Chapter 33 — Quiz
Twenty-five questions to check that the audio-post workflow, dialogue cleanup, EQ and leveling, music and licensing, sound design, and loudness mastering have stuck. Answer from memory first; the key with one-line explanations and a "topics to review" map is at the end.
Question types: MC (multiple choice), TF (true/false — justify in one sentence), SA (short answer), and Fix (what is wrong with this mix, and how do you fix it?).
Part 1 — The workflow and the mix
1. (MC) Why must the picture be locked before audio post begins? - A. Because the software will not let you open the audio otherwise. - B. Because every audio fix is keyed to a specific frame, and re-editing slides those alignments out of place. - C. Because color grading has to happen after sound. - D. Because the mix always takes longer than the edit.
2. (MC) In the mix's priority order, which element sits on top, and what follows from that? - A. Music, because it carries the emotion. - B. Ambience, because it sets the place. - C. Dialogue — so any element that makes a word harder to understand is too loud. - D. Sound effects, because they sync to picture.
3. (TF) The correct order of operations is to choose and place your music first, then clean the dialogue around it. (Justify.)
4. (SA) Name the four kinds of stem in a mix and one reason to keep them on separate tracks.
5. (MC) You mix a video on excellent studio headphones and it sounds perfect. What is the essential next check? - A. Nothing — good headphones are the gold standard. - B. Play it on a phone speaker, because that is what most of your audience uses. - C. Re-grade the color to match. - D. Export immediately before you second-guess it.
Part 2 — Dialogue cleanup and noise reduction
6. (MC) Broadband noise reduction works best when you can give it: - A. A louder copy of the dialogue. - B. A "noise print" — a clean sample of the background noise by itself, such as recorded room tone. - C. A high-pass filter set to 1 kHz. - D. The music bed as a reference.
7. (TF) Post-production can fully remove room echo (reverb) from a badly recorded voice. (Justify.)
8. (SA) A voice sounds "underwater" and swirly after your cleanup. What did you do, and what is the fix?
9. (MC) A high-pass filter at ~80 Hz on dialogue mainly removes: - A. Sibilant "s" sounds. - B. The presence range that aids intelligibility. - C. Sub-bass rumble (traffic, HVAC, handling) that carries little voice. - D. Electrical hum at exactly 60 Hz.
10. (SA) Put these cleanup steps in the right order: broadband noise reduction, high-pass filter, spot-fixing clicks/transients, de-hum. Why this order?
11. (Fix) A clip has a steady tonal buzz at 60 Hz and its harmonics running under the whole take. What specific tool addresses this, and why does it cost the voice almost nothing?
Part 3 — EQ and leveling
12. (MC) "Subtractive first" in dialogue EQ means: - A. Boost the highs before anything else. - B. Cut the problem frequencies (mud, harshness) before boosting the good ones. - C. Remove the low end entirely. - D. Always subtract 6 dB from the whole signal.
13. (MC) Which modest boost most improves a voice's intelligibility (its ability to "cut through")? - A. 60-80 Hz - B. 200-400 Hz - C. 3-5 kHz (presence) - D. 16-20 kHz
14. (TF) dBFS and LUFS measure the same thing. (Justify.)
15. (SA) You are leveling dialogue. What peak level do you aim for, and what is the difference between using a compressor and "riding" the levels by hand?
16. (Fix) Dialogue is clean and well-EQ'd, but it lurches from loud to nearly inaudible between shots, making the viewer ride the volume knob. What two techniques fix it, and which is more transparent?
Part 4 — Music, licensing, and the bed
17. (MC) You want to use a popular song you love under your client video. The safe, legal position is: - A. Fine as long as you credit the artist in the description. - B. Fine as long as you are not making money from the video. - C. You need a license; unlicensed use is infringement and platforms will block/mute/demonetize it. - D. Fine as long as you only use fifteen seconds.
18. (SA) What does it mean for a music bed to "duck," and name the two ways to do it.
19. (TF) Under dialogue, a dense, bright, vocal-heavy track is a better bed than a sparse instrumental one. (Justify.)
20. (SA) Name three legal sources of music for a video, and the one habit that keeps you out of copyright trouble.
