Chapter 14 — Exercises

Audio is learned through your ears and your hands, not your eyes. Most of these need only a phone and a quiet room; a few reward a five-dollar lav. Do the Ear Training first — they cost nothing and rewire how you hear video — and take the Record This items seriously, because they build the audio for your Project 1 talking-head.

Exercises are graded by effort and depth:

• ⭐ Warm-up — a few minutes; locks in a concept.
• ⭐⭐ Core — the real practice; do as many as you can.
• ⭐⭐⭐ Stretch — deeper or more ambitious; the biggest gains.

Model solutions and critiques for the starred and odd-numbered exercises are collected in Answers to Selected Exercises at the back of the book. Attempt each one before reading its answer — a recording you diagnosed wrong teaches more than an answer you only read. One rule for everything below: wear headphones. You cannot judge sound on a laptop or phone speaker.

A. Ear Training (no mic required — just listen)

These train the ear the rest of the chapter depends on. Add each to your Frame Log as a listening entry.

1. ⭐ The four-second audio test. Open any short-video feed. For the next ten videos, notice whether the sound makes you stay or leave before the picture does. For three of them, write the specific audio quality that hooked or repelled you — close and clear, or hollow and distant? (Answer provided.)

2. ⭐ Sound-off, sound-on, again. Take a video you like. Watch 30 seconds with sound off, then on. This time, name specifically what the dialogue's closeness or clarity was doing that the picture alone did not tell you. One sentence.

3. ⭐⭐ Room-tone safari. Walk through five rooms of your home or building and listen to each one's silence with your eyes closed for fifteen seconds. Write what each room's room tone actually is — fridge hum, HVAC hiss, traffic, a computer fan, "almost nothing." You are learning that no room is silent. (Answer provided.)

4. ⭐⭐ Spot the mic. For one day, whenever you watch anything with a person talking, guess the mic: lav (hidden, consistent, close), boom (open, natural, a touch of room), handheld (visible stick), or on-camera (distant, roomy). Do it ten times. Where you can see a boom shadow or a lav bump under a shirt, note it.

5. ⭐⭐⭐ Rate the audio, not the picture. Choose three videos of visibly different budgets (a phone vlog, a mid-tier YouTube channel, a broadcast piece). Ignore the image entirely and grade each one's sound 1–5 on: presence (close vs distant), noise (clean vs hissy/roomy), and levels (consistent vs jumpy/clipped). Notice whether "expensive-looking" and "good-sounding" always travel together — often they don't. (Model critique provided.)

6. ⭐⭐ The clipping ear. Find a clip you suspect has distorted audio (loud phone videos of concerts are perfect). Listen for the crackle on the peaks — that is clipping. Then find a clean recording of a similar loud source. Train yourself to hear the difference between "loud" and "clipped," because your meter won't always be in front of you.

B. Record This (field assignments)

The heart of the chapter. Record, listen back on headphones, and diagnose.

7. ⭐ The proximity proof. Record yourself saying the same two sentences twice with your phone: once at arm's length across the room's width, once six inches from your mouth. Change nothing else. Listen on headphones and write the difference in your own words. Keep this clip — it is the whole chapter in ten seconds. (Answer provided.)

8. ⭐⭐ Close, on-axis, out of frame. Record a 20-second piece to camera three ways: (1) camera built-in mic at normal framing distance; (2) a lav (or the phone) close and centered; (3) a mic held just above the top of frame, aimed at your mouth. Rank them and name what closeness and aim each added. (Model critique provided.)

9. ⭐⭐ The lav-rustle hunt. Clip a lav (or tape your phone) to your shirt and deliberately create the three classic noises: rub the mic on a collar, let the cable swing and thump, place it off-center and turn your head. Record each, then fix each (mount away from fabric, tape a strain-relief loop, center it high) and re-record. Hear the fix. (Answer discusses each.)

