← Back to Course

Module 08

Vocal Production

Recording, processing, and polishing vocals from capture to mix

Part A

Core Concepts

01

Understanding the Voice

Vocal Anatomy: Resonators Shape the SoundNasal 1-3kHzMouth 500-2kHzThroat 250-500HzChest 80-250HzVocal cordsFREQUENCY ZONESBody/Warmth80-250HzCore Tone250-500HzVowels/Formants500-2kHzPresence2-4kHzSibilance4-10kHz

Sound starts at the vocal cords, then resonates through the throat (pharynx), mouth, and nasal cavities. Each shapes the timbre. Singer technique controls the balance.

MUFFLE POINTS: Where Resonance Gets BlockedTHROAT MUFFLE200-400Hz← Congested zoneThick, swallowed soundFix: Technique, EQ cutNASAL MUFFLE800-1200Hz← Honky zoneNasal, telephone qualityFix: Decongestant, EQ cut ~1kHz

Muffle points are frequency ranges where resonance is blocked or absorbed. Throat and nasal muffle are the two main issues—both create dull, congested sound.

FORMANTS: Resonant Peaks Define VowelsF1 (250-800Hz)Mouth opennessClosed → Open"ee" low F1"ah" high F1F2 (800-2500Hz)Tongue positionBack → Front"oo" low F2"ee" high F2VOWEL MAP"ee" F1↓ F2↑"ah" F1↑ F2→"oo" F1↓ F2↓EQ changes formants

Formants are resonant frequency peaks that define vowels. F1 (low) and F2 (high) together create each vowel sound. 'Ah' has different formants than 'Ee'.

COMMON PROBLEM FREQUENCIES IN VOCALS150-250HzMuddiness, proximity boom300-500HzBoxy, cardboard quality800-1200HzNasal honk, telephone quality2-4kHzHarshness, fatiguing5-8kHzSibilance (S, T, F)8-12kHzAir / excessive breathThese are starting points—every voice is different. Use your ears!

Certain frequency ranges consistently cause problems across different voices. Knowing these helps you hunt for issues efficiently.

02

Recording Vocals

MICROPHONE SELECTION FOR VOCALSLDC (Condenser)Detail, presenceStudio standardU87, C414, AT4050Best for: Clean studiosDynamicLess room pickupRaw, aggressiveSM7B, SM58, RE20Best for: Untreated roomsRibbonWarm, smoothVintage characterR121, R84Best for: Harsh voicesMATCHING TIPBright voice +bright mic = harshDark voice +dark mic = muddyMatch for balance

Large diaphragm condensers are standard for their detail and flattering tone. But dynamic mics (SM7B) work great for aggressive styles or problematic rooms.

Vocal Mic Positioning: Distance & Angle MatterDISTANCE (side view)SingerMic8-12"Closer = more bass (proximity)Farther = more naturalANGLE (top view)SingerOff-axisStraight: Full responseOff-axis: Reduces sibilanceStart: 8-10", straight on, pop filter 4" from mic

Position affects tone dramatically. Distance affects proximity effect. Angle affects sibilance and plosives.

PROXIMITY EFFECT: Distance Changes BassCLOSE (2-4")+++ BASSIntimate, boomyRadio/broadcast voiceSTANDARD (8-12")NaturalBalanced, naturalStudio standardFAR (18"+)ThinLess bass, more roomNatural but distant

Close to a cardioid mic = bass boost. Radio announcers use this for 'broadcast voice'. Singers can use it for warmth or create mud.

ROOM SETUP FOR VOCALSTREATMENT PRIORITIES1. Behind Singer2. Corners3. CeilingFirst reflection behind singeris MOST importantROOM VIEW (top)SingerMicabsorbedA treated corner beats an untreated booth

Room reflections are the most common problem in home vocal recording. Sound bouncing back into the mic creates comb filtering and boxy tone.

03

Sibilance & De-Essing

WHAT CAUSES HARSH SIBILANCEBright Mic + Bright VoiceDoubles high frequenciesOn-Axis Mic PositionDirect S sounds hit micCompression Raises SQuiet S sounds get louderEQ Boost in 3-8kHz RangePresence boosts also boost sibilancePREVENTIONOff-axis mic, darker mic, singer technique

Some sibilance is natural—consonants need high frequencies to be intelligible. Problems occur when it becomes fatiguing or disproportionate.

