Chapter 7 Exercises — Microphones

These exercises are where placement stops being a chapter and starts being a reflex. You do not need a "good" microphone for any of them. A USB mic counts. A phone counts. The skills you're drilling — distance, axis, height, rejection — transfer to every capsule you will ever own, which is precisely the point of the chapter you've come from.

Work in this order if time is short: one exercise from Part C (the DAW Lab teaches your hands), one from Part B (the Listening Lab teaches your ears), then Part A to check the concepts. Selected answers appear in the appendix "Answers to Selected Exercises"; exercises marked with a hint include a collapsible nudge — wrestle first, peek second.

Part A — Conceptual

A1. The family interview. For each source below, name the transducer family you'd reach for first, and defend the choice with one physical property of that family (not a vibe — a mechanism): (a) a snare drum two feet from your only mic stand; (b) a whispered vocal in a well-treated closet booth; (c) a harsh, fizzy guitar amp that needs calming; (d) a podcast voice in an untreated home office. One sentence of mechanism each.

A2. Why the coil matters twice. A dynamic mic's moving coil explains both its toughness and its rounded transient response. Walk through both causal chains: mass → SPL tolerance, and mass → softened detail. Then explain why the condenser's design inverts both outcomes at once.

A3. The null audit. Sketch (or describe) the null locations of cardioid, omni, figure-8, and hypercardioid. For each pattern, give one recording scenario where its specific null geometry is the deciding reason to choose it — not its pickup lobe.

A4. Proximity, mechanically. Explain in plain words why directional mics boost bass up close while omnis don't. Your explanation must use the idea of front-versus-back comparison and must account for why the effect grows so quickly in the last few inches.

A5. The prepaid EQ argument. A friend says: "A flat mic is always better — I can add the presence peak with EQ later, so the hyped vocal mic is pointless." Give the strongest honest case for their position, then the strongest honest case against it. (The against-case should involve what happens before and during capture, per Chapter 2's permanence principle, plus the practical economics of small-speaker intelligibility.)

A6. Off-axis detective work. Your budget condenser sounds clean when you speak directly into it but "cheap and boxy" on a full take of acoustic guitar in your bedroom. Using the off-axis coloration sidebar, explain what's actually being heard, why the on-axis test hid it, and list three zero-dollar mitigations in the order you'd try them.

A7. The meters that can't see. A vocalist screams a take; it sounds crunchy on playback, yet the channel meter peaked at -16 dBFS. Explain where in the signal chain (Chapter 3 vocabulary) the distortion happened, why dBFS meters are blind to it, and the two fixes the chapter gives you.

A8. Tradeoff ledger. Complete this table from memory, one honest cost and one honest gain per row:

Part B — 🎧 Listening Lab

Each drill: listen on the best playback you own, then once more on earbuds. Keep your Chapter 4 listening journal open; write what you hear in band vocabulary (sub / lows / low-mids / mids / high-mids / air). Streaming versions are fine for all of these.

B1. Proximity hunting on real records. Cue Billie Eilish, "when the party's over," and listen to the lead vocal's first thirty seconds. The voice is breathy, enormous in the lows for how quiet it is, and feels inches from your ear — that combination is close working distance and proximity warmth doing aesthetic work (the bedroom production story behind this record was Chapter 5's case study). Now cue Johnny Cash, "Hurt": same phenomenon, different emotional job — chest weight, every mouth detail audible. Journal entry: for each track, which band carries the "closeness" cue? What would you predict happens to each vocal if the singer had stood two feet away?

B2. Dynamic vs. condenser, blind-ish. Cue Michael Jackson, "Billie Jean" — by engineer Bruce Swedien's own account, much of Jackson's vocal work in that era was cut on a large dynamic broadcast-style mic. Then cue Billie Eilish again (her early records were tracked on affordable-to-midrange condensers, as Finneas has described in interviews). Don't chase trivia — chase traits: listen to the consonants' leading edges, the amount of "air" above 10 kHz, the way breaths register. Journal: list three audible differences and map each to a transducer property from the chapter (moving mass, diaphragm weight, sensitivity). Then the honest disclaimer to write down: production choices, performance, and processing differ wildly between these records — you're training trait-recognition, not running a controlled experiment.

B3. Presence-peak safari. Pick three vocal-forward tracks: one from the 1970s (try Fleetwood Mac, "Dreams"), one current chart pop song of your choosing, and one of your own favorites. For each, focus only on 2–6 kHz — the presence band. Rank the three by how much energy lives there. Journal: which felt "closer" or more "expensive" at low volume, and does that ranking track the presence ranking? (Chapter 4 warned you how seductive that band is; this drill is also a vaccination.)

