Chapter 21 — Quiz

Closed book first pass; then check yourself. Every answer includes the why — read the explanations even for questions you got right, because the explanation is where the chapter compresses into reflexes. Scoring table at the end.

Section 1 — Multiple Choice (15 questions, 2 points each)

1. "Gain" versus "volume," as this chapter uses the words:

  • A) They're synonyms; usage varies by DAW
  • B) Gain sets what the next stage receives; volume sets what you hear at the output
  • C) Gain is analog, volume is digital
  • D) Gain is measured in dBFS, volume in dB SPL only
Answer **B.** Gain is an input-side decision — it changes what processing *sees* and therefore how it behaves. Volume is output-side — it changes loudness without informing anything upstream. The distinction is the whole reason Theo's fader couldn't fix his amp sim.

2. In a standard DAW channel, the fader sits:

  • A) Before the inserts, so plugins see its changes
  • B) After the inserts, so plugins never see its changes
  • C) Between clip gain and the first insert
  • D) Wherever the first plugin is placed
Answer **B.** Signal order: clip gain → inserts → fader → pan → bus. Pulling the fader down changes what the bus receives, not what any insert receives — which is why a slammed plugin must be fixed *upstream* of itself.

3. The earliest level control you own inside the session — upstream of every insert — is:

  • A) The channel fader
  • B) The master fader
  • C) Clip (region/item) gain
  • D) The first plugin's output trim
Answer **C.** Clip gain adjusts the level of the audio object itself before anything else touches it. It's where staging starts: it feeds the entire insert chain, which is why the clip-gain pass is step 1 of the audit.

4. "-18 dBFS average" became the common staging habit primarily because:

  • A) Audio above -18 dBFS distorts in all DAWs
  • B) Calibration standards mapped analog nominal level (0 VU/+4 dBu) into digital around -20 to -18 dBFS, and analog-modeled plugins are tuned to that neighborhood
  • C) 16-bit audio can't represent levels above -18 dBFS cleanly
  • D) Streaming platforms reject louder files
Answer **B.** It's 0 VU translated into dBFS — broadcast alignment practice pinned the working level around -18 (EBU) or -20 (SMPTE), and plugin developers modeling hardware calibrate sweet spots to the same zone. A convention with reasons and a paper trail — not a law of digital audio (A, C, and D are all false as stated).

5. Your mix delivers the 2-bus with peaks around -6 dBFS, master fader at unity. That headroom exists primarily for:

  • A) Preventing your DAW's float engine from distorting
  • B) The mastering stage, which will spend it deliberately with EQ, glue, and limiting
  • C) Making the mix quieter so it sounds more professional
  • D) Vinyl pressing requirements
Answer **B.** Headroom is a workflow: each stage leaves room for the next stage's work. Mastering (Chapters 31–32) is the stage whose whole job is to spend the mix's margin — a mix printed at the ceiling has already spent a budget that wasn't its to spend.

6. Two tracks both peak at -8 dBFS. One sounds far louder than the other. The best explanation:

  • A) One has a higher sample rate
  • B) Different crest factors — their average levels differ even though peaks match
  • C) Peak meters are broken in most DAWs
  • D) Pan law is reducing one of them
Answer **B.** Loudness tracks the signal's body (average energy), not its tallest sample. A pad and a snare can share a peak while their averages sit 15+ dB apart. This is why staging uses average meters and reserves peak meters for ceiling protection.

7. A VU meter "reads like an ear" because:

  • A) It measures frequency response
  • B) Its slow (~300 ms) ballistics integrate over time, gliding past transients to show the signal's body — roughly how hearing judges loudness
  • C) It displays true peak with oversampling
  • D) It's calibrated in LUFS
Answer **B.** The needle physically can't react to a single drum hit; it settles on the sustained energy. Crude for catching peaks, surprisingly good for estimating loudness — the opposite trade from a peak meter. (LUFS, [Chapter 33](../../part-07-mastering/chapter-33-loudness-wars-streaming/index.md), adds ear-shaped frequency weighting on top of this averaging idea.)

