995 (Chain Reaction catalogue). Seven years later they took Basic Channel’s own records and re-edited them into one continuous album: dub versioning applied to a techno label’s back catalogue. Documented: FACT describes it as re-edits of Basic Channel productions by Moritz von Oswald and Mark Ernestus. The edits are subtle and bring out the tracks’ dub roots, and the article calls it “as much a DJ mix as a remix album” (FACT: the remix album).

By ear: long blends where two sources play together for minutes, elements stripped back until only chord, delay and kick remain, and sections looped beyond their original length.

Signature moves → Live 12

  1. Re-edit, don’t remix (Documented). Use whole stereo tracks, not stems. Warp each one carefully: set the first downbeat, use Beats mode with Preserve Transients for the drums or Complex for the washes, and set the project to the original tempo so nothing is stretched much.
  2. Long loops of the best 8 bars (By ear). Find the most stripped-back 8 bars in each source, loop them in Arrangement and let them run three or four times their original length. The audience hears something familiar becoming hypnotic.
  3. Blend with EQ, like a DJ (Documented: DJ-mix form; recipe By ear). Put EQ Three on each source and swap the lows at the phrase point: kill the outgoing track’s low band exactly where the incoming one’s low comes in. Never let two kicks or two subs run at full level together.
  4. Add only dub (Documented: stresses dub roots). The only new processing is a shared dub delay return fed from both sources during blends. Ride it in the transitions, not in the middle of a passage.
Re-edits of other people’s records need a licence before release. Clear the rights with the owners before you release anything like this.

What to steal: the most dub thing you can do to a track is to version it. Long loops, EQ swaps and one shared echo are enough.

Deadbeat — “New World Observer” (2005): echo as the lead voice, all in the computer

29 March 2005 · ~scape, sc27cd (Documented: MusicBrainz, Exclaim!) · BPM and key undocumented · Tools: computer only, software instruments and samples (Documented); Ableton Live (Documented for his setup around 2004) · Level: intermediate

Quick start

Load
Simpler, Echo, Shifter, Auto Filter, Hybrid Reverb
Set
  • Simpler Slice mode by Transient, high sensitivity, fragments of 20–80 ms, Fade 2–5 ms
  • Echo send, 3/16, Feedback around 70%, time automated in steps (1/8, 3/16, 1/4)
  • Shifter in the return so the repeats change pitch
  • Auto Filter HP 300 Hz
  • Hybrid Reverb Spring, 1.5 s
You should hear
A riddim of tiny sample shards, with a single shard's echo stepping through times to play the melody.
Live edition
Suite

Why it matters. Scott Monteith’s album on Pole’s ~scape label put dub techno back in touch with roots dub and with politics, and it did so with no hardware at all. Documented: he composes his tracks entirely on the computer, with software instruments and samples. “Abu Ghraib” is built around a degraded recording of a talk-radio host, cut off by a spoken rebuke, and the singer Athesia appears on two tracks. He said working with a singer was harder than working with samples (Georgia Straight, 2005). Exclaim!’s review hears microsampling, with the echo used as a lead voice in the manner of King Tubby (Exclaim!, a review), and XLR8R’s review praises its precisely placed, heavily manipulated samples (XLR8R, a review). In an interview from the previous album’s era, he says he uses Ableton Live exclusively, in the studio and on stage, and builds instruments in Tassman and Reaktor (Igloo Magazine). That interview describes his setup around 2004, not this album specifically.

Signature moves → Live 12

  1. Microsamples as the riddim (Documented: review; recipe By ear). Load a short loop into Simpler in Slice mode, slice by Transient at high sensitivity, and keep only fragments of 20–80 ms. Program a skank and a percussion part from these shards. Set each slice’s Fade to 2–5 ms so the cuts don’t click.
  2. The echo plays the melody (Documented: review). Send a single short shard to Echo set to 3/16 with Feedback around 70%, then automate the Echo time in steps (1/8, 3/16, 1/4) so the repeats outline a phrase. The Shifter dub pitch tricks work well here: put Shifter in the return so the repeats change pitch.
  3. A found voice as a political statement (Documented). Use a spoken recording degraded with Redux (8-bit, downsample 4) and a band-pass at 1–3 kHz, then cut it off hard with a clip edit. The interruption is the point. Only use speech you are legally allowed to use.
  4. Everything inside the computer (Documented). No external gear: build any extra instruments in Live’s own environment, with Operator, Meld or Max for Live devices, as he built his in software.