HOW DE-ESSERS WORKFREQUENCY-SELECTIVE COMPRESSORSIDECHAIN4-8kHz band→COMP→OUTOnly compresses when S sounds hitTWO TYPESSPLIT-BANDOnly S bandcompressedMore transparentWIDE-BANDFull signalcompressedMore obvious

A de-esser is a frequency-selective compressor. It only compresses when the sibilant frequency range exceeds a threshold. Transparent when set correctly.

DE-ESSER SETTINGS GUIDE1. FIND FREQSolo sidechainListen for S soundsUsually 5-8kHz2. THRESHOLDOnly trigger onactual S soundsNot on singing3. REDUCTIONStart subtle2-6dB maximumLess is more4. CHECKListen in mixnot solo✓Process BEFORE compression—compression raises S sounds

Start subtle. Over-de-essing is worse than slight sibilance. The goal is natural S sounds, not their elimination.

MANUAL DE-ESSINGWAVEFORM VIEWSSSelect S region → Reduce gain 3-6dBWHEN TO USE✓ Important vocals✓ De-esser artifacts✓ Surgical precisionTime-consuming but precise

For surgical control, manually reduce individual sibilant sounds using automation or clip gain. Time-consuming but precise.

04

Breath Control & Editing

BREATH EDITING PHILOSOPHYKEEPEmotional breathsBetween phrasesIntimate/jazz stylesAdds humanityREDUCELoud gaspsDistracting breathsPop/produced styles-6 to -12dBREMOVEOnly obviousgasping soundsthat distractRarely remove all

Breaths communicate emotion: desperation, relaxation, intensity. Total removal creates uncanny valley. Selective editing preserves humanity while reducing distraction.

REDUCING BREATH VOLUMEBEFORE / AFTERBeforeBRAfter (-6dB)brMETHODS• Clip gain: Select, reduce• Automation: Draw dip• Plugin: Auto-detect6dB reduction often enough

Instead of removing, lower the breath's volume by 6-12dB. Keeps timing natural but makes breaths less prominent.

REMOVING BREATHSCUT TECHNIQUEWordcutWordxfadeWordWARNINGRemoving all gapsmakes phrases runtogether—unnaturalSpace IS timing

For complete removal, cut the breath and close the gap (or leave room tone). Be careful about timing—songs need space.

BREATH EDITING TOOLSiZotope RXBreath ControlDe-noiseIndustry standardWaves DeBreathAuto-detectReduce/removeQuick workflowVocal RiderCan duck breathsAutomatic levelMulti-purposeManualClip gainStill best forcritical vocals

Modern tools can detect and process breaths automatically, speeding up the workflow significantly.

05

Pitch Correction

TRANSPARENT PITCH CORRECTIONSETTINGS FOR NATURAL SOUNDSpeed: 20-50ms (slow)Amount: 50-80%Fix only wrong notesAPPROACH• Go through vocal note by note• Fix only obviously wrong notes• Keep vibrato and slides• Goal: Listener never knows

For natural sound, use slower correction speeds and less aggressive settings. The goal: fix mistakes without revealing the tool.

AUTO-TUNE AS EFFECTTHE T-PAIN / CHER SOUNDSpeed: 0-10ms (instant)Amount: 100%Key: Set to song keyPITCH GRAPHInstant pitch snapSinger should still be close to pitch

The 'Cher effect' or T-Pain sound uses fast correction that snaps pitch instantly, creating the robotic warble that's become a genre staple.

MANUAL PITCH EDITINGMELODYNE / FLEX PITCH VIEWC4D4E4?drag upF4G4BEST FOR✓ Important vocals✓ Specific problems✓ Preserve characterTime-consuming but precise

For surgical control, manually adjust individual notes using graphical editors like Melodyne, Flex Pitch, or VariAudio.

FORMANT PRESERVATIONWITHOUTPitch Up = ChipmunkFormants shift tooPitch Down = MonsterVowels change characterWITHPitch changes naturallyFormants preservedVowels stay naturalVoice character intact ✓WHENLarge shifts(3+ semitones)Not needed forsmall correctionsCreative use too!