8. Which source typically has the largest crest factor?

  • A) Sustained synth pad
  • B) Heavily distorted rhythm guitar
  • C) Close-miked snare drum
  • D) Organ chord
Answer **C.** Percussion lives almost entirely in its transient: peaks in the high teens of dB above the body. Saturation and sustain flatten crest — the distorted guitar and the held pad/organ have already had their spikes compressed away by their own sound.

9. "Unity gain" means:

  • A) All faders set to the same position
  • B) A stage whose output level equals its input level — it changes nothing
  • C) Maximum gain before clipping
  • D) The gain setting that produces 0 VU
Answer **B.** Unity is the do-no-harm setting — the "0" on your fader, or a plugin trimmed so bypass and active match level. The unity-in/unity-out discipline keeps every insert handing the next a level it expects.

10. The matched-loudness bypass discipline exists because:

  • A) Plugins sound better at high levels
  • B) Louder reliably sounds better to human hearing (equal-loudness effects), so any A/B with a level difference tests volume, not processing
  • C) DAW bypass buttons introduce latency
  • D) It prevents inter-sample peaks
Answer **B.** A plugin adding 2 dB on the way through will win an A/B even if it processes nothing — the level difference flatters lows and highs and reads as "better." Match the levels and your ear finally judges the *change* instead of the loudness. This rule governs every honest comparison from here through mastering.

11. Inside a 32-bit float mix engine, a bus running at +6 dBFS:

  • A) Is permanently distorted
  • B) Is mathematically intact — but still clips your D/A while monitoring, slams modeled plugins, and will shear off if printed to a fixed-point file
  • C) Is automatically limited by the DAW
  • D) Causes aliasing
Answer **B.** Float's enormous internal range preserves the waveform above 0 dBFS — survivable, not free. The hard edges live at the borders: converters, exports, and plugins that respond to level by design. The engine forgives; the surroundings don't.

12. You hear crunchy distortion during playback but every channel and file is clean. The first sixty-second check:

  • A) Reinstall your plugins
  • B) The 2-bus/master meter and monitor path — the D/A may be clipping
  • C) Increase the buffer size
  • D) Bounce the session and listen to the file
Answer **B.** A hot master clips the converter feeding your speakers: distortion that's real to your ears and absent from the session. It's a monitoring-chain failure — it damages decisions, not data — and it's the top of the suspect list whenever you hear crunch you can't locate.

13. Pan law exists because:

  • A) Stereo files are larger than mono files
  • B) A centered signal reproduced by two speakers sums louder than the same signal from one, so mixers compensate by lowering the center
  • C) Hard-panned signals distort more easily
  • D) Vinyl can't reproduce centered bass
Answer **B.** Without compensation, every pan move would lurch the balance. Common laws lower center by 3, 4.5, or 6 dB — and because defaults differ between DAWs, sessions that travel can arrive with shifted centered-versus-panned relationships. ([Chapter 25](../chapter-25-panning-stereo-field/index.md) goes deep.)

14. During the red-light audit, your DRUMS bus peaks at -3 dBFS while other buses sit near -8. The chapter's recommended response:

  • A) Crush the drum channels' dynamics until the bus reads -8
  • B) Pull the master fader down 5 dB
  • C) Trim the bus level (or its inputs) down — and expect drums to run the hottest peaks of any bus, since spiky sources crest high
  • D) Mute the overheads
Answer **C.** Targets are habits, not laws — drums *should* lead the peak race (crest factor), but a bus well above the neighborhood gets trimmed at the bus or its inputs. Option A destroys transients to please a meter; option B hides the reading without fixing the stage.

15. After a full audit, the master fader's correct position is:

  • A) Wherever it stops the red light
  • B) -6 dB, always
  • C) Unity — untouched until mastering, because a healthy 2-bus level is built upstream
  • D) +4 dB to maximize resolution
Answer **C.** If the master fader is doing work, some stage upstream is wrong — it's the smoke alarm, not the fire extinguisher. Staging delivers the 2-bus at a healthy level with the master at unity; loudness becomes a deliberate decision later, in Chapters 32–33.

