Chapter 28 Exercises

This chapter was really two chapters holding hands — the relationship tools, then the architecture that holds them — so the set splits its weight accordingly. Part A makes sure the detector logic and the three bus species actually stuck. Part B retrains your ears on records that made these moves famous, and on mixes that hide them completely (both outcomes are the technique working). Part C is the evening that matters: the duck, the crush, the M/S audition, and the ritual that decides what survives. C1, C3, and C7 are load-bearing — Chapter 29 builds the vocal chain inside the architecture you finalize here, and the duck and crush you build tonight are assumed working by next chapter's first page.

You'll need: your automated mix from the Chapter 27 checkpoint (rides written, lifts drawn, session stable), your three Chapter 4 reference tracks, an EQ with an M/S mode and a compressor with a key input (most stock bundles include both — Appendix E translates the control names across DAWs), the correlation meter from Chapter 25, and your listening journal.

Two laws govern everything below. First, the one you've now obeyed for six chapters: matched-loudness bypass A/B on every decision. Second, the one this chapter added: the bypass-test ritual is the exit door for every Part C exercise — full mix, busiest section, two passes each way, one sentence said out loud. No sentence, no insert.

Exercises marked ★ have worked answers in the back of the book (Answers to Selected Exercises). 🎧 means focused listening at your marked level; 🎛️ means the DAW is open. Time estimates are honest averages. Spread Part C across at least two sittings — architecture decisions made on fatigued ears are the ones you re-litigate in March.

Part A — Conceptual Understanding

A1. The two-part machine. ★ (10 min) From memory: every dynamics processor in this book is a detector plus a gain element. Write one sentence defining each half. Then fill in the family tree — for each of these five processors, state what the detector listens to and what the gain element acts on: (a) the ordinary Chapter 23 compressor; (b) the kick/bass ducker; (c) a 2-bus compressor with its sidechain high-pass filter engaged; (d) a dynamic EQ band; (e) the ghost-trigger pump. Finish with the sidebar's one-line moral about detectors and the book's first law.

A2. The axis diagnosis. (15 min) For each fight below, name the axis it's on and the lowest-rung tool that solves it — static EQ, dynamic EQ, sidechain compression, parallel processing, or M/S — with one line of reasoning each: (a) the rhythm guitar and keys overlap at 250 Hz all song long; (b) the bass line is perfect except one note that leaps out every fourth bar; (c) the kick loses its knock whenever the 808 sustains under it; (d) the drums need their transients untouched and a thick, sustaining floor at the same time; (e) the mix's edges are muddy as a group while the centered kick and vocal are clean; (f) the vocal plate smears the lyrics during phrases, but the gaps feel too dry at any lower send level.

A3. The aesthetic dial, reconstructed. ★ (15 min) Rebuild the transparent-duck versus audible-pump table from memory: job, ratio, attack, release, gain reduction per hit, what you hear, where each lives. Then do the math row honestly: at 96 BPM, compute the quarter-note and eighth-note durations and explain why a 90 ms release reads as transparent there while a 300–400 ms release reads as a surge. Recompute both note values at 124 BPM and state what happens to your release choices — and which end of the dial a four-on-the-floor arrangement invites, and why.

A4. Why punch survives the crush. (10 min) Explain parallel compression's blend math in your own words: why the loud moments of the summed signal are dominated by the dry path, why the quiet moments are dominated by the crushed path, and why the result earns the name upward compression. Then answer the sharp-reader question: a snare passes a 4 dB punch compressor, then 2 dB of drum-bus glue, while a parallel crush runs 10 dB beside it — how much serial gain reduction did the listener's snare transient actually pay, and where did the 10 dB go?

A5. M/S plumbing. (10 min) Write the encode and decode equations, and list what typically lives in M and what lives in S for a modern pop mix. Then answer: why is the mono fold-down of a mix the M channel alone, and what does that imply about every decibel of level you add to S? Classify each as an L/R problem or an M/S problem: a too-bright guitar panned 70% left; muddy low-mids across all the wide pads and stereo reverbs; a lead vocal that needs half a decibel of presence without brightening the doubles at the edges.

A6. Three species, three questions. ★ (10 min) Name the three species of bus and the question each answers. For each: what happens when you mute it mid-song, and what does its fader mean? Then diagnose the species confusion in each setup and prescribe the fix: (a) a reverb inserted directly on the GTR group; (b) a "parallel drum bus" built by duplicating ten drum tracks; (c) a return track that somehow carries dry signal.

