Chapter 29 Exercises

The vocal chain is the most settings-dense thing this book has built, which makes this set the easiest one in the book to do wrong: copy the numbers, skip the listens, call it done. Don't. Every number in Chapter 29 was a habit, not a law, and every exercise below exists to make you earn your version of those habits on your own voice, in your own mix. You'll need: your Chapter 28 checkpoint mix (advanced pass done, bus architecture final), the raw lead-vocal comp underneath it, your marked-up lyric sheet from Chapter 11, your three reference tracks from Chapter 4, and access to your full verification fleet — studio headphones or treated-room monitors, consumer earbuds, your phone, a car, and a door you can stand on the wrong side of.

Ground rules, standing as ever. Every chain judgment happens in the full mix, at matched loudness, against the busiest section — the Chapter 22 law governs all eight slots, and the chain is where most producers quietly break it. Build the chain in one designed sitting (Part C makes this literal); judging de-essers and presence boosts on fatigued ears is worse than useless, so split the set across short sessions and respect Chapter 4's hygiene rules. Exercises marked 🎧 are focused listening; 🎛️ exercises need your DAW open. Exercises marked ★ have worked answers in the back of the book (Answers to Selected Exercises).

Part A: Conceptual Understanding

A1. The double-edged expert. ★ (10 min) Explain why the lead vocal demands more control than any other element in the mix and, at the same time, demands that the control be less audible — and why both demands come from the same fact about your listener. Then define sit precisely enough to explain Jaylen's two fader failures: why +2 dB produced karaoke and -2 dB produced a singer at the bottom of a swimming pool, and why no fader position could fix what was wrong.

A2. The chain from memory. (15 min) Without looking back: write out all eight slots of the vocal chain in the book's default order, with one job per slot and the working gain-reduction or dose range for each. Then test your understanding of order as engineering — for each of these three swaps, state what the downstream stages now receive and what audibly breaks: (a) de-esser moved to the last insert slot; (b) saturation moved before the de-esser; (c) presence EQ moved to the front of the chain, before the compressors.

A3. Plumbing versus performance. ★ (10 min) Clip gain and fader rides both move level, and confusing them is the most common structural mistake in amateur vocal mixes. Define each layer: where it operates relative to the inserts, what its job is, and which chapter owns it. State the diagnostic symptom that tells you a vocal has a clip-gain problem rather than a compressor problem ("no threshold works everywhere"). Then explain the chapter's autopsy finding: Demi's channel already had forty-some rides — so why did it still fail?

A4. The de-esser debate, argued honestly. (15 min) Write the strongest possible case for de-essing before any compression (school one), then the strongest case for de-essing after the leveler (school two) — each in three or four sentences, as if you belonged to that school. Then state what this book's default seat (after the leveler, before everything that brightens) buys from each side, and name the two situations where the default isn't enough: what do you add, and where?

A5. Serial gentleness, the arithmetic. (10 min) Two compressors doing 3 dB each deliver roughly the same total control as one doing 6 — so why don't they sound the same? Build the explanation from the chapter's 🔍 block: why a compressor's misbehavior compounds rather than scales, why every set of time constants is wrong for something, and why two stages with different time constants hide each other's errors. Finish by connecting this to Chapter 28's stacking doctrine and naming what changes when the instrument under the compressors is a human voice.

A6. The earned boost. ★ (10 min) Chapter 22 preached subtractive-first as doctrine, yet stage 5 of this chain boosts 3–5 kHz on the most exposed element in the record. Give both reasons the boost is safe here and was dangerous at the front of the chain. Then explain the dynamic-EQ option: what specific problem does a dynamic node at the presence frequency solve that the static boost cannot, and why do belted phrases create that problem in the first place?

A7. The empty stage. (5 min) Jaylen swept the 900 Hz–1.2 kHz honk zone, found nothing, and cut nothing. Defend "a chain stage is a question, not an obligation" in two or three sentences — what does every unnecessary cut cost? Then name the genre from the Project Checkpoint's adaptations where the whole chain shrinks to almost nothing, and explain why that shrinkage is a success rather than a shortcut.

