Chapter 17 Exercises

This chapter's skills are dosage skills — every tool here works, which is exactly why the exercises keep asking you to measure, mute, and cut. Work Part A to load the vocabulary, do the Listening Lab before the DAW Lab (you can't ration candy you've never counted on a record), and treat C7 — the restraint pass — as the assignment the whole chapter was building toward. Keep your listening journal open and your reference playlist loaded; every A/B in this set happens at matched loudness, because "bigger or louder?" is the question under everything below. Time estimates are honest averages. Exercises marked ★ have worked answers in the back of the book (Answers to Selected Exercises).

Part A: Conceptual Understanding

A1. Three passes, three questions. ★ (10 min) Chapter 16, this chapter, and Chapter 27 each own a different kind of move. Write one sentence per chapter stating the question each pass asks of the song, then classify these four moves to the correct chapter, with one clause of justification each: (a) muting the pad for all of verse one; (b) drawing a filter rise across the pre-chorus; (c) riding the vocal up 1.5 dB on two buried syllables; (d) sending the pre-chorus's last word into a long reverb.

A2. The committee interview. (15 min) You're layering a chorus lead synth that feels thin. Write the "job description" interview for the committee: what question do you ask before browsing for layers, what does each hired layer need to be able to name, and which three checks from the layering recipe protect the committee from mud and cancellation? Finish by explaining why "the same patch, duplicated" fails the interview no matter how good the patch is.

A3. Tight, wide, whispered — the ledger. ★ (15 min) Build a three-row tradeoff ledger for the vocal-double configurations. For each: what it buys, what it costs or risks, which song section it typically dresses, and one named record or artist aesthetic where you've heard it (your Part B work feeds this). Then answer: why does the wide pair specifically need real performances rather than a duplicated-and-delayed copy — what's frozen in the copy, and what does the freeze do in mono?

A4. The boundary memo. (10 min) A collaborator asks why you're "already automating" when mixing doesn't start until Part V. Write the three-sentence memo: the definition of automation as this chapter introduced it, the one-sentence boundary rule that separates this chapter's automation from Chapter 27's, and one example of each side of the line.

A5. Candy economics, defended. ★ (10 min) Your friend's beat has a vocal chop, a riser-ette, or a sparkle in every single bar — "more is more, it keeps things interesting." Using the chapter's reasoning, explain in a short paragraph why the opposite happens: name the devaluation mechanism, the budget rule of thumb and why it's a direction rather than a law, the placement doctrine (call-and-response), and the felt-not-heard test that tells you when a candy is working.

A6. The Haas bill, itemized. (15 min) Explain the Haas trick mechanically: what you duplicate, what you delay, what the precedence effect does with the result. Then itemize the bill: what physically happens when a mono system sums the two sides, why the artifact is a comb specifically, and why a real double survives the same summing that wrecks the fake. Close with the two usage rules the sidebar gave for using the trick anyway.

A7. Transition triage. (15 min) For each seam below, choose one tool from the transition inventory (or "nothing — bare cut") and defend the choice in a sentence, citing the tool's "what it says" and its bill: (a) verse into pre-chorus, sections already strongly contrasted; (b) the drop into your song's biggest chorus; (c) chorus back down into a sparse verse two; (d) the fourth section boundary in a row where you've already used a riser twice. Then state the chapter's warning about decorating every doorway.

A8. The cut as decision. ★ (10 min) "Every idea you cut is a production decision — the only kind nobody can hear and everybody can feel." Unpack this in a paragraph: what the restraint log records, the difference between cut and park (with the chapter's parked example and its destination), why the morning re-pass can't happen the same evening, and what the chapter claims your accumulated cut lists will eventually describe.

Part B: 🎧 Listening Lab

Headphones, full attention, matched levels when comparing anything to anything. You're about to discover that records you've heard a hundred times are crowded with employees you've never noticed. That's the point — and the bar your own polish should aim for: felt, not heard.

B1. Count the layers — the maximalist census. ★ (45 min) Pick two maximalist pop choruses and run a layer census on each. Strong candidates: Katy Perry's "Teenage Dream" (Case Study 1 walks this one — do your own count first, then compare), The Weeknd's "Blinding Lights," Dua Lipa's "Don't Start Now" or "Physical," Carly Rae Jepsen's "Run Away with Me." Method: loop the first full chorus; do one pass per Chapter 4 band (sub, lows, low-mids, mids, hi-mids, air) listing every distinct element you can hear in that band; then a width pass (what's center, what's edges); then a count. Rules: a layer counts when you'd notice its absence; doubles count separately from leads when you can hear the smear or the width that proves them. Log the number, then write the real finding: how many of your counted layers had you consciously heard before tonight, and how many were doing their jobs invisibly?

B2. The candy audit and the 8-bar economy. ★ (35 min) Choose one modern pop or hip-hop track you know well (Billie Eilish's "bad guy" and Ariana Grande's "no tears left to cry" are both generous hunting grounds, but anything current works). Map one verse and one chorus bar by bar, marking every ear-candy event — chops, sparkles, reverses, one-shot FX — with its bar number and what gap it occupies relative to the lead. Compute the actual candy-per-8-bars rate. Then assess the economics: do candies repeat identically or rotate kinds? Do they answer the lead or talk over it? Find the single best-placed candy in the song and write one sentence on why that gap, that moment.