Shifting pitch naturally shifts formants, changing vowel character. 'Formant correction' preserves the original vowel sound when pitch changes.

06

Vocal Processing Chain

STEP 1: CLEANUPFIX BEFORE SHAPINGClicks/PopsRemoveNoiseReduceBreathsEditPitchCorrectTimingWHY FIRST?Compressionamplifies problemsNoise gets louderFix before compressing

Before any tonal processing, fix the raw recording. Remove clicks, reduce noise, edit breaths, apply pitch correction.

STEP 2: SUBTRACTIVE EQCOMMON CUTSHP 80-120Hz → Remove rumble200-300Hz → Cut if boomy400-600Hz → Cut if boxy800-1.2kHz → Cut if nasal3-4kHz → Cut if harshWHY BEFORE COMP?Compressor makeseverything louderIncluding problemsCut first!

Remove problem frequencies before compressing. A compressor working on muddy audio will lock in the mud.

STEP 3: COMPRESSIONVOCAL COMPRESSION SETTINGSRatio: 2:1 to 4:1 (gentle-moderate)Attack: 10-30ms (transients through)Release: Auto or ~150msTARGETS3-6dB reduction on peaksIf 1 isn't enough, use 2 in series(2-3dB each, not 6dB on one)

Compression evens out volume variations, keeping the vocal consistently present. Settings vary by style and performance.

STEP 4: ADDITIVE EQ & ENHANCEMENTCOMMON BOOSTS (gentle!)100-200Hz: Warmth (+1-2dB)3-5kHz: Presence (careful!)10-12kHz: Air and sparkleTHEN...→ De-esser (presence boosts sibilance)→ Saturation (optional warmth)→ Final limiter (optional peaks)Keep boosts small: 1-3dB usually enough

Now shape the tone: add presence, air, warmth. De-esser typically goes after additive EQ since boosting 3-5kHz can increase sibilance.

07

Vocal Doubles & Harmonies

DOUBLE-TRACKING TECHNIQUEREAL DOUBLES• Sing the part again (don't copy!)• Variations ARE the point• Pan L/R for width, center for powerIF YOU MUST FAKE IT• Short delay (10-30ms)• Slight pitch shift• Never sounds as good

Singing the same part twice creates natural thickness from tiny timing and pitch variations. Pan left and right for width, or keep centered for power.

VOCAL HARMONIESRECORDINGRecord to tuned lead vocalMatch lead's phrasing closelyCommon: 3rds, 5ths, 6thsMIXING• Lower level than lead• Often darker EQ• Tuning MUST match lead exactly

Harmonies sing different notes that complement the lead. Common: 3rds above/below, 5ths, sometimes 6ths. Must match lead vocal's tuning.

VOCAL STACKSSTACK = MANY TAKES4-8 takes, panned across fieldCRITICALTight timing essentialCareful tuning (errors multiply)Can sound phasey if sloppyCreates choir-like thickness

Stacking is recording many performances of the same part. 4, 8, or more takes create a thick, choir-like effect. Common in pop choruses.

VOCAL ARRANGEMENTBUILD THROUGH THE SONGVERSELead onlyPRE-CHORUS+harmonyCHORUSFull stackBRIDGEDifferenttreatmentFINALEVERYTHING

Where and when vocal layers appear affects the song's dynamics. Build from verses to choruses. Don't put everything everywhere.

08

Vocal Effects & Character

VOCAL REVERBREVERB TYPESPlate: Classic vocal, smoothRoom: Natural, intimateHall: Big, dramaticKEY SETTINGSPre-delay: 20-50ms (keeps vocal upfront)Decay: Match tempo/feelUse sends, not inserts

Reverb places the vocal in a space. Short reverb = intimate room. Long reverb = hall or dramatic. Too much = lost in wash.

VOCAL DELAYDELAY TYPESSlapback (50-120ms): Elvis vibe1/8 note: Subtle rhythmDotted 1/8: U2 style, polyrhythmTIPSSync to tempo for musical feelEQ delay darker = less obviousAutomate for specific words

Delay adds rhythmic repetition. Synced to tempo for musical echoes. Or short (slapback) for doubling effect.

