Acid / 303

TD-3 via External Instrument: MIDI out, audio in, latency compensated

The simplest way to play the TD-3 from Live clips as if it were a plug-in: one device sends MIDI out and brings the audio back on the same track, with delay compensation.

1. Connect the TD-3's MIDI In to your interface (DIN or USB; the TD-3 is a USB MIDI class device). Audio out to an interface input. In Live's preferences, enable Track and Sync for the TD-3's MIDI output port.
2. Drop External Instrument on a MIDI track. MIDI To: the TD-3 port, channel 1 (the TD-3's default). Audio From: the interface input. Gain to taste.
3. Set Hardware Latency: play a tight 1/16 pattern alongside a Live drum rack, record the TD-3 to an audio track, measure the offset in the clip and dial it in (typically 1–5 ms, negative or positive depending on your interface). Live 12's track delay is a fallback.
4. The TD-3 responds to note velocity for accent: velocity above its threshold (around 100 in practice) triggers an accent. Slide is triggered by overlapping notes (legato). Draw the clip with overlapping notes for slides and velocity 127 for accents.
5. Audio effects after External Instrument process the TD-3's audio — Overdrive, Pedal, Roar and delays all work as if it were a software synth.

MIDI clock sync, CV Tools gate/CV, and recording the TD-3

Running the TD-3's own sequencer from Live clock gives the authentic slide and accent behaviour. For eurorack owners, CV Tools can also drive the TD-3's CV/Gate inputs directly from a Live clip, with sub-millisecond timing through a DC-coupled interface.

1. Clock: Preferences → Link/Tempo/MIDI, set Sync on for the TD-3's MIDI output port. On the TD-3 set sync mode to MIDI. Press play in Live; the TD-3 pattern runs in lock. Use the MIDI clock sync delay setting in the same preference pane to nudge for the TD-3's start latency.
2. Record: audio track, input from the interface channel, monitor Off, arm, record 8 or 16 bars. Set the clip to Warp off after recording (it is already in tempo) and trim to the bar.
3. CV option: with a DC-coupled interface (or a eurorack interface module), use the CV Instrument device: it converts a MIDI clip into a 1 V/oct pitch CV and a gate on two outputs. Patch those into the TD-3's CV In and Gate In. Calibrate with CV Instrument's calibration routine against the TD-3's output so the pitch tracks across octaves.
4. CV Instrument's gate output drives the TD-3's gate; slides via CV are set by the TD-3's slide behaviour on legato gates. Accents via CV are not available on the TD-3's rear panel — use MIDI for accented lines or layer accent hits from a second pass.
5. CV Clock Out can replace MIDI clock for the TD-3's sync input if you prefer a hard analog clock from the same device that clocks the modular.

Audio effect chains for the TD-3: Overdrive, Pedal, Roar, Echo, Auto Filter

The TD-3's raw output is thin and clean. The classic recordings are the box through a guitar distortion pedal into a tape delay; here are the Live equivalents and the order that works.

1. Overdrive first: its bandpass filter before the clipping (set the band to 500 Hz – 5 kHz) focuses the drive on the resonance peak so the squelch distorts and the fundamental stays cleaner. Drive 35%, Dynamics 60% keeps the accents louder than non-accents.
2. Pedal in Overdrive type (Distortion and Fuzz types for techno): Gain around 4, Bass −1, Treble +2, Sub switch on to keep the low end from thinning out. This is the "guitar pedal" stage.
3. Auto Filter with its envelope follower (the Envelope section in the 12.2 redesign) at 30%: the cutoff opens on accents, a second layer of squelch that tracks the hardware's own dynamics.
4. Roar Single, Diode 4 dB — a last asymmetric stage for even harmonics — with post LP 8 kHz so the fizz does not fight the hats.
5. EQ Eight: HP 45 Hz; a small dip around 800 Hz if the box sounds nasal through all that distortion.
6. Send to a Return with Echo: time 3/16 dotted (left) and 1/4 (right) for stereo, Feedback 45%, the Echo filter HP 300 Hz / LP 4 kHz so repeats are darker than the dry line, Modulation a touch for tape wobble.

In-the-box 303 with Analog: accent via velocity, slide via Glide

Analog is the closest native fit: a saw or square, a resonant filter with its own envelope, velocity routing to filter envelope depth, and Mono/Legato/Glide for slides.