A8. Support, not competition. (10 min) Unpack "darker, wider, tucked." For each of the four choices that make a lead sit — bright, dry-ish, center, dense-with-audible-detail — name the backing stack's reverse choice and the masking reason behind it. Then state the ad-lib lane rule, and describe the standing Static Bloom complication it had to resolve: who else was already answering Demi's gaps, and how did Jaylen assign the lanes?

Part B: 🎧 Listening Lab

Ears first, always: matched playback levels for every comparison, listening journal open. This Lab is also an investment — the era-aesthetics work in B1 and the quiet-test calibration in B5 feed directly into Chapter 30's troubleshooting ear, exactly as the chapter's closing promised.

B1. The two theologies, heard. ★ (45 min) Four records, two eras. Modern dry-loud-close: Billie Eilish, "when the party's over"; Lorde, "Royals." Eighties hall: Bonnie Tyler, "Total Eclipse of the Heart"; Heart, "Alone" (or Berlin, "Take My Breath Away"). For each record, log: How far from you is the voice — inches, feet, across a hall? What happens in the gap after a phrase ends — silence, or bloom? Can you hear breaths and mouth detail? Are the consonants knife-clean or softened by wash? What's manufacturing the size — reverb tail, or stacked doubles and harmonies? Finish each with one sentence: what does this record believe intimacy sounds like? Keep the notes; case study 1 is the debrief, and exercise D1 will ask you to pick a side on purpose.

B2. De-ess detection and the lisp hunt. (30 min) Train the ear that hears de-essing. First, three polished pop or R&B records: listen only to the esses. Are they present but soft? Do any esses change color rather than level (the split-band flicker)? Can you find a single ess that's louder than its neighbors — one the engineer arguably missed? Then go where the failures live: amateur mixes on SoundCloud or YouTube. Hunt both failure modes and log one timestamped example of each — untamed sibilance that stabs on earbuds, and the sock-mouthed lisp (s gone to th) of an overdosed de-esser. Hearing the lisp in the wild once will protect your own mixes forever.

B3. Backing-vocal architecture, mapped. (30 min) Pick one record with a stacked chorus you love. Draw the chapter's vocal-stage diagram (front lip to back, left to right) and seat what you hear: where's the lead? Are the doubles hard-panned? Can you hear any stack member breathe, or are they flattened? Are the harmonies darker than the lead — does anything in the stack sparkle at 3–5 kHz, and if so, does it fight the lead's diction? Then the poor-man's mono check: play the chorus on your phone speaker and note what the stack collapses into. Judge everything in context only — the whole point of the doctrine is that the stack is built to sound wrong alone and right behind a lead.

B4. The throw inventory hunt. ★ (30 min) Across five songs in or near your genre, find every momentary vocal FX event and classify it against the chapter's menu: quarter-note throw, dotted-eighth crumb, reverse swell, hall blast, telephone section, hard-tune-as-instrument, formant move. For each find: timestamp, the word or phrase, your classification, repeat count if it's a delay, bright or dark tail, and what the event is doing (punctuating a chorus exit? lifting a bridge entrance?). Then count events per song and write the rationing observation: how rarely do great records fire these weapons, and where in the song structure do almost all of them land?

B5. The quiet test, calibrated on professionals. (20 min) Play your three reference tracks, plus one of B1's eighties hall records, at conversation volume from the next room — door open, you on the wrong side of it. Log what survives of each: the vocal? the groove's skeleton? the reverb? Compare the modern dry-close vocal against the hall voice at a whisper: which one keeps its words, which one keeps its size, and what does that tell you about density versus level as the source of vocal presence? Write the one-line rule you'll apply to your own mix in C7.

B6. The conversation layer. (20 min) One hip-hop or R&B track with a living ad-lib culture. Map the ad-libs in time — which gaps do they occupy (line-ends, held notes, the bar before the hook)? — and in space: which side, how far off-center, how much darker and wetter than the lead? Count the collisions, moments where an ad-lib talks over the lead's words; on a well-mixed record you should run out of patience before you run out of zeros. Then contrast with an indie-folk or singer-songwriter record: find the texture layer at the edge of perception and note how the same craft — gaps, sides, one row back — operates at a completely different forwardness.

