Chapter 15 Exercises

Editing is a hands chapter. Reading about crossfades is like reading about swimming — the knowledge isn't real until your fingers have placed a hundred of them, so the DAW Lab is the center of gravity here, and the four-take comp drill (C1) is the single most valuable exercise in Part III. Work Part A first to load the doctrines, do the Listening Lab before the DAW Lab (you'll edit better once you've failed to hear professional edits), and keep your listening journal from Chapter 4 open throughout. Time estimates are honest averages; editing rewards patience and punishes marathon sessions, so split the DAW Lab across two or three sittings before your ears go flat. Exercises marked ★ have worked answers in the back of the book (Answers to Selected Exercises).

Part A: Conceptual Understanding

A1. Defend the two-stage filter. ★ (15 min) A friend plans to comp their fourteen takes by "listening to everything carefully a few times, then picking the best parts." Using the chapter's reasoning, write a short paragraph on why this fails, a second on why the opposite extreme (auditioning all fourteen versions of all thirty phrases) also fails, and a third explaining what the two-stage workflow — fast triage to survivors, then phrase-level assembly — preserves that both extremes burn. Include the rough audition math and the Chapter 4 concept that makes "listen to everything repeatedly" self-defeating.

A2. The seam hierarchy, ranked and argued. (15 min) From memory, list the four seam-placement rules in order of preference (gap, breath, unvoiced consonant, never-inside-a-connected-gesture). For each, write one sentence on why the signal there forgives a cut — what's physically or perceptually discontinuous at that point — and one example of a lyric moment where you'd use it. Then name the single worst place to splice two takes and explain what two performance properties refuse to line up there.

A3. Who owns the breath? ★ (10 min) You're splicing take 3's phrase A (ending) to take 7's phrase B (starting), with an inhale sitting between them. Diagram the edit in words: where exactly the cut lands, whose breath survives, and why. Then describe what a listener perceives — without being able to name it — when you get this backwards. Finish with the exception: when does a breath get transplanted from a third take entirely, and what's the rule for moving it?

A4. The retune ledger. (15 min) Build a two-column tradeoff ledger for pitch correction's two governing parameters — retune speed and correction strength. For each, write what you gain as it increases, what you spend, and the by-ear protocol the chapter gave for setting it. Then state where "bad tuning" lives on this ledger — not at either extreme, but where — and why the chapter calls that zone a failure of both craft and candor.

A5. The grid memo. (10 min) Jaylen's acoustic guitar reads as 20 ms "late" against the DAW's editing grid, but Chapter 13 deliberately parked the Static Bloom drums behind the grid. Write the three-sentence memo to a collaborator explaining why snapping the guitar to the grid would make it rush, what the correct timing reference is, and the chapter's one-line doctrine that covers every case like this.

A6. The elastic-audio bill. ★ (10 min) Name the three recognizable artifact families of time-stretching (what happens to transients, to sustained vowels, and to everything at extreme stretch amounts), which source material gets punished worst and why, and the doctrine that decides when stretching is acceptable. Then name one genre context where the worst artifact on your list is the aesthetic, and explain what changes ethically and practically when the artifact is chosen instead of suffered.

A7. Write the ghost story. (15 min) Every working engineer carries a session ghost story — the irreversible edit that came back to haunt. Invent a plausible one in three or four sentences (the flattened comp, the deleted take, the consolidated file with the click baked in), then write the three damage-control habits that would have prevented it and the chapter's argument for why reversibility exists to fund boldness, not to indulge fear.

A8. Continuity versus the highlight reel. (10 min) The chapter warns that "a comp sheet that looks like a checkerboard is usually a comp that sounds like a séance." Explain the failure mode in performance terms: what does a listener experience when nearly every consecutive phrase comes from a different take, even when every individual phrase is the best available? Name the Chapter 11 concept about full takes that explains why long runs from one take carry something no sequence of winners can, and state the chapter's comping priority in one line.

Part B: 🎧 Listening Lab

The chapter's closing promise, kept: this lab sends you hunting for edits on records you love, and you are going to lose. That's the design. Headphones, full attention, journal open — and matched listening levels when you compare anything to anything (Chapter 4's law doesn't pause for editing).

