Strings, Brass & Physical Modelling

Bowed string in Tension (string model from excitator to body)

Tension is a Karplus-Strong/waveguide string: an Excitator (Bow, Hammer, Hammer bouncing, Plectrum) drives a string with adjustable damping, a Termination (fret/finger model), a Pickup position and a Body (Piano, Guitar, Violin, Generic in several sizes). The result is less "realistic" than a sample and far more alive, because every parameter is a physical knob you can modulate.

1. Load Tension. Excitator type Bow. Force ~50–65 %, Friction ~50 %, Velocity sensitivity 30 % so hard notes dig in. Bow Position around 0.1–0.2 (near the bridge = brighter, more nasal).
2. String section: Decay 70–85 %, Damping ~45 % (higher damping = darker, dies faster), Inharmonicity 5–15 % (the slightly stiff steel-string feel; 0 is a pure waveguide), Ratio 100 %. Enable Damper with a mask around 50 % and small Damping velocity so the string stops when you release.
3. Termination: Fret stiffness 40 %, Finger mass 30 % — adds the fingered-note softness on attack. Pickup: enable, position 0.3–0.45; moving it changes the comb-filter colour exactly like moving a mic along a string.
4. Body: Violin, size Medium to start, Decay 60 %, Low cut low (keep the body's low resonance), Hi cut around 8 kHz, Mix 40–60 %. Mix at 100 % often sounds boxy — the body is a resonator you blend in.
5. Filter section: LP 24, cutoff 4–6 kHz with the filter envelope adding +20 % on attack to simulate bow pressure opening the tone. Vibrato: Delay 300 ms, Rate 5.5 Hz, Amount 10–20 %, Error 20 % for the human wobble.
6. Set Voices to 4–6 and keep it poly for sections, mono + legato (Legato on, Portamento 60 ms) for solo lines.

Variations: Excitator Plectrum with Protrusion 30 % and Stiffness 70 % = koto/harp pluck; Hammer with Mass 60 % = hammered dulcimer / clavinet-ish; Body Piano at Large with a Hammer excitator is a decent felt-piano stand-in.

Collision for struck and sustained resonators (cimbalom, bowed metal, string-ish hits)

Collision is two mallet/noise excitators into two resonator banks (Beam, Marimba, String, Membrane, Plate, Pipe, Tube) that can run serial or parallel. "String" resonators with a Noise excitator give a bowed-metal / sustained-string hybrid you can't get from samples.

1. Excitator: turn Mallet off and Noise on. Noise filter BP, Freq ~2 kHz, Width 40 %. Noise Volume envelope: Attack 150 ms, Decay 800 ms, Sustain 70 %, Release 400 ms — this is the "bow".
2. Resonator 1: Type String, Decay 80 %, Material 40 % (low = darker, string-like; high = metallic), Radius 30 %, Listening position 0.25, Hit position 0.1. Tune to keyboard (Tune 0, Fine 0). Brightness 30 %.
3. Resonator 2: Type Plate, Decay 50 %, Material 60 %, Tune −12 st, Mix lower (−8 dB). Routing Parallel (1+2) so the plate adds body around the string rather than filtering it.
4. LFO 1 → Noise Freq, ±10 %, 0.3 Hz — a slowly wandering bow position. LFO 2 → Res 1 Material, tiny amounts for life.
5. Add Roar after it: Multiband with only the mid band driven (Soft Sine, 3–5 dB) to give the sustain some grit without exciting the low resonance.

Variations: Resonator Tube + Noise excitator = breathy flute/pan-pipe; Membrane + Mallet, Decay 20 % = taiko-ish hit for cinematic percussion.

Classic synth brass (Operator, Wavetable or Analog): filter envelope + fast attack + detune

The Jupiter/Prophet brass formula: 2 slightly detuned saws, low-pass filter with a snappy positive envelope, amplitude attack of 10–30 ms, a little pitch dip on attack. Works identically in all three synths — the recipe is about envelopes, not oscillators.