Part 5 — Sound design, mix, and mastering
21. (MC) Beyond adding realism, a continuous ambience bed also: - A. Raises the true peak of the mix. - B. Replaces the need for dialogue EQ. - C. Hides dialogue edits by keeping the background from dropping to dead silence at cuts. - D. Is only needed for exterior scenes.
22. (MC) You crush your master with a limiter to -8 LUFS to be "loud." A streaming platform that normalizes to about -14 LUFS will: - A. Play it louder than everyone else's. - B. Turn it down to its target — so you get no extra loudness and a flatter, over-compressed mix. - C. Reject the upload. - D. Boost your dynamics automatically.
23. (SA) What two numbers must you watch to master a mix correctly, and what does each control? What is the standard true-peak ceiling?
24. (Fix) A finished mix sounds great on headphones but on a phone speaker the dialogue nearly disappears, and in mono a doubled effect vanishes. Name the likely cause of each and the QC step that would have caught them.
25. (SA) Give the integrated loudness targets for (a) streaming/social video and (b) European broadcast (EBU R128), and explain in one sentence why you master to a target rather than as loud as possible.
Part 6 — Mixed review
26. (MC) Put the audio-post steps in the correct order: - A. Music bed → dialogue cleanup → EQ → master → ambience. - B. Organize/sync → dialogue cleanup → EQ & level → music → SFX & ambience → mix & master. - C. Master → mix → clean dialogue → add music. - D. Whatever is most fun first, usually music.
27. (TF) A high-pass filter at ~80 Hz is the right tool to remove a steady 60 Hz electrical hum. (Justify.)
28. (SA) What is "Foley," and where does it sit in the mix relative to the dialogue?
29. (Fix) A creator's dialogue is clean and level, but under it the music bed is static — it never changes level — and it slightly buries the words in the busiest section. Two problems; name the fixes.
30. (SA) Why does the professional habit say "cut before you boost" in EQ, and why "reduce just enough, then stop" in noise reduction? What single principle unites both?
Answer Key
Reveal answers and explanations
**1. B** — Audio fixes (noise prints, cues on cuts, synced effects) are frame-locked; re-editing the picture moves them and forces re-work. Lock picture first. **2. C** — Dialogue is always the boss of the mix; if anything obscures a word, it is too loud. **3. False** — The order is dialogue-first: organize/sync → clean → level → *then* music, effects, ambience → mix/master. Music placed before clean, level dialogue leads to a soundtrack that fights itself. **4.** Dialogue, music, effects (SFX), ambience. Separate tracks let you control each group's level independently and deliver a music-and-effects (M&E) version for re-voicing/translation. **5. B** — Most viewers listen on a phone speaker, which cannot reproduce low end; a mix must survive it. Also check mono and other systems. **6. B** — Noise reduction subtracts a learned "noise print"; the cleanest source of that print is room tone recorded on set (Ch.15). **7. False** — De-reverb can *reduce* echo but not truly remove it; echo is the voice smeared across time and tangled into itself. This is why close-micing and taming the room on set matter. **8.** You over-processed the noise reduction; it chewed into the voice and created artifacts. Back it off to "just enough" (~8-10 dB) and mask any residual with a slightly louder ambience bed. **9. C** — A high-pass at ~80 Hz rolls off sub-bass rumble (traffic, HVAC, handling) where the voice mostly is not; it is not a hum notch or a sibilance tool. **10.** Order: spot-fix clicks/transients → high-pass → de-hum → broadband noise reduction. Do the surgical and filtering steps first so the broadband step has a smaller, cleaner job and is not confused by isolated events. **11.** A de-hum tool notches out the exact hum frequency (50/60 Hz) and its harmonics; because those notches are very narrow and frequency-specific, they remove the buzz with almost no loss to the broadband voice. **12. B** — Cut problem frequencies before boosting good ones; cutting is cleaner and adds no loudness or harshness. **13. C** — The presence range, ~3-5 kHz, adds consonant clarity/intelligibility. Overdone it becomes harsh (de-ess). **14. False** — dBFS measures instantaneous peak level (clipping at 0 dBFS); LUFS measures perceived loudness averaged over time. You level by peaks and master by loudness. **15.