What to steal: let the delay write the melody. One tiny sample and an echo whose time you change by hand make a lead line.

Yagya — “Rigning” (2009): one long track, ten movements, rain on a city

2009 · Sending Orbs SO 011; reissued 25 May 2018 by Delsin, DSR/X14 (Documented: MusicBrainz, Bandcamp: Rigning) · BPM and key undocumented · Tools: undocumented · Level: intermediate

Quick start

Load
Operator, Auto Filter, Vinyl Distortion, Compressor, Saturator
Set
  • Operator Osc A Noise, sustained
  • Auto Filter band-pass at 3–5 kHz, LFO 0.05 Hz
  • Vinyl Distortion Crackle at a low Density
  • Compressor sidechained from the kick, 1–2 dB
  • Saturator on the master or a group, Analog Clip at 2–3 dB
You should hear
A moving rain layer breathing lightly with the kick, under a warm fuzz that covers everything.
Live edition
Suite

Why it matters. It is the gentle, almost ambient end of dub techno. It proves that the genre’s parts (soft kick, sustained chord, endless echo) can carry an hour-long narrative, not only a 12" groove. Documented: the Icelandic producer Aðalsteinn Guðmundsson’s album is described by its label text as “essentially one long track split into movements”, depicting a city asleep while gentle rain falls. Its ten tracks are simply numbered, from Rigning Einn to Rigning Tíu, and the sound is described as wrapped in a warm fuzz (Bandcamp: Rigning, label text). Rigning is Icelandic for rain. The gear and recording method are undocumented.

Signature moves → Live 12

  1. Write one piece, then cut it into movements (Documented concept; recipe By ear). Build a single long Arrangement of around 60 minutes with one tempo and one key centre. Mark the movement boundaries with locators and export each part separately. Each movement brings in or removes one element, so the whole keeps one identity.
  2. A rain layer that moves with the track (Documented theme; recipe By ear). Record rain if you can. Otherwise use Operator with Osc A Noise into Auto Filter band-pass at 3–5 kHz, plus Vinyl Distortion Crackle at a low Density for drops. Sidechain it very lightly from the kick (1–2 dB) so it breathes with the groove.
  3. Warm fuzz over everything (Documented description; recipe By ear). On the master or a group, use Saturator in Analog Clip mode at 2–3 dB, then a gentle high shelf of −2 dB at 8 kHz with EQ Eight. See Tape, Lo-fi & Analog Character for more tape colour options.
  4. Pads that rise rather than chords that stab (Documented: the label text describes upward-reaching pads; recipe By ear). Use the genre page’s drone pad with a slow Auto Filter opening over 32 bars per movement.

What to steal: give an album a single image and a single long timeline, and cut it into parts at the end, not the beginning.

Deepchord presents Echospace — “Liumin” (2010): Tokyo, and a reduced second disc

June 2010 · Modern Love, LOVE064, 2×CD (Documented: Discogs release 2328489, MusicBrainz) · BPM and key undocumented · Tools: undocumented for this record · Level: advanced

Quick start

Load
Auto Filter, Saturator, Echo, EQ Eight
Set
  • City recording Texture mode, −12 st, low-passed at 2 kHz
  • Saturator Analog Clip at 3 dB
  • Echo Noise 10%, Wobble 15%, 100% wet, very short time
  • EQ Eight high shelf −3 dB at 10 kHz; commit with Bounce to New Track
You should hear
Traffic and voices turned into a tonal wash under the chord, printed through a tape-like chain.
Live edition
Suite

Why it matters. It is the Echospace album after The Coldest Season (which is on the first page), and it adds the city-recording idea to that tape-delay sound. Documented: the sleeve notes say the album contains sound phenomena “culled from the aether above Tokyo”, recorded in Echospace in 2009–10. Rod Modell and Stephen Hitchell are credited as performers, and Modell took the photographs. Lupo mastered it at Dubplates & Mastering. The second disc, Liumin Reduced, is described as an atmospheric distillation of sounds taken from the first disc, and was included only with the first pressing (Discogs). Hitchell’s Discogs profile says he produces and records onto reel-to-reel tape, though it does not name a record (Discogs: Stephen Hitchell). How the Tokyo material was recorded is not documented.