1. Analog version: Osc 1 Saw, Osc 2 Saw, Osc 2 Detune +7 cents, Osc 2 Octave 0 (or −1 for fat brass). Filter 1 LP4 (24 dB), Freq 900 Hz, Res 10 %. Filter envelope amount +40 %: Attack 15 ms, Decay 250 ms, Sustain 35 %, Release 200 ms. Amp env: A 20 ms, D 300 ms, S 80 %, R 250 ms.
2. Pitch env (Analog: Osc Pitch Env): −30 cents, time 40 ms — the lip-slur "blat" into the note. Tiny but it is 50 % of the brass character.
3. Vibrato: Rate 5 Hz, Amount 6 %, Delay 250 ms so the first part of the note is straight.
4. Wavetable version: Osc 1 Basics>Saw, Osc 2 Basics>Saw transposed −12 with 7 ct detune, Filter 1 MS2 LP 24, cutoff 800 Hz, Env 2 → Filter freq +45 % with the same times. Unison Classic 3 voices, amount 20 % for section brass.
5. Operator version: Algorithm with A+B parallel (all four side by side), A Saw, B Saw with Fine +7, C Square at −18 dB for a reedy edge. Filter LP 24 with Env amount 40 %, Filter env A 15 ms.
6. Chorus-Ensemble in Classic mode, Rate 0.4 Hz, Amount 15 %, Dry/Wet 30 % — subtle. Then Saturator Analog Clip 2–4 dB to glue the detuned voices into one brass.

Variations: Grime/trap brass stab: same patch, amp decay 150 ms, sustain 0, Roar Hard clip after; Synthwave: Osc 2 at −12 with PWM square instead of saw.

String ensemble from Wavetable unison + Chorus-Ensemble

The Solina/ARP String Ensemble approach: a bright saw through a multi-voice chorus. Wavetable's unison does the voice-stacking, Chorus-Ensemble's Ensemble mode does the three-phase BBD chorus that defines the sound.

1. Osc 1 Vintage category, pick a saw-ish table (the Vintage tables have the slightly uneven harmonic ladder of real analog), position 20 %. Osc 2 same table, −12 st, level −6 dB for cello support.
2. Unison: Classic, 6 voices, Amount 30–40 %. Try Position Spread mode too — each voice gets a different wavetable position = each voice is a different instrument. That is closer to a real section than detune alone.
3. Filter 1 OSR LP 12 dB (gentler slope keeps the top lively), cutoff 3 kHz, Env 2 → cutoff +15 % with a 200 ms attack so the top end swells in slightly after the fundamentals.
4. Amp env: A 180 ms, D 500 ms, S 85 %, R 600 ms. Strings need the long release — notes overlap physically because bows do not stop instantly.
5. Mod matrix: LFO 1 (sine 5.2 Hz) → Osc 1 and 2 pitch 5 %, with the modwheel → LFO 1 amount so vibrato is performed, not constant.
6. Chorus-Ensemble: mode Ensemble, Rate 0.5–0.8 Hz, Amount 50–70 %, Width 100 %, Dry/Wet 50 %. This device's Ensemble mode is a 3-phase chorus like the Solina — the single most "string machine" effect in Live.

Variations: Disco strings: add Osc 2 at +12 with the same table, cut everything below 250 Hz; Lo-fi: Redux after it with Sample rate ~18 kHz; MPE: pressure → unison amount so chords bloom under your fingers.

Orchestral multisamples in Sampler: zones, velocity layers and round robin

Sampler is the right tool for your own orchestral library: key zones, velocity zones for dynamics (p/mf/ff), and Sample Select zones used as round robin so repeated notes do not machine-gun. All of it in the Zone tab.

1. Drop your samples in; Sampler builds a zone per sample. Open the Zone tab. Key: set root and zone range (one sample every minor third is fine with ±1 st stretch, the Fade handles crossfade). Enable key fades of 1–2 semitones to hide zone joins.
2. Velocity: stack 2–3 recordings of the same note (p, mf, ff) and set vel ranges 1–50 / 51–100 / 101–127 with ~15-value crossfades so a crescendo does not step.
3. Round robin: Sampler has no RR switch, but it has a Sample Select zone editor (0–127). Give four variants of the same note sample-select ranges 0–31 / 32–63 / 64–95 / 96–127, then in the Modulation tab route an LFO (sample-and-hold, retrigger on) or a Random MIDI device → Sample Selector. Each note picks a different take. The MIDI tool Random set to Chance 100 % on a CC alternatively works via a CC → Sample Selector mapping.
4. Filter: Morph LP 24, cutoff 8 kHz, velocity → cutoff +20 % for brightness scaling. Pitch/Osc tab: if the samples are dry, tiny vibrato via LFO 2 → pitch 4 cents.
5. Amp env Attack 0 (the sample has the real attack) but Release 400 ms so short MIDI notes still tail. Map modwheel → volume −12 dB range for swells (next technique).

Variations: Simpler in Classic mode for one-sample legato cello with Glide on; Drum Rack with one Simpler per articulation (spiccato / sustain / pizz) on separate pads for "keyswitching" by pad.

Swells and dynamics with modwheel, Expression Control and MPE pressure

The difference between a string pad and a string performance is that loudness and brightness are being played continuously. Route CC1 (or pressure) to volume, filter cutoff and a touch of vibrato depth, record it as a lane, then edit the curves.