B4. Off-axis and bleed in the wild. Cue Nirvana, MTV Unplugged in New York — any song. This is a live recording in a room full of open microphones: every mic is catching every instrument from some angle, mostly off-axis. Listen past the performances to the space: the way the room hangs around the guitar, how the vocal carries a halo of everything else. Journal: describe the difference between this vocal sound and a modern studio pop vocal in terms of direct-to-room ratio and off-axis content — not "worse," different, and chosen.

B5. One mic or two? Width listening. Cue Nick Drake, "Pink Moon" (title track): an acoustic guitar recorded intimately, with a narrow, centered image. Then cue any modern produced folk/pop acoustic track you like and listen to the guitar's width. Journal: is the modern guitar one perspective or two? Can you hear the left and right differing — pick attack on one side, body on the other? Fold to mono if your player allows (or one earbud): does the wide guitar survive? Connect what you hear to the XY/ORTF/spaced tradeoff table.

B6. The Levee re-listen. Return to Led Zeppelin, "When the Levee Breaks," which Chapter 5 gave you as history. Hear it again as this chapter: the drums are far-field placement in an extraordinary space — the room share is the aesthetic. Journal: using the distance-ladder diagram, place this drum sound on the ladder, then name one record in your own genre that lives at the opposite end, and one production reason each choice serves its song.

B7. The podcast distance audit. Queue up three podcasts you already follow — no famous names needed; your own subscriptions are the dataset. For each host's voice, listen for distance discipline: does the low end stay constant phrase to phrase, or does it swell and thin as the host moves (proximity effect riding an unlocked distance)? Does the room appear between sentences — a hollow halo that says the mic is too far — or is the voice dry and close? Journal: rank the three voices by apparent working distance, note which one would benefit most from a pop filter used as a distance-keeper, and write the one-sentence placement note you'd send that host. This is Aisha's entire Chapter 3 problem, heard from the outside — and it's the fastest way to convince yourself that listeners can hear placement, even when they can't name it.

Part C — 🎛️ DAW Lab

Ground rules for every lab: record into the session template you built in Chapter 6, name every take according to your naming convention (gtr_12thfret_8in_take1, not audio_047), and level-match before comparing — Chapter 4 taught you that the louder take wins by default, and that bias does not care about your education. Any mic works: USB, interface mic, phone voice-memo app. If using a phone, treat its mic position (usually bottom edge) as the capsule, record in a voice-memo or field-recorder app, and airdrop/transfer the files into your DAW session afterward — the comparisons happen on playback, so the capture device's pedigree is irrelevant to the lesson. Phones are roughly omni, which actually sharpens two labs: the axis sweep (C2) becomes a control experiment, and the null hunt (C4) becomes a demonstration of what no nulls sounds like.

C1. The distance ladder. Record the same 20 seconds of source — your speaking voice reading a paragraph, or a repeated guitar figure — at four distances: 2 inches, 6 inches, 12 inches, 3 feet. Same axis, same height, same performance energy. Level-match all four in the DAW. Listen blind if you can (shuffle, eyes closed). Deliverable: a four-row journal table — for each distance, the band where the biggest change lived, the room's audibility (none / present / co-starring), and one production context where that exact sound would be the right answer. What you should hear: proximity bass swelling at 2 inches, the room walking into the take somewhere past 12 inches.

C2. The axis sweep. Fix distance at 6–8 inches. Record the same source three times: dead on-axis (0°), about 45° off, about 90° off (speaking past the mic, not into it). Level-match, compare. Deliverable: describe the treble shading from 0° to 90° in band vocabulary, and note what happened to any breathy or spitty edges. Bonus for phone users: phones are roughly omni — document how much less the axis sweep changes, and connect that to the pattern table.

C3. The height experiment. Distance and axis fixed; record your voice three times with the capsule at chest height, lip height, and forehead height (tilt to keep aim consistent). Acoustic instrumentalists: do the guitar version instead — soundhole, 12th fret, behind the bridge (this is case study 1 in miniature). Deliverable: one sentence per position naming which part of the source's recipe the height favored. What you should hear: the source is not a point; height is a recipe dial.

C4. The null hunt. Put a steady noise source somewhere in the room — a fan, or a phone playing rain noise at fixed volume. Keep the mic stationary and rotate it through 360° in 45° steps while recording a few seconds at each step (announce the angle aloud as a marker). Watch your DAW's meters and listen back. Deliverable: a rough polar sketch of your mic — at which rotation did the noise drop the most, and by roughly how many dB on the meter? Does the dead zone match the pattern the mic claims? Cardioid owners: confirm the rear null is real but not a force field, and check whether it's weaker on a bassy noise source than a hissy one.

C5. The one-ear walk, verified. Have your source play (or play a speaker through your phone as a stand-in source) and walk the room with one ear cupped shut. Mark the spot where the source sounds best and the spot where it sounds worst. Record 20 seconds at each spot, same distance/axis/height. Level-match, compare blind. Deliverable: did your ear's prediction survive the recording? Journal what the bad spot added, in band vocabulary. This is the cheapest A/B in the book and the one you'll use forever.