VOCAL SATURATIONSATURATION TYPESTape: Warm, compressed feelTube: Even harmonics, smoothTransistor: Odd harmonics, edgeAMOUNTSubtle: Warmth without being obviousModerate: Edge and presenceHeavy: IS the sound (lo-fi, aggressive)

Saturation adds harmonic content, making vocals feel warmer, fuller, or grittier depending on the type and amount.

CREATIVE EFFECTSBEYOND STANDARD EFFECTSTelephone/radio: Band-pass + saturationVocoder: Robot/synth vocalReverse reverb: Swelling into phrasesPitch shift: Octave doublerGranular: Glitchy, experimentalWHEN TO USESpecific wordsTransitionsStyle momentsNot entire vocals usually

Beyond standard effects: filters, pitch shifting, vocoder, formant manipulation, and more for special moments.

Context

Analog vs Digital

A/D

Analog Approach

Digital Approach

Deep Dive

The Science

SCI

Vocal Science

In Practice

Iconic Vocal Productions

How legendary vocal sounds were achieved.

EX

Billie Eilish - 'Bad Guy' (2019)

Pop
Whisper Vocals: Extremely close miking for intimate, ASMR-like quality
Minimal Processing: Very little compression, natural dynamics
Creative Effects: Pitch processing, stacked doubles, unusual panning
Breath as Feature: Breaths left in as artistic element

T-Pain - 'Buy U a Drank' (2007)

Pop/R&B
Heavy Auto-Tune: Fastest setting, 0 retune speed
Effect as Signature: Not hiding the processing—it IS the sound
Simple Chain: Auto-Tune doing most of the work
Defined Genre: This sound became a genre marker

Adele - 'Someone Like You' (2011)

Pop Ballad
Natural Sound: Minimal tuning, real performance
Dynamic Preservation: Dynamics intact, not over-compressed
Room Sound: Some natural room in the recording
Emotion First: Technical perfection secondary to feeling

Bon Iver - 'Woods' (2009)

Experimental
Heavy Processing: Vocoder/harmonizer creating multiple voices
No Traditional Band: Voice only, processed to create everything
Auto-Tune as Instrument: Extreme settings for effect
Layering: Multiple processed vocal layers creating texture

Avoid These

Vocal Production Mistakes

Common problems and how to avoid them.

ERR
!

Over-Processing

Symptom: Vocal sounds artificial, lifeless, or fatiguing
Cause: Too much EQ, compression, de-essing, or pitch correction
Fix: Back off on processing. Compare to raw recording frequently. Less is usually more.
!

Sibilance Lisping

Symptom: S sounds like 'TH', unnatural consonants
Cause: De-esser set too aggressive or at wrong frequency
Fix: Reduce de-esser threshold, check frequency setting, use manual de-essing for critical vocals.
!

Pitch Correction Artifacts

Symptom: Warbling, unnatural pitch movement, 'chipmunk' sound
Cause: Wrong settings, too fast for style, formant issues
Fix: Slow down correction speed, use manual editing, enable formant preservation for large shifts.
!

Drowning in Reverb

Symptom: Vocal is lost, words unclear, mix is washed out
Cause: Too much reverb, too long decay, no pre-delay
Fix: Reduce reverb level/decay, add pre-delay (30-50ms), EQ the reverb to remove lows and harsh highs.
!

Bad Recording Fixes

Symptom: No amount of processing makes vocal sound good
Cause: Fundamental recording problems: room, mic position, performance
Fix: Re-record if possible. A well-recorded vocal with basic processing beats a bad recording with expensive plugins.
!

Ignoring Breath Control

Symptom: Harsh gasping sounds between phrases, or unnatural silence where breaths were removed
Cause: Either not editing breaths at all, or removing them completely
Fix: Reduce breath volume by 6-12dB rather than deleting. Some breath is natural and helps phrasing. Only remove truly distracting gasps.

Train Your Ears

Vocal Production Listening Tests

Train your ears to hear vocal processing.