1. Put Expression Control (Max for Live, in the MIDI effects) before the instrument. Source: Modwheel. Map 1 → instrument volume (Utility gain or Wavetable amp level), range −18 → 0 dB, curve slightly exponential.
2. Map 2 → filter cutoff, range 800 Hz → 6 kHz. Map 3 → LFO 1 amount (vibrato depth), 0 → 8 %. Now one wheel does loud + bright + expressive.
3. On Push 3 or an MPE controller, choose Pressure as the source instead, or additionally — Expression Control has Modwheel / Pitch Bend / Velocity / Aftertouch / Keytrack / Random sources.
4. Record the performance. In the clip, show the Modwheel or Pressure envelope lane in the Stacked detail view and draw/smooth the swells. Hold Alt while dragging envelope segments to curve them.
5. Add an Envelope Follower on a parallel "bow noise" chain (see below) driven by the same swell so the noise rides with the dynamics.

Variations: Use the LFO M4L device in Random or Sample-and-hold mode at very slow rates on the same three targets for an "unattended" string texture that breathes on its own.

Trap / grime orchestral stabs

Short, aggressive, dark orchestral hits — brass/string marcato samples pitched down, layered with a synth saw for consistency, then heavily compressed and saturated so they cut through 808s.

1. Chain 1: Sampler with brass/string marcato samples, Transpose −5 to −7 st (darker, bigger). Amp env Decay 350 ms, Sustain 0 so every hit is the same length regardless of MIDI note length.
2. Chain 2: Operator, A Saw, B Saw Fine +9, parallel algorithm, LP 24 at 1.5 kHz with env +30 %; amp env Decay 300 ms, Sustain 0. −8 dB under the sample. This fills in the pitch definition the sample loses when transposed.
3. Glue Compressor on the rack output: Ratio 4, Attack 1 ms, Release 0.2 s, aim 5–8 dB gain reduction. The compressor flattens the sample's natural decay into a block.
4. Roar: Serial routing, stage 1 Soft Sine 3 dB, stage 2 Hard 4 dB, post-filter LP 8 kHz. Tone to taste.
5. Pattern: stabs land on 1 and the "and" of 3, or follow the 808 pattern. Sidechain the stab bus to the kick with a short 30 ms duck so the kick keeps its transient.

Cinematic braam: Operator + Roar + Hybrid Reverb

The trailer "BRAAAM": a massive, slightly detuned low brass cluster with a long swell-in, huge distortion in the mids and a reverb that is part of the sound. Operator for the core because its per-oscillator envelopes let the harmonics arrive one at a time.

1. Operator, algorithm all-parallel. Osc A Saw, level 0 dB, env A 400 ms. Osc B Saw, Fine +7, env A 900 ms (arrives later — the swell). Osc C Square, Coarse 0.5 (one octave down), level −8 dB, env A 1.5 s. Osc D Noise, level −20 dB, env A 2 s, LP'd by the Operator filter.
2. Filter LP 24, cutoff 500 Hz, filter env +50 % with Attack 1.8 s — the whole thing opens up over almost two seconds. Pitch env −2 st over 1.2 s on all oscs (the sagging-pitch "doom" effect).
3. Spread 40 % in Operator for width without an external widener. Voices 4, play a fifth or an octave-and-fifth cluster in the C1–C2 range.
4. Roar: Multiband. Low band clean (keep sub tight), Mid band Fold 6–10 dB, High band Soft Sine 3 dB. Feedback section: Amount 15 %, pitch −12 st for a subharmonic growl, Dry/Wet 70 %.
5. Hybrid Reverb: Dark Hall, Decay 6–10 s, Pre-delay 40 ms, Blend 50/50 Convolution (a large hall IR) + Algorithm, Dry/Wet 40–50 %. Vintage mode and a little Shimmer-less darkness: EQ the reverb's high shelf down.
6. Limiter at the end, ceiling −1 dB, so the swell hits a wall rather than clipping the master. Then resample it — a braam is a one-shot sample in the finished track; you need the tail baked in to reverse it, chop it or pitch it.

Variations: Industrial techno: 3-bar braam at the breakdown, reversed into the drop; Dubstep: tune to the track key, cut to one bar, add Shifter +12 blended at 20 % for a horror edge.

MPE bowing with Drift or Wavetable (pressure = bow, slide = position)

With Push 3 or any MPE controller, per-note pressure is a bow: it controls level and brightness after the note has started, and slide (per-note Y / CC74) moves the "bow position" (filter or wavetable position). The patch is simple; the mapping is everything.