Signature moves → Live 12

  1. A city as the source (Documented: Tokyo material). Treat a city recording as the pad: warp it in Texture mode, transpose it −12 st, and low-pass it at 2 kHz so traffic and voices turn into a tonal wash under the chord. The street-recording rack in the Hash-Bar Loops section works here too.
  2. Print to tape, or fake it honestly (Documented for Hitchell in general; recipe By ear). Print each stem through a tape-like chain: Saturator in Analog Clip mode at 3 dB, Echo with Noise at 10% and Wobble at 15% at 100% wet but with a very short time, and an EQ Eight high shelf of −3 dB at 10 kHz. Commit with Bounce to New Track and edit the printed audio, not the MIDI.
  3. Make the reduced disc from the main disc (Documented). When the album is done, resample the finished tracks and remove the drums. Use Spectral Time with Freeze on and Hybrid Reverb at 100% wet, then stitch the results into long, beatless pieces. It is a distillation of material you already have, not new composition.
  4. Two performers (Documented credits). As with Porter Ricks, split the roles: one plays the sources and the other plays the delay returns and the tape chain.

What to steal: finish the album, then make a second one from its own sound with the drums taken out.

Deepchord — “Hash-Bar Loops” (2011): a hidden recorder and synths as triggers

27 June 2011 · Soma, SOMACD091 (Documented: MusicBrainz, Wikipedia: Deepchord) · BPM and key undocumented · Tools: see the interview below (Documented, general) · Level: advanced

Quick start

Load
Operator, Echo, Auto Filter, Chorus-Ensemble, Hybrid Reverb
Set
  • Operator a plain sine stab, used only as a trigger
  • Echo 3/16, FB 70%, Noise 15%, Wobble 15%
  • Auto Filter LP 1.8 kHz, slow LFO
  • Chorus-Ensemble Chorus 0.2 Hz, then Hybrid Reverb Dark Hall 6 s, 35%
  • Street take a section with an event in it, Texture mode, looped at −18 to −24 dB
You should hear
A boring sine stab turned into a smeared, wobbling chord by the effects chain, over a night street recording.
Live edition
Suite

Why it matters. It is Rod Modell’s clearest statement of dub techno as a portrait of a place: Amsterdam, at night. Documented: in an interview with Inverted Audio, Modell says most of the album was recorded in a small apartment near Amsterdam’s centre and in its streets, working mainly at night. For field recordings he hides a small WAV recorder, sits about 50 yards away and reviews the result later. He wants recordings of something significant, not “field recordings for field recording sakes”. He sees himself as a conduit that turns the essence of a place into sound, and says good headphones are enough, with no need for a formal studio (Inverted Audio: Deepchord).

Documented tools (his setup in general): past key gear includes Linn and Forat 9000 drum machines, SCI Studio 440s, Prophet VS units, an Eventide Orville, and Publison and Korg delay units, most of which he has sold. Synths now serve mainly as triggers for his effects chains, so the source sound matters less. The computer plays back samples of his own hardware and replaces effects units and recorders, and he says he has never used a computer to make a sound (Inverted Audio).

Signature moves → Live 12

  1. Hide the recorder, then walk away (Documented). Record long takes without being present, so people and traffic behave naturally. In Live, keep the whole take and pick the section that has an event in it: a voice, a tram, a door. Warp in Texture mode and loop it under the groove at −18 to −24 dB.
  2. Synth as trigger, chain as instrument (Documented). Play a plain Operator sine or a short Analog pulse, then do all the design in a long effects chain: Echo with Noise and Wobble, Auto Filter, Chorus-Ensemble, and Hybrid Reverb. The source patch should be almost boring. See tape delay emulation with Echo for the delay settings.
  3. Sample your own hardware (Documented). If you own a synth, record multisampled notes into Sampler (or Drum Sampler for hits) and play those, rather than a soft synth. The computer replays real gear and replaces the recorder and effects units, as in his setup.
  4. Mix on headphones, at night (Documented working conditions). Check the sub on speakers or with a Spectrum analyser afterwards, because headphones mislead below 60 Hz. See Mixing for sound systems.