1. Analog: Osc 1 Saw (or Square — the 303 has both). Osc 2 off. Noise off.
2. Filter 1: LP 12 (closer to the 18 dB original than LP 24). Cutoff 300 Hz as the "floor". Resonance 55%.
3. Filter envelope: Attack 0, Decay 250 ms, Sustain 0, Release 100 ms. Envelope amount to cutoff around 50%. This decay is your main "Env Mod/Decay" knob — map it to a macro.
4. Velocity: Analog has velocity-to-filter-envelope amount (the Vel control in the filter envelope section) — set it to 60%. Also velocity to amp level around 40%. Now velocity 127 = accent, velocity 90 = normal note.
5. Amp envelope: A 1 ms, D 300 ms, S 60%, R 80 ms. Mono voice mode, Legato on, Glide 50–80 ms. Overlapping notes slide without retriggering envelopes; separated notes retrigger.
6. Saturator Analog Clip 10 dB, then Overdrive 25% with its band at 400 Hz–4 kHz. EQ HP 40 Hz.

Drift 303: Shark Tooth, MS2 filter, Legato and Glide

Drift's MS2 filter has a slightly rounder resonance than Analog, and its Shark Tooth shape sits between saw and square — an in-between tone that reads as a "modded" 303.

1. Drift: Osc 1 Shark Tooth (or Saw; try Rectangle for the square flavour). Osc 2 off. Filter type MS2, LP, cutoff 250 Hz, resonance 60%.
2. Envelope 1 (filter env): A 0, D 200 ms, S 0, R 50 ms, amount to cutoff 55%. Use the mod matrix: Velocity → Env 1 amount (the "Vel" entry) 70%, Velocity → Volume 30%.
3. Voice mode Mono, Legato on, Glide 50 ms. Drift's Drift amount at 10% for slight analogue pitch wander — this is one thing it does better than Analog.
4. Mod matrix extra: Env 1 → Filter resonance +15% so the resonance spikes with the envelope on accents, approximating the original's accent resonance behaviour.
5. Roar Single Soft Sine 8 dB, pre filter HP 100 Hz on the distortion path so the fundamental stays clean while the squelch grinds.

Operator 303: saw with a 24 dB filter and a pitch envelope slide

Operator's filter section has its own envelope and the 24 dB LP with high resonance is aggressive — the "big room acid techno" sound rather than the vintage one. Its per-oscillator pitch envelope can fake slides even in polyphonic contexts.

1. Operator: Osc A Saw only, level 0 dB, other oscillators off. Osc A envelope: A 0, D 300 ms, S 50%, R 60 ms.
2. Filter section: LP 24, cutoff 250 Hz, Resonance 70%. Filter envelope: A 0, D 180 ms, S 0, R 80 ms, amount to cutoff 50%. Velocity to filter envelope 50% (the "Vel" in the filter's envelope). The filter drive at a low setting warms it before the external distortion.
3. Global: Voices 1 (mono), Glide on at 40 ms with the Legato-style behaviour — in Operator, glide between overlapping notes gives the slide; retriggered notes restart envelopes.
4. Alternative slide trick for poly contexts: Osc A pitch envelope −2 semitones with Decay 60 ms, so every note swoops in from below. Not a true slide but reads as "acid movement" in a chord-acid or layered context.
5. Pedal, Distortion type, Gain 5, then Auto Filter LP 12 at 6 kHz with the envelope follower set to open on transients — a de-harshing filter that still moves.

Sequencing accents and slides with Live 12 MIDI tools: Seed, Rhythm, Velocity Shaper, Ornament

Acid lines are generative by nature — the original was programmed step by step with accidental results. Live 12's MIDI tools reproduce that: generate the rhythm, generate the notes within a scale, shape the velocities into accents, and add slides by overlapping notes.

1. In a 1-bar MIDI clip, open the MIDI Tools tab. Generate with Rhythm: grid 1/16, density around 70%, accent pattern off for now, note length 1/16. This writes the step pattern.
2. Transform with Seed (or generate with it on an empty clip): pitch range A1 to A3, Scale on (A minor or A phrygian), weight the root and the octave. Seed's random seed value is an integer you can change for a new line with the same settings — write it in the clip name if you like the result.
3. Transform with Velocity Shaper: draw a shape with peaks on steps 1, 4, 7, 11 and 14 (3+3+4+3+3 style grouping), max 127, min 85. These peaks become accents in Analog/Drift (velocity → filter env) and on the TD-3 (velocity above threshold).
4. Slides: select a few notes that lead into a step above or below and extend their length to overlap the next note by a 1/32 — in Mono + Legato mode this is a slide. The Ornament transformation can add short grace notes before selected notes; set it to a very short duration one semitone below for a "flick" into accents, used sparingly.
5. Quantize transformation last, 1/16, to clean up anything the overlaps shifted. Then Time Warp on a copy to stretch the last bar for a half-time break.
6. Apply the Live 12 clip Scale mode: set the clip to A minor (or match the set's scale); the piano roll highlights in-scale notes and Seed/Chord follow it. Fold to scale when you edit by hand.