1. Enable MPE on the track (track title bar → MPE / expression settings) and make sure the controller sends MPE (Push 3 does in its MPE layout). Drift and Wavetable both receive per-note Pressure, Slide and Pitch Bend natively.
2. Drift: Osc 1 Saw, Osc 2 Tri at −12, Noise 10 %. Filter MS2 LP, cutoff 600 Hz, Res 15 %. Amp env: A 60 ms, S 100 %, R 500 ms. Mod matrix: Pressure → Filter Freq +50 %, Pressure → Amp (via Env 2 amount or LFO amount) +40 %, Slide → Osc Mix (tri to saw = soft to hard bow), Pitch bend range ±2 st for vibrato by wiggling.
3. Set Drift's voice mode to Poly with Drift amount 20 % — each voice's analog-style instability reads as different players. Legato off (MPE handles the per-note gestures).
4. Wavetable alternative: Osc 1 Formant table with Pressure → wavetable position (the "pressure opens the vowel" effect) and Pressure → filter cutoff; Slide → Unison amount so sliding up widens the section.
5. Play: put the finger down softly (quiet, dark), lean in (loud, bright), slide up for harder bow. Record it; each note's lanes are editable in the Stacked detail view.

Variations: Tension does not receive MPE natively; wrap it in an Instrument Rack and map pressure via Expression Control set to Aftertouch for a mono-ish bowed line. Granulator III with a cello sample and pressure → Grain Size works beautifully.

Bow-noise layering: the parallel rosin chain

Take any synthesized string (Wavetable, Analog, Drift) and add a separate quiet noise chain that follows the same envelope and the same dynamics. The noise is what tells the ear "friction".

1. Instrument Rack. Chain A = your string patch. Chain B = Wavetable with Osc 1 from the Noise category (or Simpler with a real bow-noise/breath recording looped), Filter 1 BP, 3–4 kHz, Res 20 %, Filter 2 HP 1.5 kHz.
2. Chain B amp env mirrors chain A's but with Attack 20 % longer and Sustain lower (0.4) — the scratch is loudest when the bow starts moving.
3. Chain B level around −18 to −24 dB relative to A. If you can clearly hear it solo'd as "noise" in the mix, it is too loud; it should only be missed when muted.
4. Put an Envelope Follower (M4L) on chain A, mapped to chain B's Utility gain, Rise 20 ms, Fall 200 ms. Now loud phrases get more rosin automatically — also works from the modwheel swell above.
5. Slight LFO (0.2 Hz, random) on chain B's BP frequency ±15 % so the noise colour wanders like a real bow changing position on the string.

Legato lines, glide and portamento for solo strings/brass

Solo instruments are mono: a cellist cannot play two notes on one string. Mono voice mode + legato retriggering + a short glide makes a synth line phrase like a player instead of a keyboard.

1. Voice mode Mono (Drift, Wavetable, Analog all have it; Sampler has a Voices = 1 setting with Retrigger options; Operator has Voices 1 with Legato).
2. Enable Legato: overlapping notes do not retrigger the envelopes — the second note inherits the first note's open filter. This is the "slur". Non-overlapping notes retrigger = "détaché".
3. Glide/Portamento 30–80 ms. Drift has a Glide knob; Analog has Glide with Legato mode; Wavetable has Glide in the global section. Keep it short; above 120 ms it is a theremin.
4. In the MIDI clip, overlap the notes you want slurred by a 32nd and leave gaps where the bow changes. Use the Note Length MIDI effect with a small positive Length (10–20 %) on a whole clip to legato-ise a quantized line fast.
5. Brass: add a Pitch env of −20 cents, 30 ms, which only triggers on non-legato notes — tongued notes get the blat, slurred notes do not. That alternation is what phrasing sounds like.

EQ positioning: where strings and brass sit against the mix

Orchestral sounds are wide-band and will eat everything. Carve them into the gaps and use width, not level, to make them big.

• Strings pad/section: HP 100–150 Hz, −2 to −4 dB at 250–400 Hz, leave 1–2 kHz for leads/vocals (−2 dB if needed), +2 dB shelf at 8–10 kHz for air. Width 120–140 % via Utility, mono below 150 Hz.
• Brass stabs: HP 80 Hz, cut 300–500 Hz, keep 1–3 kHz (that is the stab), LP at 10–12 kHz so they do not fight hats. Width narrow (80–100 %) — brass is a point source.
• Solo cello/violin: keep the body (200–400 Hz) — that is what makes it real — but cut the same band on the pad behind it. Dead centre or slightly off-centre, never wide.
• Sidechain: 1–3 dB of duck from the kick on the whole string bus keeps big sustained chords from blurring the groove, without the pumping of a bass sidechain. Long release (300–500 ms).