10. ⭐⭐ Off-axis rejection, felt. If you have any directional mic (even a shotgun on a phone rig, or a cardioid headset), point it at a steady noise source (a fan) and record; then rotate it so the noise is off-axis (to the side or behind, per the pattern) while a voice stays on-axis. Listen to how much the noise drops. If you only have an omni lav, note that it can't do this — and why that's fine when it's close.

11. ⭐⭐⭐ The room-tone recording. Pick a room. After "recording" a few spoken sentences, hold dead still and silent and record 30 seconds of pure room tone. Then, in any editor, lay that room tone under a gap between your sentences. Notice how the gap stops sounding "dead." You have just previewed the single most useful habit in Chapter 15. (Guidance provided.)

12. ⭐⭐ Set a level from scratch. Using your phone or recorder's manual level meter, set the gain so your normal speech peaks around -12 dBFS, then deliberately deliver your loudest likely line and confirm it stays below 0. Now set it too hot on purpose and hear the clip. Write the gain setting that worked. (Answer provided.)

13. ⭐⭐⭐ Mic a second person. Record a short two-person exchange (an interview question and answer) with a single mic. Decide: one lav on the main speaker, or a mic you move/aim between them, or a boom-style mic held close to whoever is talking. Report what was hard about serving two mouths with limited gear — this is the problem Chapter 15's booming solves. (Model critique provided.)

C. Hear the Sequence (critique the SOUND field)

14. ⭐ Re-read FIGURE 14.1. Without looking, write the two SOUND fields from memory — the built-in-mic take and the close-mic take. What single change produced the entire difference? (Answer provided.)

15. ⭐⭐ Write the SOUND field. Watch any 20–30 second scene and write a full Described Shot for its opening shot, but pour your effort into THE SOUND field: name the dialogue's presence, the room tone/ambience, any music or effects, whether it's sync or off-screen, and audio's job in the shot. (Model answer provided.)

16. ⭐⭐ Diagnose the mic from the sound. Listen (eyes closed) to three different talking-head videos and, from the sound alone, argue what mic each used and roughly how far it was. Justify from presence, room, and consistency — not from any visible clue. (Answer provided.)

17. ⭐⭐⭐ Rewrite a failure as a fix. Take FIGURE 14.1 Take A (the roomy built-in-mic version) and rewrite its SOUND and EFFECT fields as if you had fixed it on set — name the exact mic and placement change you'd make and what the new SOUND field would read. (Answer provided.)

D. Five Ways

18. ⭐⭐ One line, five treatments. Record the same single sentence five ways, changing only the audio approach: (a) camera built-in across the room; (b) camera built-in up close; (c) omni lav on the chest; (d) a directional mic aimed from just off-frame; (e) the same directional mic mis-aimed at the wall behind you. Listen back-to-back and rank them. Which surprised you? Write one sentence on why (e) — right mic, wrong aim — can lose to (b). (Model discussion provided.)

19. ⭐⭐⭐ Five rooms, one mic. Record the identical sentence with the identical close mic in five different spaces: a carpeted bedroom, a tiled bathroom, a kitchen, outdoors, and a closet full of clothes. Rank them for cleanliness. Notice that the closet often wins (soft surfaces kill reflections) and the bathroom loses (hard surfaces) — proof that the room is part of your signal. (Answer provided.)

E. Fix the Shot (diagnose bad audio)

20. ⭐ Fix the plan. A friend says: "I'll film my testimonial tomorrow — I'll set the camera up on a tripod across my office and just talk." Name the audio problem baked into that plan and the one-sentence fix. (Answer provided.)

21. ⭐⭐ Fix the recording. A described clip: a subject sits by a lovely window; the picture is gorgeous, but the voice is hollow and distant, the air conditioner is as loud as the words, and one emphatic sentence crackles. List every audio problem and the specific fix for each, sorted by cause (placement / gain / room). (Answer provided.)

22. ⭐⭐ Fix the shotgun. Someone mounts a shotgun on their camera hotshoe to interview a person seated across a small, tiled kitchen and complains it "sounds echoey and far." Explain both reasons it fails and prescribe the fix. (Answer provided.)