B7. Hard-tune, committed. (15 min) Start at the documented source: Cher, "Believe" (1998), the effect's public debut. Then one modern rap or hyperpop track that wears hard-tune as a lead aesthetic. Listen for commitment: the zero-speed snap between pitches reading as an instrument, not an error. Then find (or recall) a track where tuning sits at half-strength — audible but timid — and log the difference in how your ear assigns intent. Where does each record place the effect: lead, ad-libs, hooks only?

Part C: 🎛️ DAW Lab

The centerpiece is C2 — the chain itself, built stage by stage on your own lead vocal. Do C1 immediately before it, in the same sitting, because the teardown writes the job description the build will follow. C3 is the serial-versus-single A/B the chapter's Deep Dive called the single most instructive hour in this chapter; treat that claim as a hypothesis and test it. Save a version before every exercise. Fresh ears or no verdicts.

C1. The teardown ritual. (20 min) Save your session as a new version, then bypass everything on your lead vocal — every insert, every send, one channel-wide bypass if your DAW offers it (the Chapter 28 ritual, applied to a whole channel). Loop your busiest section. A/B the raw comp against your current processed mix at matched loudness, and write the three-gap job description: where exactly does the raw voice fail? Level evenness? Density against the dense arrangement? Esses? Placement — dry and hovering, or washed and distant? Those three gaps are your chain's job description, and every stage you build in C2 must answer to one of them. Keep the old processed channel archived; C7 ends by A/B-ing your rebuild against it.

C2. The full chain, stage by stage, checkpointed. ★ (90 min — split across two sittings if your ears fade; the chain doesn't expire overnight) Build in the book's order, and after every stage run the listen: full mix, matched loudness, busiest section, and the stage either audibly earns its slot or gets bypassed without sentiment.

1. Clip gain: phrase by phrase against the lyric sheet until no phrase jumps more than about 3 dB out of the pocket. Log your three biggest moves and what they were (a belt? a trail-off? one cracked consonant?).
2. Subtractive EQ: high-pass swept up in context until the voice thins, then backed off; sweep-and-destroy for mud (200–350 Hz) and box (400–600 Hz) — cut only what the sweep convicts, wide and shallow. The listen: clearer, not thinner.
3. Leveler: 2–4:1, attack 10–20 ms, release medium or auto, threshold so loud phrases catch 2–4 dB and quiet ones sail under. The listen: lurch gone, consonants alive.
4. De-esser: install at modest settings for now — C4 is the dedicated craft session; just get 3-ish dB on esses only, centered wherever your first sweep suggests.
5. Density compressor: faster and more opinionated — 3–6:1, fast attack and release, 1–3 dB touching most of the time. The listen: closer, not louder. If you can't hear the closeness, toggle it five times before deciding.
6. Presence EQ: +1–2 dB, wide, at the frequency your ear picks between 3 and 5 kHz; air shelf +1–2 dB if the voice wants it. The listen: more intelligible, not more bright.
7. Saturation: drive in until the voice audibly thickens, match the output, then halve the drive. You should only be able to hear it leave.
8. Sends: plate with 20–40 ms pre-delay, set in the mix until felt-not-heard; a crumb of the band's tight room; at most one throw for now.

Deliverable: your own one-page chain printout in the format of case study 2 — slot, setting, reason, the listen that approved it. That document is the artifact this chapter wants you to own.

C3. Serial versus single — the honest A/B. ★ (60 min) Duplicate your lead vocal channel. Version A: your C2 chain as built, both compressors and the clip gain intact. Version B: same clip gain and EQ, but bypass both compressors and insert one compressor alone, then chase the same total control — aim for 6–8 dB of gain reduction on the loud phrases — using any settings you like. Spend a real fifteen minutes making B as good as you possibly can; this experiment only teaches if the single compressor gets a fair fight. Bounce both versions, matched loudness, and A/B blind if you can recruit a friend or shuffle the files. Listen specifically to: the breaths in the gaps between phrases, the esses against the vowels, the stability of your longest held note, and where your attention goes during a full chorus. Journal which artifacts appeared, and at what gain-reduction depth. Bonus round, and the real lesson: lower the single compressor's workload until it is fine — there's a depth where one compressor is all you need, and knowing where that line sits on your voice is worth more than any doctrine.