Section 2 — True/False with Justification (5 questions, 3 points each: 1 for the call, 2 for the why)

16. True or false: If every channel fader is below 0 dB, the 2-bus cannot clip.

Answer **False.** Two reasons. First, faders are the *last* channel stage — hot clip gain and hot plugin outputs ride straight through regardless of fader position. Second, summing: many individually healthy tracks stack the bus far above any one of them (a pile of similar-level tracks grows roughly as 10·log₁₀ of the count). Theo's +3.8 master with all-negative faders was arithmetic, not a glitch.

17. True or false: At 24-bit, staging tracks around -18 dBFS average costs you audible fidelity compared to running them near 0.

Answer **False.** Bit depth sets the noise floor ([Chapter 2](../../part-01-sound-fundamentals/chapter-02-digital-audio/index.md)), and 24-bit's floor is so far down that an 18 dB margin spends nothing audible. There is no fidelity prize for hot levels in a modern session — the headroom is free, which is exactly why the convention survives.

18. True or false: The proper way to set a track's level during staging is to watch its peak meter.

Answer **False.** Stage by averages, protect by peaks. Peak meters know nothing about loudness; crest factor means a snare and a pad at identical peaks sit at wildly different loudness. The average (RMS/VU-style) reading parks the body in the working zone; the peak meter only confirms no ceiling is threatened.

19. True or false: Driving extra level into a saturation plugin is always a gain-staging error.

True or false — Answer **False.** How hard you hit a nonlinear processor *is* part of its sound — drive is a legitimate flavor decision ([Chapter 26](../../part-06-advanced-mixing/chapter-26-saturation-harmonic-color/index.md) lives there). The discipline isn't "never drive"; it's *drive on purpose, then trim the output to unity* so the extra loudness never votes in your A/B and never piles onto the bus. The error is accidental drive, not drive.

20. True or false: Once you've gain-staged the session, the job is done for the life of the mix.

Answer **False.** Gain creep is continuous: every plugin added from [Chapter 22](../chapter-22-eq/index.md) onward is a new gain stage, and small "little louder" decisions accumulate exactly the way they did in Theo's four-year archaeology. The audit's final steps become a thirty-second end-of-session glance, with a full re-run whenever the master meter drifts.

Section 3 — Short Answer (4 questions, 5 points each)

21. Explain the staging hierarchy clip gain → compression → automation for a vocal: which kind of level problem each tool is for, and why feeding a wildly uneven vocal straight into a compressor is asking the wrong tool for help.

Answer Clip gain handles the big, slow moves — section-to-section (even phrase-level) inconsistencies, fixed before any insert sees the signal. Compression ([Chapter 23](../chapter-23-compression/index.md)) handles the fast, moment-to-moment dynamics no human would ride by hand. Automation ([Chapter 27](../../part-06-advanced-mixing/chapter-27-automation/index.md)) handles the *musical* moves — rides that serve phrasing and arrangement. A vocal swinging 12 dB between sections forces the compressor into huge, audible gain reduction to fix what a minute of clip gain solves transparently: the compressor ends up doing a level-matching job violently that an earlier stage does invisibly.

22. Your friend's "stems" arrive peaking at -0.1 dBFS. Name two distinct problems this creates for the person mixing with them, using this chapter's vocabulary.

Answer Any two: (1) No headroom — the moment two of those stems sum, the bus is over the ceiling; the whole session starts life pressed against the wall. (2) Mis-staged processing — every analog-modeled plugin they're fed into sees a level far above its design zone, so thresholds, drive, and meters all behave wrong until everything is trimmed down first. (Also acceptable: any fixed-point print that hot risks true-peak overs, and the receiving engineer must begin with a corrective clip-gain pass — time spent un-doing rather than mixing.)

23. Why do meters get the infrastructure questions and ears get the musical questions? Give one concrete example of each question from a real mix.