A7. The stacking doctrine. (10 min) "2 dB three times beats 6 dB once." Give the three compounding reasons the gentle stack wins. Then name the doctrine's failure mode — the one with no single culprit — and the sixty-second audit that catches it. Finish by identifying which lane in the "Static Bloom" architecture runs double-digit gain reduction without violating the doctrine, and exactly why it's exempt.

A8. The tax return. (15 min) List four costs of the complexity tax that never appear on a CPU meter. Then write the bypass-test ritual's five steps from memory, including the exact question step three asks — and a two-sentence explanation of why "can I hear it?" is the wrong question and which bias it flatters. End with the rule for an insert that can't state its sentence: both halves, the deletion and the MIX NOTES exception.

Part B — 🎧 Listening Lab

Reference playlist open, marked listening level from Chapter 8, journal ready. This lab's skill: hearing relationships — level that moves because another sound moved it.

B1. The French-house session. ★ (30 min) You ran a pump safari in Chapter 23 to hear the surge; tonight you run it as an engineer. Cue Daft Punk's "One More Time," Stardust's "Music Sounds Better with You," and Modjo's "Lady (Hear Me Tonight)" — or any filter-house tracks you love. For each, work past the hook and log: (a) what ducks — the whole sustained layer, or only part of it? (b) the release length, estimated by ear against the tempo — does the swell crest right before the next kick, or come home earlier? (c) what happens to the breathing in the breakdown, when the kick drops away; (d) the depth — closer to a few decibels of motion, or closer to "the floor is inhaling"? Deliverable: a four-line pump anatomy per track. Pair this session with case study 1 — this lab is that story's evidence locker.

B2. The transparent-duck hunt. (20 min) Now the opposite end of the dial. Cue two modern hip-hop or pop records whose low end you admire — kick and 808 sharing the bottom octaves all song. Loop the busiest eight bars of each on your best headphones and listen only to the hand-off: does the kick's knock stay articulate when the 808 sustains under it? Can you catch the 808 dipping — or only the result, a low end that's huge and articulate at once? Be honest in the journal: you can't audit their sessions, and the same polish can come from a duck, a keyed dynamic band, or immaculate arrangement — that ambiguity is the lesson. Log confidence, not certainty. You're training the detector.

B3. Parallel-crush spotting. (25 min) Cue one recording from the minimal-processing era — the jazz kit from Chapter 23's lab works, or any sparse acoustic recording where the drums' dynamics arrive roughly as played — and one modern rock or pop record whose drums sound expensive. In the second, listen for the parallel signature: transients that still crack at full size while the between-hits layer — ghost notes, room decay, cymbal bloom — sits impossibly loud and dense. That pairing (peaks intact, valleys lifted) almost never happens in nature, and downward compression alone can't manufacture it. Log three moments where quiet detail sits louder than physics suggests, then one sentence on what the older recording keeps and gives up by comparison.

B4. M/S width listening. (25 min) Most M/S-capable EQs and many metering plugins offer mid and side solo — find yours (Appendix E). Run all three reference tracks through it. Solo the sides and inventory what lives out there: doubles? hats? reverb? pad spread? Then listen specifically for how much low end the sides carry. Then solo the mid and notice how complete the song still is — that's the fold-down your phone-speaker listener gets. Journal per track: three things found in the sides, the side low-end verdict (clean, slight warmth, or mud), and whether the mid alone would survive as a record. Expect the pro mixes to keep their sides bright, busy, and nearly bass-free — and now you'll know it's a choice, not luck.

B5. Duck-detection under vocals. (20 min) Cue two vocal-forward pop or R&B ballads with audible reverb or delay tails. Listen across phrase boundaries: does the space swell the instant the phrase ends and tuck itself away when the next phrase starts? Log three bloom moments with timestamps and a size guess. Then write the honest caveat this chapter taught you: that move can be built two ways — Chapter 27's drawn send rides or this chapter's vocal-keyed duck — and from the listener's seat you cannot tell which. What you can log is the musical function: size in the gaps, clarity on the words. The function is the lesson; the implementation is a workflow choice.

B6. The breakdown tell. (15 min) In dance music, the pump's key source leaves fingerprints. Find one track where the breathing continues through a breakdown with no audible kick, and one where the pump stops the moment the kick does. What does each imply about what the detector is listening to? Connect the first to the ghost-trigger trick, and journal one sentence on why a producer would want the room to keep breathing through the drop.

B7. The processed-and-alive audit. (15 min) Final calibration before Part C hands you more tools. Play ninety seconds of your densest, most maximal reference, then ninety seconds of a sparse acoustic record you love. Both are professional; one is heavily managed, one is mostly left alone. Journal: which one your ears relax into, where your attention drifts in each, and one honest sentence about where your own mix sits on that spectrum tonight — before the ritual moves it.

