Sound Recipe - Vocoder & Pitched Vocal Textures

Target

Synthetic, choral, and robotic vocal layers used as instruments and texture: vocoders, harmonized stacks, pitched/formant-shifted voices, and granular vocal beds. The electronic-art-pop signature of treating the human voice as a synth. Expensive through clean carrier design, tasteful tuning, and layered width.

These textures support the other recipes (Icy Electronic Art-Pop, ionnalee, DM). This page is the technical reference for how to build them.

The textures should feel:

• Synthetic but musical
• Choral or robotic by intent
• Wide and layered
• Cohesive with the synth world (shared tone)

Follows the house method: intention → source → preset/plugin start → routing → settings → automation → taste checks.

Useful References

• The Knife / Fever Ray (pitched, gender-shifted, robotic voices)
• ionnalee / iamamiwhoami (layered breathy/choral stacks)
• Grimes (ethereal pitched harmony clouds)
• Imogen Heap / Daft Punk / Air (classic vocoder)
• Bon Iver (Messina-style formant harmonizer stacks)

Steal the jobs: voice-as-synth, harmony clouds, robotic leads, and granular vocal atmosphere.

Best For

• Hooks and counter-melodies
• Choruses needing width and lift
• Robotic/processed lead moments
• Synthetic backing beds and pads-made-of-voice
• Transitions and ear candy

Technique 1: Vocoder

Intention

Robotic/synthetic vocal where a synth (carrier) is shaped by the voice (modulator). Classic electronic-pop move.

Tools + setup

EVOC 20 setup (fast path)

1. Put EVOC 20 PolySynth on a software-instrument track
2. Set Side Chain to your vocal track (the modulator)
3. Play chords on the instrument track (the carrier)
4. Start with a bright saw carrier; adjust Formant and Bands (more bands = more intelligible)
5. Blend U/V Detection so consonants come through

Carrier design (the "expensive" part)

• Use a rich carrier: bright saws (Diva), wavetable (Pigments), or a pad — feed it as the carrier in VocalSynth, or use EVOC's internal synth
• The better the carrier, the more expensive the vocoder

Insert chain (on vocoder output)

1. Pro-Q 4 — HPF ~150 Hz, presence for intelligibility 2–5 kHz
2. UAD Studio D Chorus / Eventide TriceraChorus — width
3. Reverb/delay send (shared with synths)

Settings anchors

• Mic for the modulator: UT Twin87 (modern voicing) close, or an SM58 in an untreated room — clear consonants matter more than tone for vocoding
• Sing/speak clearly and rhythmically into the modulator
• Hold carrier chords slightly longer than the words for smear

Automation

• Bring vocoder in for hooks/choruses; automate formant for movement

Taste checks

• Intelligible enough to feel intentional. If muddy, fewer/clearer carrier notes + more bands.

Technique 2: Harmonized Stacks

Intention

Lush harmony clouds or tight robotic harmony — generated or pitched from one performance.

Tools + setup

Quadravox/Octavox setup

1. Insert on the vocal (or a duplicate)
2. Set Key and Scale to the song
3. Choose intervals per voice (e.g., +3rd, +5th, +oct, −oct)
4. Add small delay offsets per voice for width/lushness
5. Pan voices across the field

Insert chain (harmony bus)

1. Pro-Q — HPF ~150 Hz, thin low-mids (stacks get muddy)
2. Gentle bus comp (LA-2A) for cohesion
3. Reverb send (lusher than lead); MicroPitch for extra width

Settings anchors

• Harmony bus lower and wider than lead; less presence
• For "clouds": detune/offset voices; for "robotic": tight, quantized, formant-shifted

Automation

• Swell the harmony bus into choruses; mute in verses

Taste checks

• Lead stays on top. Widen, don't just raise. Thin 200–500 Hz if the stack muds up.

Technique 3: Pitched / Formant-Shifted Voices

Intention

Gender-shifted, chipmunk/deep, or alien lead and accent voices (The Knife/Fever Ray signature).