EAR
01

Pitch Correction Speed Test

1. Apply pitch correction to a vocal
2. Set to slowest speed (50ms+), listen
3. Gradually speed up to fastest (0ms)
4. Notice: Slow = natural, fast = robotic
5. Find the speed that matches your style goal
Purpose: Understand how correction speed affects naturalness
02

De-Esser Threshold Hunt

1. Solo the de-esser's sidechain (listen mode)
2. Adjust frequency until you hear primarily S sounds
3. Set threshold so only S sounds trigger reduction
4. Listen in context: Do S sounds fit naturally?
5. Adjust until sibilance is controlled but not eliminated
Purpose: Learn to target sibilance accurately
03

Compression Amount Test

1. Apply compressor with 4:1 ratio
2. Start with high threshold (no compression)
3. Gradually lower threshold, watch gain reduction
4. Listen: At what point does vocal sound 'controlled' but natural?
5. Go too far—hear the 'pumping' and lifelessness
6. Back off to the sweet spot
Purpose: Find the compression sweet spot for different styles
04

Reverb Pre-Delay Test

1. Add reverb to vocal with 0ms pre-delay
2. Notice how vocal blends into reverb immediately
3. Add 30ms pre-delay, listen again
4. Add 50ms, 70ms, 100ms
5. Notice: Pre-delay keeps dry vocal upfront before reverb starts
Purpose: Understand how pre-delay affects vocal presence

Implementation

Vocal Production in Your DAW

Practical vocal processing tools and techniques.

DAW

Logic Pro

Key Synths/Tools:Vocal tools
Tips:
  • Flex Pitch: Double-click audio region for pitch editing
  • De-Esser: Channel strip has dedicated de-esser
  • Vocal Transformer: Creative pitch/formant effects
  • Sample Alchemy: New spectral editing for advanced processing
  • Comping: Quick Swipe Comping for take selection

Ableton Live

Key Synths/Tools:Vocal tools
Tips:
  • Audio to MIDI: Extract melody for harmony reference
  • Vocoder: Built-in vocoder effect
  • Tuner: Check pitch live or after recording
  • For pitch correction: Use third-party (Melodyne, Auto-Tune)
  • Comping: Take lanes available in Arrangement View

FL Studio

Key Synths/Tools:Vocal tools
Tips:
  • NewTone: Built-in pitch correction/editing
  • Pitcher: Real-time pitch correction plugin
  • Vocodex: Advanced vocoder
  • Edison: Recording and detailed editing
  • Fruity De-Esser: Basic but functional

Pro Tools

Key Synths/Tools:Vocal tools
Tips:
  • Elastic Pitch: Built-in pitch editing
  • Playlist Comping: Industry-standard take management
  • VocAlign: Timing alignment for doubles/harmonies
  • Native plugins: De-esser, dynamics available
  • Third-party: Pro Tools users often use Waves, iZotope, etc.

Third-Party Essential

Key Synths/Tools:Common plugins
Tips:
  • Melodyne: Industry standard for manual pitch editing
  • Auto-Tune: Real-time or graphical pitch correction
  • Waves Vocal Bundle: De-esser, doubler, rider
  • iZotope RX: Vocal cleanup, de-noise, breath control
  • Soundtoys: Little AlterBoy for formant/pitch effects

Self-Check

Vocal Production Quiz

Q
Q1

What is a 'muffle point' in vocal terminology?

Show Answer
Q2

Why should de-essing typically come AFTER additive EQ in the chain?

Show Answer
Q3

What makes real vocal doubles sound different from copied/processed doubles?

Show Answer
Q4

What does pre-delay do on a vocal reverb?

Show Answer
Q5

What causes the 'chipmunk' effect when pitch-shifting vocals up?

Show Answer
Q6

What's the typical sibilance frequency range for most singers?

Show Answer
Q7

Why is pitch correction typically done before compression in a vocal chain?

Show Answer

Remember

Key Takeaways

KEY
Source

Get It Right at the Source

80% of vocal sound comes from recording: mic choice, position, room, performance. A well-recorded vocal needs minimal processing. You can't 'fix it in the mix'—not really.

Subtle

Less Processing is Usually Better

Over-processed vocals sound artificial and fatiguing. Small moves on multiple processors beats heavy processing on one. Always compare to the raw recording.

Chain

Order Matters

Clean/tune first, then shape (EQ, compress), then enhance (effects). Each processor affects what comes after. Logical order = better results with less fighting.

Style

Match Processing to Style

Heavy Auto-Tune for T-Pain, transparent correction for Adele. Intimate for Billie Eilish, powerful for Beyoncé. The 'right' processing depends entirely on the artistic goal.