B1. The comp-seam hunt — you will fail, and the failing is the lesson. ★ (40 min) Pick three commercially released vocal performances you know intimately — your own favorites first, and if you want suggestions in the modern-pop lane where phrase-level comping is standard industry practice, try Adele's "Hello," The Weeknd's "Blinding Lights," or Ariana Grande's "thank u, next." Your mission: find the comp seams. Listen phrase by phrase at a comfortable level, then loud, then quiet. Log every spot where you suspect a seam — a breath that feels placed, a word whose energy steps, a vowel whose color shifts mid-line. Be paranoid. Accuse freely. Then read your suspect list back and grade your confidence honestly: can you actually hear an edit, or are you inferring that one must exist because the chapter told you they're everywhere? Almost everyone ends this exercise with a page of suspicions and zero convictions — and that's the finding. Write the journal entry: what does it mean that performances assembled from dozens of pieces survive your most hostile listening? Connect it to the chapter's claim that the work is invisible by definition — and to what that demands of your own crossfades. Then, as a consolation prize with a real answer: cue up The Beatles' "Strawberry Fields Forever" and listen closely right around the one-minute mark, where two performances in different keys and tempos were famously joined with a razor blade. Most listeners sail past it; knowing it's there, you may catch the song's color darkening. Sixty years of listeners, one of recording's most audacious splices, and it still mostly hides. That's the bar.

B2. The hard-tune spectrum. ★ (45 min) Build a playlist that walks the entire aesthetic range of pitch correction, from can't-point-at-it to proudly synthetic. A working lineup, in roughly the order you should audition them: Adele's "Easy On Me" and Billie Eilish's "everything i wanted" (the transparent end — if correction is present, you can't locate it, which is the point); Post Malone's "Circles" and Juice WRLD's "Lucid Dreams" (tuning as a light gloss the genre expects); Travis Scott's "goosebumps" and Future's "March Madness" (the robot as full texture); Kanye West's 808s & Heartbreak — try "Heartless" — (hard-tune as an emotional instrument, the machine voice carrying grief); Lil Wayne's "Lollipop" and T-Pain's "Buy U a Drank (Shawty Snappin')" (the signature era, pitch quantization as lead-vocal identity); Cher's "Believe" (the origin — listen for the flicker on the verse lines); and Bon Iver's "Woods" (the voice pushed all the way into being a synthesizer). For each track, log: can you hear correction at all? Where — sustained notes, note transitions, fast runs? Does vibrato survive? Do you hear the stairstep snap between pitches? Then rank the tracks along the spectrum and defend two of your placements in a sentence each. Finish with the genre-contract question: pick the two tracks at the extremes of your ranking and write what each one's vocal promises its audience — and what would break that promise.

B3. The pulse autopsy: hearing over-quantized drums. (35 min) Three listens, one diagnosis skill. First, Led Zeppelin's "When the Levee Breaks" — a human pocket so heavy it built half of hip-hop; listen to where the snare sits against your foot-tap and how the feel breathes around the fills. Second, D'Angelo's "Untitled (How Does It Feel)," where the famously dragged, behind-the-beat feel — Questlove has described the Voodoo sessions' deliberately loose pocket in many interviews — is the entire point; notice how "late" reads as deep, not sloppy. Third, pick a mainstream rock or metal track from the 2000s onward in your own library and run the over-quantization tells from the chapter: does the backbeat ever waver? Do fill entrances land with identical machine confidence every time? Does the hat pattern breathe, or does it sound like the same bar photocopied? Does the drummer ever get audibly excited into a chorus? (You're diagnosing, not convicting — some bands play that tight live. The skill is hearing the question.) Then the control test: pull up your own Chapter 13 exports — the flat-grid version versus the pocketed version of your Static Bloom drums — shuffle them blind, and verify you can identify which is which in under ten seconds. Log what gives it away fastest.

B4. The breath census. (25 min) Three vocals, three breath cultures. Billie Eilish's "when the party's over": count the audible breaths in one verse and chorus and describe what they do to intimacy — that's a production keeping breaths on purpose. A maximal modern pop chorus of your choice (Dua Lipa's "Don't Start Now" works): how many breaths survive, and at what level relative to the voice? A hard-tuned melodic rap track from your B2 playlist: are breaths present at all, or chopped and flipped into texture? Write the genre-contract conclusion: what does each treatment promise the listener, and which treatment does your current project's genre expect? File the answer where you'll find it during your C4 pass.

B5. The audible edit, on purpose. (20 min) The chapter's rule is inaudible or intentional — now collect the intentional end. Listen for edits designed to be heard: the stuttered vocal chops in Skrillex's "Scary Monsters and Nice Sprites," the sliced-and-resequenced vocal hooks all over modern hyperpop (100 gecs' "money machine" is a dense tour), tape-stop and stutter transitions in any current pop or trap playlist. For each audible edit you catch, log what makes it read as style rather than mistake — level, placement, rhythm, repetition. Derive the pattern in one sentence. (You'll find it's the same answer the chapter gave for hard-tune: commitment. Half-measures read as accidents; full measures read as choices.)

