Definition
Vocal production is the work that takes you from recorded takes to a finished vocal track, ready for the mix. The three main operations are comping (building the best take from several takes), tuning (pitch correction) and timing (timing correction and alignment between tracks). To these are added cleanup and leveling (clip gain).
1. Before recording
- Gain: average level around −18 dBFS, peaks at −10/−6 dBFS (at 24-bit there is enough headroom).
- Headphones: monitor mix with little reverb and the voice a little loud; a singer who hears themselves well sings in tune.
- Microphone and distance: 10–20 cm with a pop filter; a constant distance makes comping easier.
- Section list: record by sections (verse, pre, chorus) 3–6 takes each, plus 1–2 full takes for emotional continuity.
2. Comping
- Listen to all the takes line by line and choose the best for each phrase, sometimes for a single word or syllable.
- Criteria in order: emotion and intention, timbre, pitch, timing. An emotional, slightly imprecise take beats a perfect, cold one (pitch and timing can be corrected, emotion cannot).
- Cuts should be made at breaths, pauses or on consonants (best before a plosive or sibilant consonant), with short crossfades (5–20 ms).
- In DAWs with a comping function (Pro Tools playlists, Logic Quick Swipe, Cubase lanes) you select by dragging; in FL Studio you can spread the takes over several Playlist tracks and assemble the final take by cutting and moving clips onto a “Comp” track (recent versions offer dedicated take functions: check the manual for your version).
3. Cleanup and editing
- Breaths: keep them (they are part of the performance) but lower them by 6–12 dB with clip gain; remove them on doubles and harmonies.
- Noises: mouth clicks, mic bumps, background noise (iZotope RX De-click, De-noise).
- Plosives: HP below 80–100 Hz on the individual clip or manual correction.
- Clip gain: level phrases and words before compression, so the compressor works less.
4. Timing
- Align words that fall out of the groove only if they bother you: the voice is naturally “elastic”.
- For doubles and harmonies: align attacks and word endings to the lead with VocAlign (Synchro Arts) or Revoice Pro, or manually.
- In rap rhythmic precision matters more: align syllables and consonants to the grid or to the drums.
5. Tuning
- Melodyne (Celemony): graphical note-by-note editing of pitch, vibrato, transitions, formants; the standard for transparent correction.
- Auto-Tune (Antares, introduced in 1997): automatic real-time correction. With a very fast retune speed (0–10 ms) it produces the robotic effect made famous by Cher — Believe (1998) and later by T-Pain and trap; with values of 20–50 ms and “humanize” it gives a more natural correction.
- FL Studio: Newtone (graphical editing, similar to Melodyne) and Pitcher (real-time correction, also a harmonizer).
- Rules for natural correction: correct the center of the note, leaving the transitions and vibrato; don't bring every note exactly to 0 cents; check long notes and the final notes of phrases (where out-of-tune notes are most noticeable).
- Correct scale: set the right key and scale in the plugin (in A minor: A B C D E F G); with chromatic melodies use the chromatic scale or manual editing.
6. Recommended workflow
Comping → cleanup → clip gain → timing → tuning → print (render) the corrected track → mix processing (EQ, compression, de-esser, space). Printing the corrected vocal before the mix lightens the project and “freezes” decisions.
Ethics and style
The amount of correction is a stylistic choice: mainstream pop and trap use obvious correction; soul, jazz, folk and singer-songwriter music prefer minimal correction. Communicate with the artist about what you want to achieve.
Practice exercise
Record 4 takes of an 8-bar verse. Comp the best take with at least 6 cuts, lower the breaths by 8 dB, correct the pitch with Newtone or Melodyne only on the long notes, then record a double and align it to the lead. Compare the vocal before and after and check that it doesn't sound artificial.
Links
Doubles and Triples, Vocal Stacks, Ad-libs and Vocal Runs, Gain Staging, Compression.