** Aim dialogue to peak around -12 dBFS with headroom below 0 dBFS. A compressor automatically turns down whatever crosses a threshold (evens dynamics continuously); riding by hand means manually drawing level automation up/down on specific moments (more transparent, best for big swings). Use both. **16.** Compression (automatic, evens the overall swing) and riding/automation by hand (drawing gain up on quiet bits, down on loud). Hand-riding is more transparent; a gentle compressor handles the rest. **17. C** — You need a license regardless of credit, profit, or clip length. Automated content-matching enforces it. **18.** Ducking = the music drops a few dB under every line of dialogue and lifts back in the gaps. Two ways: manual volume automation, or a sidechain from the dialogue signal onto a compressor on the music. **19. False** — A sparse instrumental bed leaves the presence range clear for the voice; a dense, bright, vocal-heavy track competes with dialogue for the same space and buries it. **20.** Any three of: royalty-free/production library, Creative Commons (per its exact terms), commissioned original, platform-cleared library. Habit: use only music you can prove you are licensed to use — terms read, receipt kept. (Full treatment: Ch.38.) **21. C** — A continuous ambience bed keeps the background from "blinking" to silence at cuts, hiding edits, in addition to establishing place. **22. B** — Loudness normalization turns your over-loud master down to the house target, so you gain no loudness and keep a flatter, over-compressed mix. Master *to* the target instead. **23.** Integrated loudness in LUFS (set to the destination's target — about -14 for streaming) and true peak in dBFS/dBTP (keep below the ceiling, standard -1 dBFS). Loudness governs how loud it feels; true peak guards against distortion. **24.** Phone dialogue disappears because it was leaning on low-frequency energy the phone speaker cannot reproduce, or was under-EQ'd for presence; the mono effect vanishes because of phase cancellation between doubled/stereo-spread sounds. A four-system QC plus a mono check would have caught both. **25.** (a) Streaming/social ≈ -14 LUFS; (b) EBU R128 broadcast = -23 LUFS. You master to a target because platforms normalize loudness — mastering louder than the target just gets turned down, costing you dynamics for no gain in playback loudness. **26. B** — Organize/sync → dialogue cleanup → EQ & level → music → SFX & ambience → mix & master. Dialogue-first, master last; music is never first. **27. False** — A high-pass at 80 Hz removes broadband sub-bass rumble, not a specific tonal hum. A 60 Hz hum and its harmonics need a de-hum notch; a broad high-pass would miss the harmonics above 80 Hz or gut the voice. **28.** Foley is everyday human sound — footsteps, cloth movement, handling — performed and recorded in sync to picture. It is a kind of spot effect and sits *below* the dialogue in the mix, adding realism without competing with the voice. **29.** (1) The bed is static → add ducking (automation or sidechain) so it lifts in gaps and bows under lines. (2) It buries the words in the busy section → the bed is too loud and/or too dense there; pull its level and, if needed, swap to a sparser track that leaves the presence range (3-5 kHz) clear. Dialogue is the boss. **30.** Both cut/reduce *conservatively* because every processing move trades one thing for another — an EQ boost adds loudness/harshness and stacks energy; noise reduction past "just enough" adds artifacts ("underwater" voice). The uniting principle: do less, and A/B against the raw so the cure never becomes worse than the disease.Topics to review by question
| If you missed… | Review |
|---|---|
| 1, 2, 3, 4, 5, 26 | §33.1 — the workflow, the mix, stems, priority order, honest monitoring |
| 6, 7, 8, 9, 10, 11, 27 | §33.2 — dialogue cleanup, noise reduction, high-pass, de-hum, de-reverb |
| 12, 13, 14, 15, 16, 30 | §33.3 — EQ (subtractive first, presence), leveling, dBFS vs LUFS |
| 17, 18, 19, 20 | §33.4 — music selection, licensing, the ducked bed (and Ch.38 for the law) |
| 21, 28, 29 | §33.5 — ambience, spot effects, Foley, hiding edits, ducking |
| 22, 23, 24, 25 | §33.6 — the mix, mastering, loudness targets, normalization, QC |