Resampling Workflow

Resampling input: capture the whole master with your hands on it

The classic. Set an empty audio track input chooser to Resampling, arm it, record while you perform. You capture returns, master effects, Push macro moves, Looper overdubs — everything.

1. Create an audio track, input chooser top: Resampling. Set Monitor to Off (In would feed back the master into itself).
2. Bypass any master limiter or print with it — decide. For design material bypass it; you want headroom for later processing.
3. Arm, hit record on a Session clip slot or in Arrangement, perform 30–60 s of movement (filter sweeps, Echo feedback, Spectral Time freeze).
4. Stop, rename the clip immediately (see naming technique below), drag it to the browser or a Simpler.

Freeze / Flatten: commit with an undo

Freeze renders the track (devices included, sends excluded) to a temporary file and greys out the devices. Flatten replaces the chain with that audio permanently. Freeze first, listen, flatten only when you stop missing the knobs.

1. Right-click the track name, Freeze Track. Frozen tracks still let you move clips, change clip gain and automate mixer volume/pan/sends.
2. Duplicate the frozen track before flattening (Cmd+D). One copy stays frozen as the source, the other gets flattened.
3. Right-click, Flatten. Clips now carry the tail of release/reverb from the device chain (Freeze renders tails up to a limit; long delay tails get truncated — resample those instead).

Bounce in Place and Bounce to New Track (12.2)

Live 12.2 added the operations samplers have had for years. Select one or more clips, right-click, Bounce in Place renders them through their track's device chain and swaps the MIDI clip for the audio on the same track (devices stay, now bypassed by having audio in front of them — move the clip to an audio track or Flatten if you want them gone). Bounce to New Track puts the render on a new audio track and leaves the source alone.

1. Select the clip(s). Multiple clips across tracks bounce in one go, each to its own result.
2. Right-click, Bounce to New Track. The new track is named after the source; rename it now.
3. On the audio copy: reverse (R on a selected clip in Arrangement), warp mode Texture, Grain Size 60–120 ms, Flux 20–40% for a smeared ghost of the pad.
4. Play both together: dry synth plus reversed-texture copy 8–12 dB lower is an instant two-layer pad.

Bounce Group and Paste Bounced Audio (12.3)

Bounce Group renders the group output — every child track, its mixer, group devices — to one audio track. For a layered bass (sub + mid + top on three tracks under a group with Glue Compressor and Saturator) that is the whole stack in one clip, ready to be the next cascade stage. Paste Bounced Audio lets you copy a MIDI clip and paste it onto an audio track as the rendered audio, skipping the intermediate track entirely.

1. Right-click the group track header, Bounce Group. Choose whether the group is kept; keep it (muted) until the arrangement is done.
2. For the paste route: select the MIDI clip, Cmd+C, select an audio track and a slot, Edit menu Paste Bounced Audio. The clip is rendered through the source track's chain and lands as audio.
3. Name the result with the group name and a version suffix.

The resample cascade: bass, FX and textures in five passes

The method behind every "how did they make that" neuro bass and industrial texture: render, re-ingest in a sampler, process, render again. Each pass adds artefacts that no single chain produces, because the sampler's pitch/warp step sits between distortion stages.

1. Pass 1 (source): any 2-bar phrase with movement in it. A Reese with slow filter automation, or a TD-3 acid line recorded through External Instrument. Bounce to New Track.
2. Pass 2 (pitch and warp): drop the render into Simpler on a new track. Play it one octave down (−12 st) with Warp on, mode Complex Pro, Formants −20 to +20 to taste. Then Roar in Multiband mode, drive 10–15 dB on the mid band only. Bounce again.
3. Pass 3 (modulation and space): Sampler with the pass-2 render across three key zones (root, +7, +12) so chords are possible. Add Spectral Time with Freeze set to retrigger on Onsets, Fade In 50 ms, feeding Delay with Tilt +30%. Record via Resampling input while you play it, 16 bars.
4. Pass 4 (rhythm): Simpler in Slice mode (Transient, Sensitivity 50%) on the pass-3 capture. Program the slices from Push as a drum part; Drum Buss Transients +30%, Boom 60 Hz at 20% if it became a bass. Bounce.
5. Pass 5 (finish): EQ Eight (HP 30 Hz, tame 2–4 kHz if the cascade built a shelf there), Utility Bass Mono 120 Hz, Saturator Soft Sine 3 dB. Consolidate.