23. ⭐⭐⭐ Fix the levels. An editor opens a file where the dialogue peaks sit around -35 dBFS — clean but far too quiet — and another where the peaks are pinned flat against 0 dBFS. For each, say what happens when they try to fix it in post, and which one is salvageable. (Answer provided.)

24. ⭐⭐ Fix the lav. A hidden lav adds a rhythmic scritch on every hand gesture and a dull, drop-out quality when the subject looks left. Name the two separate problems and the two separate fixes. (Model answer provided.)

F. Settings Drill (gain and mic scenarios)

25. ⭐ Read the meter. Sketch a dBFS meter and mark: the clipping point, where dialogue peaks should sit, where room tone should sit, and the "headroom." Label each in one phrase. (Answer provided.)

26. ⭐⭐ Write the audio settings. For each scenario write the mic choice, placement, and a starting gain target, from this chapter only: (a) a seated 60-second talking-head for a website; (b) a walk-and-talk outdoors on a windy day; (c) a person-on-the-street interview at a busy market; (d) a two-person scripted scene shot from six feet back. (Model answers provided.)

27. ⭐⭐ Auto vs manual. Explain, to someone who insists "auto levels are easier," exactly what AGC does in the pauses of speech and why it makes a recording sound worse. Then give the two-word instruction. (Answer provided.)

28. ⭐⭐⭐ Budget the two-mic safety. You're about to shoot an important interview with one lav and one shotgun. Describe how you'd run both at once, what each is for, and which you'd treat as the "safe" track and which as the "preferred" sound. Connect it to why pros do this. (Model answer provided.)

G. Recreate It

29. ⭐⭐⭐ Recreate a podcast's intimacy. Podcasts and ASMR channels get an extremely close, present voice. Study one, then recreate that intimacy with your own mic: get very close, kill the room (soft space), set a clean level, and speak quietly. Report what distance and space it took to match the reference's presence. (Answer discusses what to expect.)

H. Cut This (a first taste of the audio edit)

You don't need advanced skills to feel how audio choices pay off in the edit.

30. ⭐⭐ Smooth a cut with room tone. Take two spoken takes and cut between them. Hear the tiny "hole" at the join. Now lay your recorded room tone (Exercise 11) continuously underneath both. Hear the hole vanish. Write one sentence on why this works. (Guidance provided.)

31. ⭐⭐⭐ Match or fail. Deliberately record two versions of one sentence at different mic distances (close, then far), then try to cut them together as if they were one continuous answer. Hear the presence/room jump at the cut. This is why consistent placement on set matters — write the lesson in your own words. (Answer provided.)

I. Interleaved (bring back earlier chapters)

32. ⭐⭐ Light it and mic it. Set up your Chapter 13 window-lit talking-head and mic it properly in one pass. Confirm the mic (and its shadow, and a boom) stays out of your composed frame (Chapter 6, §6.2) and out of the window light. Note any conflict between where the light wants the subject and where the mic wants to be. (Answer discusses the light-vs-sound tension.)

33. ⭐⭐ Sound for coverage. Recall shooting coverage — a wide and a matching medium/close (Chapter 7, §7.5). If you re-mic between sizes, how do you keep the sound matching so the sizes cut together? (Answer provided.)

34. ⭐⭐⭐ The exposure/audio parallel. In one short paragraph, draw the parallel between judging exposure on a waveform/zebras (Chapter 5, §5.4) and judging level on a dBFS meter — including the one "ceiling" value each has that you must never hit (blown highlights vs 0 dBFS clipping). (Model answer provided.)

35. ⭐ Frame Log — the listening entry. Add today's Frame Log entry as a sound observation: for one video, name the mic, the presence, and one thing the sound did that the picture couldn't. Commit to one listening entry a day for a week. (Answer discusses the habit.)

When you've done these — especially the recording ones, on headphones — you have crossed the line most video makers never cross: you now hear your own work the way an audience does. Take that ear into Chapter 15, where we run real location sound and make clean dialogue a reliable habit rather than a lucky take.