B3. Throws and the delay that plays. ★ (40 min) Two hunts. First, throws: cue any two modern pop or melodic hip-hop tracks and hunt phrase-tails — the word that suddenly blooms into space and hangs while everything else stays dry (section boundaries are the high-probability zones; Travis Scott records are dense with thrown ad-libs). Log each throw: which word, which boundary, how long the tail hangs into the next section. Second, delay as rhythm: U2's "Where the Streets Have No Name" or "With or Without You" — listen to the guitar and try to separate played notes from delay repeats. Notice how sparse the actual playing is, then write the T2 conclusion in your own words: what did The Edge leave out, and what filled the space?

B4. The doubles census. (30 min) Three voices, three doubling cultures. (1) Nirvana's "In Bloom": A/B the verse vocal against the chorus vocal — what happens to the voice count at the chorus line, and what does it do to the size? (2) Any Billie Eilish ballad ("when the party's over" works): listen for the stacked, whispered intimacy — layers as closeness rather than power. (3) A modern country or pop-rock chorus of your choice: hunt the wide pair (hard-panned doubles around a center lead — collapse your player to mono if you can, and notice what narrows). For each, log: how many voices you can defend hearing, and whether the doubles read as one fat voice or a crowd.

B5. The build inside the section. (35 min) Cue three long sections and map their internal slope against the four levers (subdivision, filter, width, register). Suggestions: the long filtered passage and reopening in Daft Punk's "One More Time" (filter as form), any Avicii or Calvin Harris build (subdivision and filter firing together), and the final chorus of Robyn's "Dancing On My Own" (register and density). For each: list which levers fire, in what order, roughly every how many bars. Then the inversion: find one moment in any of the three where energy jumps because something was removed. Write down what left and what its absence did to the next downbeat.

B6. Telephone hunting. (20 min) Find three band-limited moments in your own library: a filtered intro that previews the hook, a verse or bridge squeezed to telephone width, a lo-fi section that makes the full-range return enormous (The Strokes' Is This It vocal sound is a whole-album version — "Last Nite" shows it in three minutes). For each, log what was withheld (which bands), for how long, and what the restoration felt like. One-sentence conclusion: what does the trick spend, and what does it buy?

Part C: 🎛️ DAW Lab

Save a new session version before you begin — passes 1 through 4 happen tonight, and C7 happens tomorrow morning. That split is not optional; it's the exercise. Appendix E translates control names into your DAW's dialect.

C1. Build a three-layer kick. ★ (40 min) Extend Chapter 13's two-layer recipe to a full committee: a body layer owning 50–110 Hz, a knock layer owning the 1–5 kHz presence click (high-passed so it owns nothing below), and an air layer — a short filtered-noise tick above 6 kHz for texture. Workflow per the recipe: jobs first, then browse; high-pass everything above the body; polarity-flip test all three pairings; tune the body to your song's key; balance in context against your 808 or bass. Then the honest test: mute the air layer and A/B at matched loudness — does it earn its job, or was two the right number? Either verdict is a pass; log it in the restraint format (idea/promised/cost/verdict).

C2. Tight vs wide vocal double comparison. ★ (50 min) Record three doubles of one hook line from your own track (Chapter 11 mic craft; doubles are performed tight — consonants first): two full-voice takes and one whispered take. Build three configurations and bounce each: (a) lead alone; (b) lead + tight double tucked 6–12 dB under, same pan; (c) lead + wide pair hard L/R + whisper tucked center. A/B all three at matched loudness, then collapse (c) to mono and listen for what survives. Log: which configuration suits your verse, which your chorus, and what the whisper added that you can feel but couldn't point to. No recorded vocal in your track? Run the drill on a synth lead: real re-performances on the keyboard, not duplicates.

C3. One-section energy build. (45 min) Take your longest high section (chorus or drop) and score its internal slope using the four levers, one firing roughly every 4–8 bars: a subdivision add, a filter open, a width entrance, a register lift — in whatever order your song wants, drawn from the chapter's double-chorus timeline diagram. Then add the subtraction trick: drop the kick and sub for the final two bars so the next downbeat detonates. Bounce before/after, A/B at matched loudness, and write down which single lever bought the most.

C4. The motion pass. (30 min) Draw the minimum viable automation set on your track: one filter open into your biggest peak, one close-down into your deepest valley, one send swell across your biggest boundary. Play the full song top to bottom once. Log what changed about sections you didn't touch — and check for seasickness: if you added more than the three arcs, name which one is the architecture and which is decoration.