C4. The frequency-find de-ess session. ★ (30 min) By sweep, not by chart. Loop your worst essy line — find it by reading your lyric sheet for s, ts, and ch clusters. Method one: narrow EQ boost, sweep 4–9 kHz until the ess turns into a knife; note the frequency; remove the boost — it was a flashlight, not a move. Method two, if your de-esser has a listen/solo mode: sweep the detection band until you hear pure ess and almost no vowel. Center the de-esser there. Now the calibrated overdose, on purpose: push the threshold down until you manufacture the lisp — listen to s become th and learn the sound of taxidermy — then back off until natural ess returns. Finish with the mode test: split-band versus wideband on your voice, in the mix; log whether you can hear the esses changing color (split-band's flicker) or the voice dipping whole (wideband's grab), and pick by ear. Target when you're done: 3–6 dB on esses only, zero lisp.

C5. The tightrope, with and without a net. (20 min) On your stage 5: A/B a static +1.5 dB presence boost against the same boost with a dynamic node at the same frequency, set to shave a decibel or two only when the band spikes. Audition both against your most belted phrase and your quietest verse line, in the mix, matched loudness. Log which phrases wanted the net and which never triggered it. No dynamic EQ in your DAW? Do the manual version: keep the static boost and write a -1 dB automation dip on the boost's gain across the belted phrases (a Chapter 27 move standing in for a Chapter 28 tool) — then check Appendix E for what your DAW does have.

C6. Glue and the single throw. (25 min) Two builds. (a) The room crumb: send your lead to the band's existing tight room — start so low the meter barely flickers. Toggle the send on and off in the full mix and journal what changes, in sticker-and-glass terms: does the voice bond to the record? Raise the send until you can hear the room as a place, then halve it — the destination is felt, not heard. (b) The throw: build or wake a dedicated delay return (quarter or dotted-eighth, repeats filtered dark), send sitting at zero. Pick the one word in your song that most deserves to ring — almost always the last word of a chorus — and automate the send to fire on that word only. If the repeats fight the next section's downbeat, automate the return down across the boundary. One throw. Rationing is the craft.

C7. The vocal-sits-everywhere protocol, run honest. (45 min) Axis one — every section: set the lead's fader once, in the densest chorus, then play the whole song top to tail without touching it. Log each failure as level (a ride or a missed clip-gain phrase — performance layer) or tone (fine in the sparse verse, wrong in the dense chorus — masking, an arrangement fight). Axis two — every system: studio headphones, consumer earbuds, phone speaker at arm's length, the car, and the quiet test from the next room. One named symptom per failure — "esses loudest thing on earbuds," "voice vanishes at whisper volume" — because named symptoms route to fixes and vague unease doesn't. Then the compass: A/B your vocal's seat against one reference, level-matched — louder, drier, brighter than theirs? On purpose? Write the ranked fix list, execute the top three, and re-verify. Final deliverable: A/B your rebuilt channel against the C1 archive of your old processing, matched loudness. The designed chain should beat the archaeology. If it doesn't, one of your stages is lying — find it with bypass tests, working backward from the last stage that won.

C8. The supporting-cast pass. (25 min) Apply the doctrine to your stack, relative to your lead's chain: high-pass higher (120–200 Hz), darken the top (pull the presence band or low-pass gently), flatten without guilt (4–6 dB of gain reduction — they're texture, not testimony), de-ess harder than feels polite (stacked esses sum), widen and wetten (hard-panned doubles, more plate, one row back), and tune the stack tighter than the lead. Ad-libs: into the gaps, 30–60 percent off-center, darker and wetter, one costume maximum. The acceptance test is the doctrine's own: solo the stack — it should sound processed, thin, almost rude. In the mix, the lead should grow wings. If soloing tempts you to "fix" the stack, close the solo button and re-read the last line of the chapter's backing-vocal section.