Part C — 🎛️ DAW Lab

Everything below happens on your own mix, continuing from the Chapter 27 checkpoint. Save a versioned copy before C1 and after C7. The exit door for every build is the ritual: full mix, busiest section, matched loudness, two passes each way, one sentence or deletion.

C1. The transparent duck. ★ (30 min) The chapter's centerpiece. Build the kick/bass sidechain exactly as walked through: compressor on your bass — track or bus; if your low end is layered, audition both homes and note which one the "Static Bloom" hearing chose and why — kick routed to the key input, 4:1, fastest attack, release timed to your tempo (compute your eighth note first: 60,000 ÷ BPM ÷ 2), threshold down to 2–4 dB per hit. Verify the two diagram facts by ear: no kick audio in the bass path, and nothing audible happens until both sounds play together. Then the ritual. Deliverable: your settings, your release math, and the one sentence the duck earned — or the honest note that your arrangement already pockets the low end and the duck got deleted, a legitimate outcome the chapter explicitly blesses for some genres. Defend whichever verdict you reach.

C2. The pump, then the choice. (25 min) Same routing, other aesthetic: deepen the threshold to 5–10 dB per hit and lengthen the release until the swell crests just before each kick lands. If your track isn't four-on-the-floor, build a ghost trigger — a muted track playing steady quarter-note kicks — and key from that instead; and duck the pad bed, not only the bass, so the whole sustained layer becomes a rhythm instrument. A/B the pump against C1's transparent duck at matched loudness and let your genre vote — your references are the ballot. Keep one, delete the other, and journal both verdicts with one sentence each. (Electronic producers: this exercise may be the night your genre's signature move becomes yours. Everyone else: knowing the dial exists is the point.)

C3. The parallel drum bus, from scratch. ★ (35 min) Build the crush exactly as specified: a new return named DRUM CRUSH, post-fader send from the drum bus at unity, compressor set to deliberate abuse — 8:1 or 10:1, fastest attack, fast release, threshold down to 8–12 dB of gain reduction. Solo it once to hear the absurdity; never again. Return fader from silence, full mix playing, busiest section: raise until the drums thicken — ghost notes surfacing, room swelling — and stop just before you can point at it as a separate sound. Expect 6–10 dB under the dry bus. Then the two audits: the polarity-flip phase check (deep null at equal level = aligned; thin phasey wash = fix the latency before blending) and the post-fader verification (pull the drum bus down 6 dB and confirm the crush follows). Then the ritual. Deliverable: blend level, both audit results, the sentence.

C4. The M/S session. (25 min) On your widest bus — synths, pads, or doubled guitars — insert an M/S-mode EQ. First, listen: solo the side channel and inventory the swamp. Then the workhorse move: high-pass the sides, sweeping somewhere in 120–250 Hz until the width tightens without narrowing; watch the correlation meter relax. Then audition the air shelf: +1 to +2 dB on the sides from 8–10 kHz up. Mono check after every side move — both moves, every time; that's the price of the tool. Compare the air shelf against the same shelf in ordinary L/R mode and journal what the M/S version spared — listen to your lead vocal during the A/B. The ritual judges the two moves independently: it's entirely legal to keep the high-pass and delete the shelf.

C5. The space duck. (20 min) Only if your vocal-versus-reverb fight is real — check first: loop a dense vocal phrase and ask whether the plate blurs words at the send level you actually want for size. If yes: compressor on the reverb return, keyed from the lead vocal, 2:1–3:1, attack 10–30 ms so the verb blooms onto each phrase before yielding, release 200–400 ms so the tail swells back into the pauses, threshold for 1.5–2.5 dB while the vocal sings. Then raise the send a touch hotter than Chapter 24's static setting allowed and verify you bought size without losing words. A/B against the Chapter 27 alternative — drawn send rides — and journal which you're keeping and why; the every-time-automatic test from the automation boundary is the tiebreaker. Repeat for the throw return if you run one.

C6. The dynamic EQ drill. (20 min) Find one genuinely intermittent problem in your mix — the boomy bass note, the vocal edge that only appears on the pushed choruses, the open hat that swells over the vocal's air. (If you can't find one, don't invent one: write that in the journal and skip to C7 — the ladder's lowest rung for "no problem" is "no tool.") Set the band with Chapter 22's sweep, set the threshold so the band sleeps until the offense, cap the range at 3–6 dB. Then build the static-cut alternative at the same frequency and A/B the two at matched loudness across both the problem moments and the innocent ones. Deliverable: which rung of the ladder this problem actually needed, in one sentence.