B6. Silence isn't silent. (20 min) Cue an intimate, sparse recording — Bon Iver's "Skinny Love" or any close-mic'd folk track you love — and listen only to the gaps between phrases: room tone, string noise, breath, chair, air. Then cue a dense modern pop or trap track and listen to its gaps: often digital black, and nobody misses the room. Now play one of your own Chapter 11 vocal takes and audit its gaps the same way: what lives in there (click bleed, rustle, the room), and — given your genre's row in the chapter's table — should your cleanup pass mute it, fade through it, or leave a bed of it? Write the one-line policy you'll apply in C6.

Part C: 🎛️ DAW Lab

The hands-on half. Appendix E translates any control name (take lanes, comping mode, elastic/warp/flex, strip silence, edit groups) into your DAW's dialect. Save a new session version before you begin — that's not a suggestion, it's the chapter.

C1. The four-take comp drill. ★ (90 min, the chapter's centerpiece) You need four full takes of one verse-plus-chorus of a song. Best source: your own Chapter 11 session — pick four takes from your 4+ (if you recorded more, triage down to four first, which is the drill's stage one running at full scale). No Chapter 11 takes? Record four passes tonight — even rough ones on a phone-adjacent setup work, because the drill is about deciding, not fidelity — or pull a multitrack with multiple vocal takes from Mike Senior's free Cambridge Music Technology library (cambridge-mt.com). Then, with a timer running, execute the chapter's workflow exactly: Stage one — triage (20 min). One real-time pass per take, no rewinding, marking ★ / • / ✗ per song section on a printed lyric sheet. Rate the performance, not the problems: star the alive-but-flawed, x the clean-but-dead. Stage two — phrase-level assembly (45 min). Loop phrase one, audition it across the four lanes, pick at conversation speed, mark the comp sheet, move on. A (?) on hard calls; no agonizing. When the sheet is full, read its shape: if it's a checkerboard, re-comp for runs — long stretches from single takes, switching at section boundaries, transplants only where decisively better. Stage three — in context (15 min). Play the comp top to bottom twice. Fix every (?) by asking which choice connects — does the energy hand off, or does the singer teleport? Stage four — dress the seams (10 min). Crossfade every boundary per the hierarchy: gaps first, breaths attached to the phrases they launch, consonant cuts faded short. Zoom in on every seam. Deliverables: the filled comp sheet (photograph it), a bounce of the comp, and a three-sentence debrief — which phrase was the hardest call, which take surprised you after triage, and where your comp's shape ended up between quilt and checkerboard.

C2. Crossfade surgery drills. (40 min) Build the reflexes at the waveform level. Drill one — manufacture the click: cut a sustained vocal note or synth pad mid-vowel, butt-spliced, no fade. Play it loud. That tick is the speaker-cone teleport from the chapter's diagram — learn its sound so well you can identify it in a moving mix. Now crossfade the same cut and hear it vanish. Drill two — the length ladder: place three edits — one in a gap, one inside an unvoiced consonant ("t," "s," "k"), one mid-sustain — and try fade lengths of roughly 3 ms, 10 ms, 30 ms, and 80 ms at each location. Log which lengths disappear and which announce themselves (a dip, a swell, a doubled-room flutter) at each seam type. Drill three — the curve check: on the mid-sustain edit, compare your DAW's equal-power crossfade against a linear one and describe the difference (one holds loudness steady; one dips). Drill four — the future-compression reveal: take a seam you believe is clean, raise the clip gain of the surrounding quiet material 8–10 dB, and listen again — this simulates what Chapter 23's compression will do to everything you edited. Any tick that surfaces now would have surfaced then, three chapters from now, at the moment of maximum embarrassment. Fix it now.

C3. Transparent versus hard, same phrase. ★ (45 min) Duplicate one comped vocal phrase — ideally one with a sustained note and some motion — onto two tracks. Track A, the documentary: run the chapter's transparent protocol. Bypass correction and listen until you can name exactly where the phrase hurts (often three notes, not thirty). Engage correction at slow retune speed and moderate strength, raise both only until the hurt stops, then back off one notch the moment you can hear the correction working — the flicker on note entries, vibrato starting to stiffen. Track B, the costume: retune speed at or near zero, strength at maximum, scale locked to the song's key. Listen for the snap between notes, the flattened vibrato, the synthetic sheen. Then the forbidden middle: deliberately set Track A's speed fast enough that correction flickers audibly on the loud notes but isn't committed enough to read as an effect. Bounce all three, label them cryptically, and listen blind tomorrow. Deliverables: the settings at which Track A first became audible (your personal flicker threshold on this voice), and one sentence on what the middle version feels like — that feeling is the sound of unowned tuning, and your job henceforth is never to ship it.