C6. The Theo test. Replicate the chapter's hook with whatever you own. Option one (two mics, even unequal ones — USB mic vs. phone): record the same performance on your better mic placed carelessly (center of the room, 18+ inches, no thought) and your worse mic placed with everything this chapter taught (dead spot behind source, deliberate distance, axis, height). Option two (one mic): careless placement vs. careful placement, same mic. Level-match. Blind-test yourself a day later, or better, blind-test a friend with no stakes in the answer. Deliverable: the verdict, in writing, plus the three biggest audible differences. If careless-expensive wins, interrogate the placements honestly — and write down why. (Either result teaches; only not running the test doesn't.)

C7. Capture-plan dry run. Take one line from the capture plan you'll write in the Project Checkpoint and execute it as a rehearsal: set up the position exactly as planned, record 30 seconds, and grade the result against the "enemy" column — did the plan's rejection strategy actually reject? Deliverable: one adjustment to the plan based on evidence rather than theory. Plans that survive contact with a recording are the only plans that matter.

Part D — Synthesis

D1. Three mics, one living room. A four-piece band — drum kit, bass amp, electric guitar amp, lead vocal — insists on recording live in an ordinary living room, and you own exactly three mics: one cardioid dynamic, one hypercardioid dynamic, one cardioid LDC. Design the capture: which mic on what, every pattern's null aimed where, who gets DI'd instead of mic'd, and which source you deliberately sacrifice to bleed. Justify each decision by rejection, not by capture. (There are several defensible answers; an indefensible one is any answer that never mentions a null.)

D2. The decision tree. Write out, as a text flowchart, the chapter's mic-selection logic — from "what's the source?" through "what's the enemy?" through "where can a null point?" to family, pattern, and the three dials. Maximum one page. Tape it where you record. You've internalized the chapter when you stop needing it; you'll know you're teaching it when you hand someone a copy.

**D3. The $300 memo.** You have $300 and a $99 dynamic mic. Write a six-sentence memo to yourself allocating the money among: a mic upgrade, a stand/pop filter/shock mount kit, and Chapter 10-style room absorption — in any proportion, including $0 lines. Every dollar must be justified by a named concept from this chapter (direct-to-room ratio, off-axis coloration, the three dials drifting, SPL handling...). There is a strong argument that the mic-upgrade line should read $0; if yours doesn't, your memo must beat that argument.

D4. Diagnose the upgrade that downgraded. A friend replaced their $99 dynamic with a $700 LDC and their vocals now sound worse — roomy, boxy, harsher. They're about to spend more. Write the diagnosis (three distinct mechanisms from this chapter that predict exactly this outcome) and the prescription (a placement-and-pattern fix list that costs nothing). End with one sentence about what the $700 mic will reward them with once the room and placement are handled — be honest in both directions.

Part M — Mixed Practice

Interleaving is how the earlier chapters stay load-bearing. No peeking back until you've attempted each.

M1. (Chapter 1.) Explain, strictly in Chapter 1 vocabulary — fundamentals, harmonics, timbre-as-recipe — why pointing a mic at an acoustic guitar's 12th fret versus its soundhole produces two different timbres of the same notes. Then the arithmetic: the guitar's low E string sounds roughly 82 Hz; name the frequencies of its next three harmonics, and say which of the chapter's frequency bands each lands in.

M2. (Chapter 2.) Record two takes of anything: one peaking near -18 dBFS, one deliberately clipping the converter. Zoom into both waveforms. Describe what clipping did to the waveform's shape, explain why no plugin can restore it (what information exists in the file?), and state what 24-bit tracking headroom buys you that "recording hotter for quality" doesn't.

M3. (Chapter 3.) Draw the complete signal chain for a condenser vocal take, capsule to DAW file: every stage, in order, with phantom power entering at the right place and labels for where the signal is mic level versus line level versus digital. Circle the first gain stage and annotate the one thing this chapter taught you can go wrong before it.

M4. (Chapter 4.) Run a single-pass structured listen on one favorite vocal recording: one sentence per band, sub through air, then one sentence on its soundstage depth (close/far). Finish by guessing — explicitly labeled as a guess — the working distance and one placement choice behind it. You now know enough for the guess to be educational either way.

M5. (Chapter 6.) Open your template. Add and configure the tracks your Chapter 7 capture plan requires: correct names per your convention, inputs assigned, color-coded into the right bus lanes (VOX, GTR, and friends), record-armed routing verified with a 5-second scratch take on each. Save the template as a new version with a changelog note. Capture plans live in sessions, not in notebooks.

Selected answers: see the appendix "Answers to Selected Exercises." The Listening Lab has no answer key — it has your journal, which is better.