Tools + setup

• Logic Vocal Transformer 🟢 — pitch + formant separately (shift formant without pitch for gender morph)
• Eventide H910 / H949 ⭐ — vintage harmonizer pitch/detune, gritty character
• Eventide MicroPitch — subtle detune doubling
• Antares Throat / Mutator — formant/throat reshaping, vocal mutation
• Auto-Tune Pro (Throat/Formant + hard retune) — hard-tuned alien lead

Recipe for a gender-shifted lead

1. Duplicate the vocal
2. Vocal Transformer or H910: shift formant up/down (±several semitones), pitch to taste
3. Blend under or replace the lead for the effect
4. EQ to sit; send to shared space

Settings anchors

• Subtle for "uncanny human"; extreme for overt robot/alien
• Keep one clean readable layer if the lyric matters

Taste checks

• Decide: is this the lead (commit) or a texture (blend low)? Don't leave it half-in.

Technique 4: Granular Vocal Beds (voice-as-pad)

Intention

Turn a vocal into an evolving pad/atmosphere — choral wash made of your own voice.

Tools + setup

• Logic Alchemy / Sample Alchemy / Quick Sampler ⭐ — load a sustained vocal note, granular-stretch into a pad
• Eventide Crystals — reverse + pitched vocal halos
• Omnisphere — granular-import a vocal sample
• Spectral Gate / Ringshifter (Logic) — spectral vocal textures

Granular pad recipe

1. Bounce a held vowel (e.g., "ahh") across a few pitches
2. Load into Alchemy granular; set slow motion/spray
3. Pitch to chords; play as a pad
4. Heavy filter + reverb send; keep low in the mix

Taste checks

• It's atmosphere — submerged and filtered, supporting the lead.

Routing Summary

• Vocoder/robot layers: own tracks → shared synth reverb/delay (sit in the synth world, not the lead-vocal world)
• Harmony bus: lush reverb, wide, lower than lead; gentle glue
• Granular beds: heavy filter → big reverb send
• Keep the clean lead in its own (drier, central) space so processed layers don't blur the lyric

Fast Path

• Robotic hook: EVOC 20, sidechain the vocal, bright carrier, chords
• Harmony lift: Quadravox in key (+3/+5/+oct), pan wide, lush reverb
• Alien accent: Vocal Transformer formant shift on a duplicate
• Vocal pad: Alchemy granular on a held vowel, filtered, reverb

Adjustment Rules

Vocoder unintelligible: more bands, clearer carrier notes, raise U/V/consonant detection, presence EQ.

Harmony muddy: HPF + thin 200–500 Hz, widen instead of louder, fewer voices.

Robot sounds cheap: richer carrier (Diva/Pigments), add chorus/width, share the synth reverb.

Texture blurs the lyric: keep one clean central lead; move processed layers wider/quieter; automate them only where they help.

Common Mistakes

• Thin/buzzy carrier (the #1 reason vocoders sound cheap)
• Harmony stacks too loud and too narrow
• Processed layers competing with the lead's lyric
• Wrong key/scale on harmonizers (sour notes)
• Using every technique at once

Closest Tools I Own

Vocoder: Logic EVOC 20 PolySynth, iZotope VocalSynth 2, UAD Softube Vocoder Harmony: Eventide Quadravox/Octavox, Antares Harmony Engine, Logic Vocal Transformer Pitch/formant: Logic Vocal Transformer, Eventide H910/H949/MicroPitch, Antares Throat/Mutator, Auto-Tune Pro Granular/texture: Logic Alchemy/Sample Alchemy/Quick Sampler, Eventide Crystals, Omnisphere, Logic Spectral Gate/Ringshifter Carriers: u-he Diva, Arturia Pigments, KORG wavestate Width/space: Eventide MicroPitch/TriceraChorus, UAD Studio D Chorus, Lexicon PCM

Practical Summary

Treat the voice as a synth: design a rich carrier for vocoders (EVOC 20 / VocalSynth), build harmony clouds in-key with Quadravox/Octavox, create alien leads with formant shifting (Vocal Transformer / H910), and make pads from your own held vowels (Alchemy granular). Keep one clean central lead, place processed layers in the synth reverb world, and widen rather than over-loud. A good carrier and tasteful tuning are what separate expensive vocal-synth textures from cheap ones.

Related Pages

• Vocals
• Sound Recipe - Vocal Processing Chain
• Sound Recipe - Grimes / Layered Art-Pop
• Creative Effects and Transitions