Part D: Synthesis & Critical Thinking

D1. Your vocal theology. (30 min) Using your B1 notes and your Chapter 5 manifesto: which theology is your record borrowing — the cathedral, the ear, or a deliberate hybrid? Write the vocal paragraph of your manifesto: where the voice lives relative to the listener, what manufactures its size, and — the part most people skip — which costs you're accepting on purpose. Cite one reference track as your exhibit, and name the one production move from the other theology you're stealing anyway.

D2. The archaeology post-mortem. (20 min) Demi's channel failed even though every plugin on it was individually defensible. But here's the tension: Part V taught you to mix in passes — EQ week, compression week, space week — which is exactly how the sediment accumulated. Was the book wrong? Reconcile pass-based mixing with chain-based design in two paragraphs: what does each approach buy, which elements of a mix tolerate accumulated passes and which (one, in particular) demand end-to-end design, and what's the trigger that tells you a channel needs a teardown rather than another adjustment?

D3. Two chains that aren't yours. (25 min) Using the Project Checkpoint's genre adaptations as a starting point, design on paper the lead-vocal chain for two genres you don't work in — slot list in order, dose ranges, and a one-sentence reason for every place your design departs from this chapter's pop default. Requirements: the folk/acoustic chain must shrink (and you must say what each deleted stage's job is now — answered by the source, or genuinely unneeded?), and the hip-hop/R&B chain must take a position on hard-tune as stage 3.5 — in or out, full strength or absent, and why.

D4. What the chain cannot install. (20 min) "Nothing in this chapter can install alive." Draw the boundary between capture problems and mix problems on a vocal: list three complaints the chain genuinely fixes (with the responsible stage) and three that are re-record territory (Chapter 11's domain), and state the test that tells you which side of the line a given complaint lives on. Then the honest cost accounting: what does pretending the chain can fix a capture problem cost you — in hours, in plugin count, and in the recording you ship?

Part M: Mixed Practice

M1. (Chapter 4) (10 min) Why does the quiet test work? Using the equal-loudness story from Chapter 4: what does your hearing physically de-emphasize at conversation volume, why does the vocal-and-groove skeleton survive when the bass doesn't, and why does a vocal that vanishes at whisper volume indicate a density problem rather than a fader problem?

M2. (Chapter 11) (10 min) Cheaper than a de-esser: name three capture-stage moves from Chapter 11 that reduce sibilance before it exists, and the cost of each. Then the judgment call: you're mixing a comp that's already printed and the esses are savage — at what point is "re-record it" the right answer instead of a third de-esser, and what non-audio factors (the performance, the schedule, the singer) weigh on that call?

M3. (Chapter 24) ★ (10 min) The throw return's math. "Static Bloom" runs at 96 BPM. Verify that the quarter-note delay is 625 ms, then compute the eighth and the dotted eighth. Then the tone question: the throw return's repeats are filtered dark — give both reasons (one about front-to-back depth from Chapter 24, one about this chapter's most contested frequency band).

M4. (Chapter 28) (15 min) Duck the rival. Design the pluck-delay duck from the chapter's last fix: what's the key source, what gets keyed, how much gain reduction are you targeting, and — the aesthetic dial — what release character do you choose here, audible pump or transparent duck, and why? Close with the principle: what does ducking fix that carving a static EQ hole in the rival cannot, and when would you still choose the static carve?

M5. (Chapter 21) (10 min) The plumbing audit. In gain-staging terms, explain why clip gain must live before the inserts while the ride lane lives after the whole chain — what does each placement guarantee about what the processors see? Then the housekeeping nobody skips twice: after a chain rebuild that added makeup gain in two compressors and a saturator, which two meters do you re-check before declaring the pass done, and what's the Chapter 21 contract they have to honor?

Selected answers and worked examples: see the Answers appendix. If you only have one evening, spend it on C2, C3, and C7 — the chain, the proof, and the verification. Everything else in this set exists to make those three honest.