What to steal: record something significant rather than ambience for its own sake, and spend your sound design budget on the effects chain, not the synth.

Topdown Dialectic — “Topdown Dialectic” (2013): a process instead of a person

2013 (Tailings); vinyl reissue 2 December 2022 · Peak Oil, PEAK15 (Documented: Bandcamp: S/T 2013, MusicBrainz) · BPM and key undocumented · Tools: laptop (Documented, label text) · Level: intermediate

Quick start

Load
Grain Delay, Spectral Time, Hybrid Reverb
Set
  • Source one 2-bar loop of chord and pulse, resampled three times
  • Prints Grain Delay Spray 40 ms; Spectral Time Freeze, Spray; Hybrid Reverb 8 s, 60%
  • Clip lengths 2, 2.25 and 1.75 bars so the prints drift and rarely realign
  • Pulse a low-passed tom or filtered click at −12 dB below the texture
You should hear
Several layers of the same material slightly out of phase, with very low dynamic contrast and a muted pulse instead of a kick.
Live edition
Suite

Why it matters. By ear, it is dub techno’s fog with every personal story removed, and it is one of the reference points for the genre’s ambient end in the 2010s (scene view). Documented: the Peak Oil text says the debut first appeared as the last release on the short-lived American imprint Tailings. It describes music built from layered laptop textures, quiet rhythms and distant, glitchy systems, and says the project presents itself as a process rather than a person, so the listener focuses on the sound alone (Peak Oil). The eight tracks are titled only by side and position, A1 to B4. Peak Oil also released a 2018 Topdown Dialectic album, PEAK09, and Vol. 3, PEAK13, in 2021 (MusicBrainz).

By ear: short pieces of 4–7 minutes, very low dynamic contrast, a muted pulse rather than a kick, and loops that seem to be several layers of the same material slightly out of phase.

Signature moves → Live 12

  1. Laptop only, stock devices (Documented: laptop textures). Start from one 2-bar loop of chord and pulse. Resample it three times through different chains (Grain Delay, Spectral Time, Hybrid Reverb) and stack the three prints at different levels.
  2. Layers slightly out of phase (By ear). Set the resampled clips to lengths of 2, 2.25 and 1.75 bars so they drift against each other and realign rarely. The polymeter method is on Ableton’s Making Music: asynchronous loops.
  3. Quiet rhythms (Documented). Replace the kick with a low-passed tom or a filtered click at −12 dB below the texture. The rhythm is implied, not driven.
  4. Name nothing (Documented stance). Number the tracks by position. This is a working method too: without a title to live up to, you stop decorating.

What to steal: remove yourself from the music. A loop, a few resampling passes and numbered tracks can be a whole aesthetic.

Porter Ricks — “Anguilla Electrica” (2017): the reunion on self-written software

22 June 2017 · Tresor, Tresor 295 (Documented: Tresor, MusicBrainz) · BPM and key undocumented · Tools: software written by Andy Mellwig (Documented, 2014–15 interviews) · Level: advanced

Quick start

Load
Corpus, Resonators, Echo, Glue Compressor, Limiter
Set
  • Core one loop, a kick with its own resonant tail
  • Corpus Tube, low decay; then Resonators tuned to the root
  • Echo dark, FB 65%
  • LFOs slow, at non-integer rates (0.013, 0.021, 0.034 Hz) on filter, feedback and send levels
  • Master Glue Compressor threshold performed by the second pilot, Limiter −1 dB
You should hear
A single core loop that keeps growing and developing on its own, steered and mixed in one stereo pass.
Live edition
Suite

Why it matters. It is their first album since Porter Ricks: Techno Animal Symbiotics in 1999, after the Shadow Boat EP in 2016 (Wikipedia: Porter Ricks), and it shows how Thomas Köner and Andy Mellwig rebuilt their submerged dub techno for performance. (Biokinetics is on the first page.) Documented, before the album: in 2014 Köner told Time Out that Mellwig writes the duo’s software, which Köner calls very creative and intuitive, and that they were developing it to generate, mix and master music live. Their sets relied heavily on improvisation, because they don’t want things to be predictable (Time Out New York, 2014). In 2015 Köner told Crack that a piece starts from a very central core element that grows as the system runs, that the two of them are like two pilots steering one boat, and that the underwater association was never intended (Crack Magazine). These interviews describe the reunion method. They do not say which tracks on the album were made this way.