Filter envelope tricks and the 18 dB vs 24 dB question

Beyond decay time: envelope shaping, a second filter stage, and where to put resonance to get the 303's particular squelch from Live's 12 and 24 dB filters.

1. Two 12 dB filters in series ≈ 24 dB with two resonance points. In Wavetable: Filter 1 MS2 LP 12, res 50%; Filter 2 OSR LP 12, res 20%, cutoff 20% higher than Filter 1. Routing Serial. Env 2 → both cutoffs (60% and 40%). The resonances sit at slightly different frequencies — a wider, more "diode ladder" peak than a single 24 dB filter.
2. Envelope curve. The 303's filter envelope is roughly exponential. In Analog the envelope slopes are fixed, but in Wavetable's Env 2 the slopes can be dragged curved: pull the decay into a fast-then-slow curve for the authentic "yow" rather than a linear "yaw".
3. Accent decay shortening. Map velocity negatively to Env 2 decay time (in Wavetable's matrix: Velocity → Env 2 Decay, −30%). Accented notes get a shorter, snappier sweep — exactly what the hardware does.
4. Resonance into distortion. Saturator Soft Sine 6 dB right after the synth. Resonance peaks that would whistle instead compress and sing.
5. Post-synth Auto Filter with its own envelope follower at 40% adds a second, audio-driven squelch that reacts to the accent loudness — more contrast between accented and plain notes.

Roar as a 303 distortion

Roar's Serial and Feedback routings, with an envelope follower on the stage amounts, give the classic "distortion that reacts to accents" more controllably than a pedal chain.

1. Roar, Serial routing. Stage 1: Diode, 8 dB (asymmetric, "transistor pedal"). Stage 2: Soft Sine, 5 dB (rounds the Diode's harshness). Tone/colour controls neutral.
2. Pre filter: HP 120 Hz so the fundamental passes cleaner (Roar's pre-filter affects only the distortion path's emphasis when set this way — if the dry fundamental vanishes, lower the HP or raise Dry/Wet mix balance).
3. Post filter: LP 7 kHz, resonance small. Modulation matrix: Envelope Follower → Stage 1 amount +4 dB (accents distort more), LFO (1/4 synced, triangle) → Post LP frequency ±10% for a subtle second movement.
4. For hard acid: switch to Feedback routing, Feedback amount 30%, feedback pitch to the root, Fold shaper — the line gets a sustained ringing under the squelch. Limiter after.
5. Glue Compressor after Roar, 4:1, Attack 1 ms, Release 100 ms, 4 dB of reduction. This is the "consecutive accent build" emulation: accented runs push into the compressor and the resonance peak comes forward.

Dub delay sends and acid dub throws

Acid lines without delay sound like a demo. A dotted-eighth stereo delay on a return, filtered and saturated, plus a second return for "throws" — momentary send automation that spins a single note into a tail.

1. Return A — the always-on delay. Echo: Left 3/16, Right 1/4 (unlinked for stereo), Feedback 50%, Echo's filter HP 250 Hz / LP 3.5 kHz, Modulation 15% for tape wobble, Noise 10% for hiss, Reverb section 10% for glue. Then a Saturator at 4 dB so repeats get grittier as they fade. Send level from the acid track: −12 dB, constant.
2. Return B — the throw delay. Delay in ping-pong, 1/4, Feedback 80% (near runaway), filter on, LP 2 kHz. Send level from the acid track normally −inf; automate it to 0 dB for a single 1/16 note at the end of a phrase, then back down. The note spins out for bars.
3. Auto Filter on Return B with LFO at 1 bar, LP, so the throw tail sweeps as it decays.
4. Sidechain both returns lightly from the kick (Compressor, 3 dB) so the delay tails pump with the beat — the dub techno trick applied to acid.
5. For performance, map Return B's send to a Push encoder and a footswitch-style momentary MIDI mapping: hold = throw.

Full dub delay design is on dub FX & delays and the dub delay send rack.

Distortion stacks for hard acid and acid breaks

When one pedal is not enough: stacked stages, each gentle, each a different kind. Amp/Cabinet for the "303 through a guitar amp" rave sound, Erosion and Redux for digital 90s grit.

1. Saturator Analog Clip 6 dB — a consistent first stage.
2. Amp in Clean or Boost mode (Heavy for rave stabs), Gain 5, Mid 7. Cabinet 4×12 with the dynamic mic off-axis to tame fizz. The amp's tone stack scoops and pushes the mids exactly where the resonance lives.
3. Erosion, Wide Noise mode, frequency 1.5 kHz, amount 20%: adds a modulated hiss that reads as "aliasing" — the digital-era acid sound.
4. Redux: 12 bit, downsample 2, for a little grit without obvious crushing. Pre-filter on so the top does not alias harshly.
5. EQ and Glue. The total: five stages, none past moderate, each contributing a different artefact.