Sound designIntermediate3 min read

Resampling

Definition

Resampling consists of recording to audio the output of a synth or an effects chain, and then reprocessing that audio as new material: cutting it, transposing it, reversing it, distorting it again, resampling it again. It is one of the main methods of modern sound design (neuro and dubstep basses, glitch, vocal chops, FX) because it allows chains that are impossible in real time and frees up CPU.

Why it works

  • It freezes decisions: the sound becomes an object to sculpt, not a patch to refine endlessly.
  • Irreversible processes: extreme time-stretching, reverse, slicing, pitch shifting with formants, granulating the result of another granulation.
  • Long chains: distortion → filter → resample → distortion → resample... each generation adds character (like multiple tape recordings).
  • CPU: a heavy synth becomes a light clip.

Typical workflow

  1. Play or program a phrase with the synth (e.g. 4 bass notes with different modulations).
  2. Render to audio (24-bit, no limiter on the master).
  3. Listen and pick the best fragments (often small portions of 100–500 ms).
  4. Cut and map the fragments in a sampler or in Slicex.
  5. Reprocess (pitch, formants, filter, distortion) and resample again.
  6. Organize the results into a personal library.

Techniques

  • Pitch down/up by octaves: a pad becomes a drone, a bass becomes an FX.
  • Reverse and reverse reverb.
  • Stretch (granular or phase-based) for textures and pads.
  • Chop & rearrange: reorder the slices to create new rhythmic patterns (a technique derived from jungle and hip hop).
  • Formant shift on vocals: vocal chops that become instruments.
  • Bounce in place of layers: merge 3 layers into a single file.
  • Drum resampling: the complete drums rendered, compressed and saturated, then blended in parallel with the kit.

In FL Studio

  • Render to audio / Consolidate: in the Playlist select the portion (or the pattern) and use the clip's consolidate/render options; alternatively export the selection from File → Export with “Split mixer tracks” if you need the individual tracks.
  • Edison: load it on a mixer insert and record the output directly (record button with “Record on input” or on play); then drag the result into the Playlist or a sampler.
  • Disk recording of the mixer: arm the insert and record to audio (by setting one insert as the input of another).
  • Slicex and Fruity Slicer for slicing; DirectWave for mapping.
  • Newtime and the clips' stretch options for working with time.

Example in A minor

Neuro bass at 174 BPM: in Serum or Vital record 8 notes of A1 with different LFOs (1/4, 1/8, 1/16, triplet), distortion and OTT. Export, slice with Slicex (each slice = a “gesture”), select 6 slices, transpose some to C2 and E2, add a flanger, resample the result and build the drop pattern by alternating the slices over a fixed sine sub on A1.

Common mistakes

  • Resampling with the master limiter on: the material is born already squashed.
  • Losing the key: note the pitch of each fragment (rename them: “neuro_A1_03.wav”).
  • Cutting away from the zero crossing: clicks; use 2–5 ms fades.
  • Not saving the original patch: always keep the source so you can redo it.

Exercise

  1. Take a trivial preset (a pad) and do three generations of resampling with different processes: note what you did at each step.
  2. Create a library of 20 sounds of your own derived from the same track (vocal chops, FX, percussion).

“Sampling and samplers”, “Granular synthesis”, “FX, risers, impacts”, “Distortion as sound design”, “Reese basses and wobble”.

Sources