Signature moves → Live 12

  1. One core element that grows (Documented). Start from a single loop: a kick with its own resonant tail, or a chord through a dark delay. Everything else is generated from it. Use MIDI generators (Rhythm, Seed) on derived parts, and make audio layers by resampling the core through Corpus and Resonators.
  2. Let the system run (Documented). Put slow Max for Live LFOs, at non-integer rates (0.013, 0.021, 0.034 Hz), on filter, feedback and send levels so the piece keeps developing without hands. You only steer.
  3. Two pilots (Documented). Split control: one person, or one controller, owns the generator and source parameters, the other owns mixing and mastering (sends, EQ Three, Glue Compressor threshold). Record the session as one stereo pass.
  4. Master as part of the performance (Documented aim). Put Glue Compressor and Limiter on the master and map Glue’s Threshold and Makeup to the second pilot’s controller. Mixing and mastering decisions then happen in time, like a dub desk.

What to steal: build a system that keeps changing on its own, then perform with someone else on top of it. Split the roles: one of you makes the sound and the other mixes it.

Patterns across the dub lane

Read the twelve together and a few decisions keep coming back. They matter more than any piece of gear.

  1. The mix is performed, then kept. Von Oswald’s live mixes to stereo tape, Fluxion’s live effects processing, Delay’s live reprocessing and Porter Ricks’ software for mixing and mastering in performance are all Documented. In Live this always means the same move: a Resampling track, one pass, kept as it is.
  2. Versioning over composing. Burial Mix pairs each vocal with its version, Main Street pairs club vocals with dubs, and Scion turned a whole catalogue into one mix: all Documented. The dub is made from material that already exists.
  3. A place, recorded. Henke’s Hong Kong and Guangzhou recordings and Modell’s hidden recorder in Amsterdam are Documented. Yagya’s rain is a Documented theme whose sound source is not documented.
  4. The source is secondary, the chain is primary. Deadbeat’s microsamples led by echo (a review’s description), Fluxion’s processing that makes the source hardly matter, Modell’s synths as triggers for effects chains and Pole’s defective filter are all Documented. Design the chain, not the patch.
  5. The cutting room is part of the sound. Dubplates & Mastering cut Huone and mastered Liumin, and Betke worked there until 2000, all Documented. The label family shared one mastering ear, which partly explains the shared sound: dark tops, mono low end, and quiet masters with plenty of headroom.
  6. Anonymity and process as aesthetics. Topdown Dialectic’s “process rather than a person” and Yagya’s numbered movements are Documented. The genre’s refusal to explain itself is part of how it sounds.
ApproachRecordsLive 12 core
Live mix to stereoNever Tell You, Huone, Vibrant Forms, Anguilla ElectricaResampling track, Macro Variations, Push or a controller
VersioningNever Tell You, Arrange and ProcessDuplicate arrangement, mutes, EQ Three swaps, shared dub return
Field recordingLantau, Liumin, Hash-Bar Loops, Rigning (theme)Texture warp, band-pass rack, Echo on a band of the recording
Faulty or trivial source, rich chain1, New World Observer, Vibrant Forms, Hash-Bar LoopsResonant Auto Filter with S&H, long effect chains, Limiter as safety net
Long formHuone, Rigning, Arrange and Process, Liumin ReducedFollow Actions, locators, one-pass recording, Hybrid Reverb Freeze
Records that were considered but left out for lack of public documentation about how they were made: Vainqueur’s “Elevation” (Chain Reaction CR-07, 1996), Round One’s “I’m Your Brother” (Main Street MSR-02, 1994), Andy Stott’s Merciless (Modern Love, Love024, 2006: the credits show only that Rashad Becker cut it at Dubplates & Mastering, per Discogs), Intrusion’s The Seduction of Silence (echospace [detroit], 2009, with Paul St. Hilaire on two tracks, per Discogs), Quantec, Bvdub and Mike Huckaby’s dub-leaning work. Catalogue details for the first two are on the Basic Channel catalogue pages.

Related