C4. The breath-management pass. (30 min) On your C1 comp (or any recorded vocal): drop a marker on every breath, top to bottom — the census usually surprises people. Then apply the doctrine: pull breaths down 6–10 dB with clip gain or per-clip volume handles until they read as punctuation rather than scuba commentary; delete only the gasps that serve nothing; and keep — at nearly full level — the run-up inhale before the biggest phrase, because that breath is part of the drama. Now render the cautionary version: duplicate the track and delete every breath. A/B the two. Log the android effect in your own words — what goes missing that you can't quite name — and your final tally: breaths kept, reduced, cut. If your genre strips breaths as an aesthetic (hard-tuned rap, slick maximal pop), do the reduce-don't-delete pass anyway, then make the strip a deliberate second move, so you've chosen it instead of defaulted into it.

C5. Tighten without flattening. ★ (45 min) Take your comped vocal (or recorded guitar) against your track's drums. Step one — map before touching: loop six to eight consecutive phrases and, by ear at performance tempo, estimate each phrase's offset from the pocket — early, late, sitting — in a sketch like the chapter's timing-offset map. Step two — classify: for each off-center phrase, ask the diagnostic question: does the deviation repeat across takes and serve the feel (intention), or does it stick out from the performance's own pattern (accident)? Step three — move only the accidents: snapping OFF, slide whole phrases by ear until they sit; resist carving into syllables (if a phrase genuinely needs internal correction, log it as a candidate for elastic audio and ask first whether a different take of that phrase comps in cleaner). Step four — verify the fingerprint survived: redraw the map after the pass. If it's a flat line of zeros, you flattened the performance — undo and try again with smaller convictions. Deliverables: both maps and a bounce. The before/after maps are the exercise; they keep you honest in a way your ears, mid-session, won't.

C6. The full cleanup pass. (40 min) Run the chapter's checklist on one complete track, in order, after comping and timing are settled. (1) Gaps: strip-silence with a conservative threshold, then hand-check every cut it made — listen for beheaded soft entrances and amputated word-tails, and fix them by extending clips and re-fading. Apply your B6 policy: cut to black, fade through, or leave a room bed, per your genre. (2) Mouth noise: find two clicks or smacks and excise each with a tiny cut-and-crossfade (a few milliseconds, faded both sides); if one is welded onto a vowel, try your DAW's or a restoration tool's de-click and listen for the smeared-lisp watermark of overprocessing. (3) Plosives and thumps: high-pass or clip-gain just the offending syllable, never the whole track. (4) The fades audit: zoom through the entire session and put a fade on every clip edge that lacks one — every chop, every bounce, every consolidated region. Count how many naked edges you found and report the number in your journal without shame; everyone's first audit is a massacre.

C7. The group-edit discipline (and the sound of breaking it). (25 min) Take any multi-mic or stereo-pair source — your Chapter 12 DI-plus-amp bass, a stereo-recorded guitar, or multitrack drums from the Cambridge MT library. First, break the law on purpose: nudge ONE channel of the pair 5–10 ms while its partner stays put. Listen in stereo (often "almost fine"), then sum to mono and hear the comb-filtering hollow-out the chapter promised — the delayed-fuse mistake that bites on phone speakers. Undo. Then install the law: create an edit group linking the related tracks, make one cut and one slide across the group, and confirm every member moved as a welded unit. Log the mono before/after in one sentence each. This drill costs twenty-five minutes once; the mistake it prevents is nearly undiagnosable three chapters later.

C8. The warp tour and the damage-control ritual. (30 min) Part one — earn the artifacts: take one sung phrase and time-stretch it by increasing amounts (start small, then go absurd). At each step, name what you hear using the chapter's vocabulary: transients smearing into "fff," vowels going watery, then full underwater warble. Find your personal threshold — the largest stretch on this voice that stays genuinely inaudible — and write it down as a habit, not a law. Try the same ladder on a drum loop and confirm it punishes stretching worse, and faster. Part two — the ritual: finish the session the way every edit session should end: save-as a new version with a stage tag (_edit or your Chapter 6 convention), confirm every losing take is archived-muted rather than deleted, and if you consolidated or rendered anything, verify the pre-commit originals sit muted underneath. Thirty seconds of ritual; it's the funding for every bold edit you'll ever make.