Variations: for textures, replace pass 2 with Granulator III (Grain Size 40 ms, Spray 30%, FM 10%) and pass 4 with Beat Repeat (Grid 1/32, Variation 50%). For FX risers, pass 3 becomes a Resampling capture of the sound with Echo feedback pushed to 110% for 4 bars then cut.

Stem Separation as a resampling source (12.3)

Live 12.3 separates any audio clip into Vocals, Drums, Bass and Other, each on its own track. Point it at your own old bounces, a jam printed through Resampling, a field recording of a band — and you get isolated layers to resample. The separation artefacts (phasey "watery" residue) are themselves a texture.

1. Right-click the audio clip, choose the stem separation command. Wait; it is offline processing. Four new tracks appear.
2. Drums stem into Simpler Slice mode, Split by Transient — your own break, re-sequenceable.
3. Bass stem: Convert Melody to MIDI if you want the notes, or pitch it −5/+7 in Simpler for new basslines of the same character.
4. "Other" stem is usually the most interesting: chords plus residue. Reverse, warp Texture, long Hybrid Reverb — instant pad.
5. Vocals stem from a non-vocal source (a modular jam) returns whatever the model thought was voice-like: resonant mids. Resample that as a formant texture.

Looper as a performable resampler

Looper records audio in tempo, overdubs with feedback, halves/doubles speed and reverses — and its drag-to-browser button exports the loop as a file. It is a resampler you can play with a foot controller.

1. Looper on the track after the effects. Set length to 4 bars, Song Control: Start Song so the first press defines tempo-locked length.
2. Record one pass, overdub two more with Feedback at 85% so earlier layers decay.
3. Hit Speed ÷2: everything drops an octave and halves tempo. Reverse. Record this to Resampling input or drag the loop to the browser.
4. The exported file goes into Simpler; now the layered loop is playable chromatically.

Clip warp modes as design tools

Warp modes were built for tempo-matching but each has a signature artefact. Pushed to extremes they are synthesis engines: Beats for stutter, Tones for pitch-crunch, Texture for smear, Re-Pitch for tape, Complex Pro for formant mangling.

1. Put the clip on an audio track, Warp on, pick the mode from the table.
2. Drive the transpose and warp-mode parameters with clip envelopes (Envelopes box: Clip, Transpose) for movement — that is automation inside the clip, printed when you Consolidate.
3. Cmd+J to consolidate the result into a new file. Now it is a clean sample with the artefacts baked in.

Convert audio to MIDI, then re-synthesize

The reverse resample: audio becomes notes, notes drive a different instrument, that renders to new audio. Great for extracting a hook's rhythm from a jam, or turning a TD-3 line into Operator FM.

1. Right-click the audio clip: Convert Melody to New MIDI Track for monophonic lines, Harmony for chords, Drums for percussion (which maps to a Drum Rack with the hits sliced in).
2. Clean the MIDI: delete ghost notes below velocity 20, quantize to 1/16 at 50% strength so the human timing survives.
3. Load any instrument. Bounce in Place. The new audio and the old audio now share rhythm and pitch — layer them.

Pitch and time tricks between passes

Where to put pitch change matters. Pitching before distortion folds harmonics differently than after; pitching a reverb tail down gives a bigger room than any Size parameter.

Naming and organizing renders so you find them again

A cascade produces 10–20 files per sound. Without a naming rule you will re-render things you already have. The rule: sound_passN_whatdidido_key_bpm, saved to a per-project folder in the User Library, tagged with a Collection colour.

1. Rename the clip before leaving the track: Cmd+R. Example: neuro_p3_spectralfreeze_F_172.
2. Drag the clip to User Library/Resamples/<project>/. Live copies the file with its name.
3. Assign a Collection colour: right-click the file in the browser, pick a colour. Red = bass, green = texture, yellow = drums, for example.
4. Keep the source track frozen and grouped under a "SRC" group at the bottom of the set, muted. Collect All and Save before closing.
5. Every render that survives two weeks goes into a "keepers" folder and gets a Simpler preset saved with it.