C7. The map and the ritual. ★ (45–60 min, fresh ears — this is the capstone) Two parts. Part one, the map: draw your complete architecture on one page — every family bus, every parallel lane, every return, every key line dashed — using the three-species table as your legend. If it won't fit on a page, that's a finding, not a formatting problem: merge or delete until it fits. Part two, the full ritual: every insert added this chapter, plus any earlier insert you've been quietly unsure about, gets the five steps — full mix, busiest section, matched loudness, in and out twice, the sentence said out loud. Delete the failures; park the ideas in MIX NOTES with dates. Then the gain-reduction audit: play your chorus, walk one loud element's path across the map, sum the GR meters out loud, and confirm the total matches a budget you'd defend. Deliverable: the map (photograph it into the session folder), the kill list with one-line causes of death, and the ledger for your two most-processed elements. Save versioned. Chapter 29 starts here.

Part D — Synthesis & Critical Thinking

D1. Technique as identity. Case study 1 traces how a broadcast utility became a genre's signature. Pick another production technique that crossed from tool to identity — Chapter 24's gated reverb and Chapter 15's hard-tune are both fair game, or choose your own — and write two paragraphs comparing its crossing to the pump's: what conditions let a technique stop being invisible and start being meaning? End with one sentence naming something your own genre currently treats as identity that began as utility.

D2. The tour test. A collaborator opens your session cold next month, with you unreachable. Write the one-paragraph tour your map should let them give — which lanes exist, what flows where, what the dashed lines do — without playing a note. Then the honest audit: name the one lane in your current session most likely to confuse them, and rule whether the confusion is the lane's fault (merge or delete it) or the map's (label and document it). Do the ruling's paperwork tonight.

D3. Your tax bracket. The "Static Bloom" hearing built seven candidates and kept five. Argue where your line sits, in two paragraphs, for two different release contexts: the Friday-upload beat and the single you'd pitch hard for playlists. Use T3 honestly — name specifically what each additional relationship tool buys (clarity, density, width, size) and what it costs (the four rents from A8) — and say why the budget differs between contexts, or defend the claim that it shouldn't.

D4. Reflex or choreography. Chapter 27 drew the boundary: automation is choreography you write; sidechain is a reflex you install. The space duck sits exactly on that line — both implementations produce the same audible weather, and B5 proved you can't tell them apart from outside. Write the decision criteria you'll actually use: when does every-time-automatic win, when does phrase-shaped-human win, and is there a case for running both at once on the same return? Defend each answer with one concrete scenario.

Part M — Mixed Practice

M1. (Chapter 21) After blending your crush return in at 8 dB under the dry drums, your 2-bus peaks rose by well under a decibel — and after the side air shelf they barely moved at all. Explain why adding a much quieter copy under a loud peak changes that peak so little, why you still re-verify the −6 dB headroom contract after every architecture change anyway, and name the two meters you check.

M2. (Chapter 22) A friend watches you high-pass the sides of your synth bus at 180 Hz and says, "Same result as high-passing the synth tracks at 180 — fewer steps." It isn't. Explain what the track high-pass removes that the side high-pass keeps, what kind of low-end problem makes each one the right tool, and which version "Static Bloom" needed and why.

M3. (Chapter 13) The duck made your kick and 808 coexist — so a producer friend concludes that Chapter 13's homework (choosing a kick whose knock sits clear of the 808's fundamental, tuning the 808 to the song's key) is now optional. Correct them using this chapter's own boundary: what does the duck handle, what does it merely bandage, and which Chapter 20 threshold concept is standing guard over the difference?

M4. (Chapter 4) Your pump sounds even and musical at your calibrated monitoring level, but lumpy and exaggerated when you check quietly late at night. Which Chapter 4 concept explains why a low-frequency surge reads differently at different playback levels, and what's the discipline that keeps you from re-balancing the duck for the wrong level?

M5. (Chapters 2 and 12) A forum post claims parallel-path comb filtering is "analog superstition — digital copies are identical, so they can't cancel." Untangle it: what makes two identical digital paths dangerous the moment one passes through a plugin (Chapter 2's latency story), what the polarity-flip null test actually demonstrates (Chapter 12's phase logic), and the DAW feature that makes the problem mostly historical — plus the two situations where it still bites.

Selected answers and worked examples: see the Answers appendix. If you only have one evening, spend it on C1, C3, and C7 — the duck, the crush, and the ritual. Those three are load-bearing: Chapter 29 builds the vocal chain inside the architecture C7 locks, and the quiz assumes all three live in your hands, not only your head.