Answer Meters are stable instruments; ears adapt to level within minutes ([Chapter 4](../../part-01-sound-fundamentals/chapter-04-listening/index.md)), so ears make poor voltmeters — but meters know nothing about music, so they make worse mixers. Infrastructure (meter) questions: "Is this plugin receiving around -18 average?" "Does the 2-bus have -6 of room?" "Is my A/B level-matched?" Musical (ear) questions: "Is the vocal present enough over the pad?" "Does the snare hit hard enough in the chorus?" The numbers set the stage; the ears make the calls.

24. State the threshold concept of this chapter in your own words, and describe the one-move demonstration that proves it on any session.

Answer Levels are relationships, not absolutes: no level means anything except relative to the floor below it, the ceiling above it, and the other tracks beside it — headroom is a workflow agreement between stages, and faders play intervals, not notes. The demo: select every track and pull them all down 6 dB together. The mix — balance, depth, groove — is unchanged, because everything that mattered was the *distances* between levels, and you preserved every one of them.

Section 4 — Applied Scenario (1 question, 11 points)

25. A friend sends you a session for help. Symptoms: every channel fader between -8 and -15; master fader at -7 "to stop the clipping"; the vocal's compressor pins its gain-reduction meter at the gentlest settings; their tape plugin "sounds like a broken radio"; the bounced MP3 they sent distorts on the choruses even though "the DAW says it's fine"; and they mention everything started sounding better after they normalized all the drums "to make them consistent."

Diagnose in order: (a) the root cause connecting all five symptoms; (b) the specific mechanism behind the pinned compressor and the broken-radio tape plugin; (c) why the bounce distorts even though the float engine "says it's fine"; (d) what normalizing the drums actually did; and (e) the rescue plan, in correct signal order, ending with where the master fader belongs.

Answer **(a) Root cause:** accumulated upstream gain — hot clip levels (the normalize pass) plus uncompensated plugin outputs — with faders and the master dragged down afterward to hide it. Levels were "fixed" at the latest stages instead of the earliest; everything upstream of the faders is still hot. **(b) Mechanism:** inserts see *pre-fader* signal. The compressor receives audio far above its design zone, so even a high threshold sits effectively deep in the signal — huge gain reduction at "gentle" settings. The tape plugin is a nonlinear model: fed 10+ dB hot, it renders its version of slamming the hardware — distortion by design, just nobody's design. **(c) The bounce:** the master fader at -7 may keep the *2-bus* under 0, but if it doesn't fully compensate the overs (or the export path/true peaks exceed the ceiling), the fixed-point MP3 clips where the float engine didn't; and lossy encoding adds its own overshoot on hot material. The engine's forgiveness ends at the border — prints are fixed-point territory. **(d) Normalizing:** maximized each drum clip's *peak*, which — because crest factors differ hit to hit and sample to sample — randomized their *average* levels (the thing loudness follows) while parking everything against the ceiling. The opposite of consistency, and the opposite of staging. **(e) Rescue, in signal order:** 1. Master fader back to unity (it's the smoke alarm, not the fix). 2. Clip-gain pass: every track's body toward ~-18 dBFS average on an RMS/VU-style meter, spiky sources peaking high and free; undo/re-trim the normalized drums by average. 3. Unity pass: every insert output trimmed to matched-loudness bypass; re-set the compressor and tape drive *after* their inputs are sane. 4. Restore the fader balance by ear. 5. Returns and buses trimmed (buses ≈ -8 peaks). 6. Confirm 2-bus ≈ -6 dB peaks, master at unity, zero red lights source to 2-bus; re-bounce and verify the file.

Scoring

Section Points available
1 — Multiple Choice 30
2 — True/False + Justify 15
3 — Short Answer 20
4 — Applied Scenario 11
Total 76
Score Reading
68–76 Staged. Run the audit on your session and move to Chapter 22.
53–67 Solid. Re-read "Gain vs. Volume" and "Peak vs. Average," then re-try the scenario.
38–52 Re-walk the channel-strip map and the headroom ladder, redo DAW Lab C1–C4, retake in two days.
below 38 Re-read the chapter with your own session open — this one only sticks when the meters are yours.