Part D: Synthesis & Critical Thinking

D1. The editing ethics position paper. ★ (40 min) Write the one-page editing policy for your own current project — the document Theo effectively gave Demi. It must contain: (1) your genre's contract, stated in one line (steal from the chapter's table or write your own row); (2) your lines, drawn on purpose, before the polish pass — what gets comped, how far timing tightens, transparent or hard tuning, what happens to breaths, and at least one thing you will deliberately keep imperfect; (3) the artist-consent clause — who hears the edit before it's final, and who holds the veto (if you're producing yourself, write the clause anyway and name the future listener who keeps you honest); (4) a disclosure stance — if someone asks "is that one take?", what do you say? Pressure-test your policy against the chapter's hardest argument: the un-edited take you'd romantically prefer is also a construction — one arbitrary evening, one mic's opinion, one room's acoustics. Then state where, for you, a performance stops being testimony and becomes product — and own the answer.

D2. The collage conversation, rehearsed. (20 min) A singer you're working with hears their comp for the first time and says, more hurt than angry: "Nobody ever sang that. That's not me." Record a three-minute voice memo of your actual reply. It must be honest (no "everyone does it" hand-waving), it must include what Theo told Demi about choosing versus generating, and it must end by handing the singer real power — the equivalent of Demi's "keep the crack" condition. Listen back to your memo and grade it on one axis: would you trust the person speaking with your own voice?

D3. Your edit-pass SOP. (25 min) Build your personal standard operating procedure for the editing stage: an ordered checklist from "open session" to "version-saved, ready for arrangement," with a time budget per stage and an exit criterion for each ("comp complete when: the sheet has runs, every seam is faded, two full-context passes made"). The chapter's order — comp, then timing, then tuning, then cleanup, then fades audit, then version — is the default; if you deviate (some engineers tune before timing, some clean as they comp), write one sentence defending each deviation. Cap the whole document with the chapter's tiebreaker rule, verbatim, as the final checkbox: if you can hear the edit and you didn't choose to, it isn't finished.

D4. The editing premortem. (20 min) Three chapters from now, compression will raise everything quiet; five chapters from now, the mix will expose every shortcut. Predict, in writing, the three most likely editing failures of yours that Part V will uncover — be specific to your habits ("I rushed the fades audit on the chopped FX," "my breath pass only covered the first chorus," "I left the doubles loose because the lead sounded fine solo"). Seal the prediction in your journal, dated. Chapter 30's troubleshooting pass grades it. A correct prediction now is a mix problem prevented then — and an incorrect one teaches you what you under-worry about, which is rarer knowledge.

Part M: Mixed Practice

Interleaved retrieval from earlier chapters — this is how the doctrines become reflexes instead of trivia.

M1. (Ch 1 + Ch 15) Two connected questions about transients. First: why does the chapter's drum-editing doctrine prefer cut-and-slide over time-stretching, in terms of what Chapter 1 said the transient carries? Second: explain the butt-splice click using Chapter 1's pressure-wave picture — what does an instantaneous waveform jump ask the speaker cone to do, and why does the ear flag it even at low level?

M2. (Ch 2 + Ch 6 + Ch 15) "Non-destructive editing" is the legal foundation for this chapter's boldness. Explain what actually happens to the audio files on disk when you cut, fade, comp, and warp in a modern DAW (Chapter 2 and Chapter 6 both contributed here), why that makes the three damage-control habits possible, and name the one editing action in this chapter that genuinely can forfeit the grace — and the habit that protects you when you do it.

M3. (Ch 4 + Ch 15) Editing is applied psychoacoustics, twice over. First: explain why seams hide inside unvoiced consonants using Chapter 4's masking. Second: explain why the triage stage trusts the first fast reaction to a take, using what Chapter 4 taught about listening as a perishable resource — and what that implies about scheduling a comp session at the end of a long studio day.

M4. (Ch 7 + Ch 10 + Ch 11 + Ch 15) ★ "Capture discipline is comping's down payment." Cash that sentence out: name three specific capture-stage disciplines from Chapters 7, 10, and 11 (think mic distance and the tool that enforces it, room consistency, and session structure) and, for each, the exact editing-stage misery that skipping it creates. Then explain why no crossfade, however well-placed, can hide a seam between a take recorded at four inches and a take recorded at ten — which Chapter 7 behavior is the unfixable culprit?

M5. (Ch 9 + Ch 13 + Ch 15) Quantization appears three times in this book so far: MIDI quantize (Chapter 9), groove and swing in drum programming (Chapter 13), and audio quantize-to-groove (this chapter). State the single taste principle that governs all three, explain what "quantize the shaky tom overdub to the drummer's own kick-and-snare feel" means and why it beats grid quantize, and connect it to why Jaylen's edit week slid his guitar against the drums' deliberate behind-the-grid pocket instead of snapping it to the editing grid.