C5. Effects as instruments, one of each. (45 min) Build three events: (a) a throw — automate a reverb send up for the last word or hit before your chorus, back to zero immediately after, tail decaying across the downbeat; (b) a delay print — set a dotted-eighth (or eighth) delay on a sparse melodic part, feedback at two to four repeats, and remove notes from the part until the delay's answers have room; (c) a telephone section — band-limit your intro or one pre-chorus to a midrange slot and let the full spectrum return at the boundary. Bounce each in context. The check for all three: do they read as events in the song, or as plugins switching on?

C6. The transition and candy budget pass. (40 min) Deploy from the inventory on its budget: one riser maximum across the whole song (you built one in Chapter 14), one downlifter after a peak, one moment of half-beat-to-full-bar silence at your strongest contrast (with Chapter 15 hygiene fades), and a fill where the drummer should turn the page. Then candy: place at most one per 8 bars, in the lead's gaps, no two adjacent candies the same kind. Mute-test every single placement. Count what you placed versus what you considered — the ratio is your first restraint data.

C7. The restraint pass. ★ (30 min tonight + 30 min tomorrow — the chapter's centerpiece) Tonight: bounce the polished version from C1–C6. Tomorrow morning, fresh ears: open the session and remove three things — not the three worst, the three you're least sure about. Mute them, bounce the leaner version, then A/B the two bounces at matched loudness, ideally blind (shuffle them in a playlist, eyes closed). Verdict per removal: did the song get poorer (restore it — it was felt-not-heard support), or cleaner (log the cut)? Fill in the restraint log for all three: idea / what it promised / what it cost / verdict. Finish with one sentence: which removal surprised you, and what does that teach you about your evening ears?

C8. Augment a drum. (35 min) Take a recorded or looped drum element (your Chapter 12 kit, or a top-loop's snare) and reinforce it with a sample: align transient-to-transient at sample zoom, polarity-flip test, tune the sample toward the original, split the jobs (one owns body, the other crack), and blend from underneath until the chorus knocks. A/B augmented vs original at matched loudness. If your track is fully programmed, run the reverse: record real hand-claps (three loose takes) and layer them over your programmed backbeat — the humanity direction of the same move.

Part D: Synthesis & Critical Thinking

D1. Minimal is also polished. (25 min) "bad guy" is sparse; "Teenage Dream" is a thirty-employee payroll. Both are immaculately produced. Write two paragraphs arguing that polish is not density: identify what the sparse production polishes instead of layering (space, candy timing, the few sounds' design), and use the felt-not-heard doctrine to explain how both records can be "full of production" while one stays nearly empty. Land on a definition of polish that covers both.

D2. Polish in a foreign dialect. (25 min) Design — on paper — a polish pass for a genre you don't make. Pick country ballad or minimal techno. For each of the five families (layers, doubles, automation, effects-as-instruments, candy/transitions), write one move that fits the genre's expectations and one move from your home genre that would violate them. You're previewing Chapter 18: conventions are contracts, and this chapter's tools are spent differently in every dialect.

D3. The fakery line. (20 min) Drum samples under live drums, whispered ghosts under synths, octave copies hiding their artifacts, Haas width from one take. A listener might call all of this deception. Drawing on Chapter 15's editing-ethics conversation, write a paragraph locating your own line: which of this chapter's moves are craft, which (if any) feel like cheating to you, and whether the felt-not-heard doctrine makes the question better or worse. There's no rubric answer — but defend your line with the chapter's vocabulary.

D4. Reverse-engineer the restraint. (25 min) Pick one producer or production team whose records sound unmistakably theirs. From three of their tracks, infer their cut list: name three moves this chapter teaches that they conspicuously don't use (or use once where others use ten). Write the inferred restraint log — and one sentence on the chapter's claim that the cut list is the style, tested against your evidence.

Part M: Mixed Practice

M1. (Chapter 12) The flipped clap. (10 min) Your layered clap-plus-snare composite from this chapter's recipes sounds thinner together than either alone. Name the problem, the five-second test, and the distinction between the switch you flip and true phase — then state which step of the layering recipe exists to catch this before you hear it.

M2. (Chapter 4) The quiet sub lie. (10 min) You balanced your three-layer kick at low monitoring volume and the body layer vanished on a louder system check. Which Chapter 4 curve explains the lie, what does it predict about balancing lows quietly, and what listening habit defends every layering decision you'll ever make?

M3. (Chapter 10) The room votes on your committee. (15 min) Your sub layer sounds perfect at your desk and bloated near the back wall. Explain what's happening in room-mode terms, why your listening position may be lying about the layer's level, and the two-system verification (from your translation habits so far) that protects the blend decision.

M4. (Chapter 14) Rebuild the riser. (20 min) From an init patch, rebuild a riser using Chapter 14's tools: noise oscillator, filter cutoff climbing (envelope or drawn automation), optional pitch rise. Print it at your tempo, four bars and two bars long. Then apply this chapter's ration: it gets deployed at one doorway. Which one, and why that one?

M5. (Chapter 6) Sends, inserts, and the throw. (10 min) Explain why a throw is built on a send to a shared return rather than an insert reverb on the vocal track: name the routing concept from Chapter 6, the practical advantage when five elements need occasional access to the same space, and what automating the send (versus the insert's mix knob) keeps independent.