Appendix F: The Ear Training Program

Here's the promise this whole book has been making, finally written as a program you can run: trained ears plus cheap tools beat untrained ears plus expensive tools, every single time, and the gap between those two producers is closed not by money but by minutes — ten to fifteen of them a day, for a while. A spectrum analyzer ships free in your DAW. Reference tracks cost the price of a streaming subscription you already pay. A stock EQ does everything the first six levels here need. The only thing that isn't free is the practice, and the practice is the part that actually matters. That's the entire thesis of the book, restated as a curriculum.

This appendix is the language lab. The chapters are the language course — Chapter 4 taught you the six bands as vocabulary, Chapter 22 taught you to act on them, Chapter 30 taught you to diagnose with them — and every chapter's Listening Lab is the applied half of one of these levels. Reading about EQ makes you fluent in EQ the way reading a phrasebook makes you fluent in French: not at all. You get fluent by doing reps with feedback. That's what this is.

The Rules

Five rules govern everything below. Break them and the program quietly stops working; keep them and the "click" arrives whether you believe in it or not.

1. Ten to fifteen minutes daily beats an hour once a week. This isn't a motivational slogan, it's how perceptual skill consolidates — in frequent, spaced touches, not heroic marathons. A daily ten-minute session you actually keep will outrun a Saturday ear-training binge you abandon by week three, and the binge violates your ear-fatigue hygiene from Chapter 4 anyway. Short and daily. That's the regimen.

2. Log every session. One line is enough: the date, the level, the drill, your score, and one sentence about what tripped you up. An observation you don't write down is an observation you lose. The log is also where you'll see the click happen — a column of scores climbing out of chance is more convincing than any pep talk, and on the bad days it's proof the bad days are temporary.

3. Guess first, then look at the analyzer. This is the single most violated rule in ear training, and violating it wastes the whole exercise. Call the band, the frequency, the direction — out loud or in writing — before you open the analyzer or reveal the answer. The analyzer is the feedback loop, not the perception. Stare at it without guessing first and you train your eyes while your ears stay exactly where they started. Guess, commit, then check. Every time.

4. The click arrives for many people around a few weeks in — keep going through the dip. For a lot of people, somewhere in the first few weeks of daily practice, frequencies stop being abstract numbers and start being places — you hear "that's the 300s stacking up" the way you hear a wrong note. Your mileage will vary; some hear it sooner, some later, and nobody can promise you a date. But almost everyone passes through a discouraging stretch first, usually in the first week or two, where it feels random and pointless and like you have no ears. That stretch is the dip, and the dip is the program working, not failing. The people who quit, quit in the dip. Don't. The log is how you out-stubborn it.

5. Matched loudness or it didn't happen. Louder sounds better — to every ear, including yours, every time. Any A/B where one version is even slightly louder is rigged before it starts, and you'll "hear" the louder one as clearer, fuller, better, when all you heard was louder. Match the levels before you judge, no exceptions. This is the law Chapter 22 built and Chapters 26 and 32 re-ratified, and it governs every drill below that involves a comparison. Most ear-training tools and DAW A/B setups can level-match for you; if yours can't, trim by ear until a fast switch reveals no loudness jump, then judge. Matched-loudness or it didn't happen.

A note on tools before the levels: everything here runs on free software. You need a parametric EQ (every DAW ships one), a spectrum analyzer (likewise), a compressor (likewise), a reverb and a delay (likewise), and a saturation plugin (most DAWs include one; free ones abound). For dedicated ear-training trainers — the kind that quiz you and keep score automatically — free and low-cost options exist; Appendix I points you at current ones, because specific tools come and go and this page outlives them. You can run this entire program with nothing but your DAW's stock plugins and a friend who'll randomize settings for you. That's the point.

The Level Ladder

   ┌──────────────────────────────────────────────────────────────┐
   │  LEVEL 7 ── COLOR & DISTORTION                                │
   │             saturation amount · harmonics vs mud · Ch 26      │
   ├──────────────────────────────────────────────────────────────┤
   │  LEVEL 6 ── SPACE & WIDTH                                     │
   │             reverb type/size/pre-delay · delay vs verb ·     │
   │             mono vs wide · the Haas trick · Ch 24/25          │
   ├──────────────────────────────────────────────────────────────┤
   │  LEVEL 5 ── DYNAMICS                                          │
   │             fast vs slow attack · amount of GR · pumping ·    │
   │             Ch 23                                            │
   ├──────────────────────────────────────────────────────────────┤
   │  LEVEL 4 ── Q & GAIN AMOUNT                                   │
   │             wide vs narrow · how many dB                      │
   ├──────────────────────────────────────────────────────────────┤
   │  LEVEL 3 ── FREQUENCY PRECISION                               │
   │             narrow it: octave, then a third                   │
   ├──────────────────────────────────────────────────────────────┤
   │  LEVEL 2 ── BOOST vs CUT DIRECTION                            │
   │             is the move up or down, roughly where             │
   ├──────────────────────────────────────────────────────────────┤
   │  LEVEL 1 ── BAND ID                                           │
   │             name which of the six bands is boosted            │
   │             ◄── everyone starts here                          │
   └──────────────────────────────────────────────────────────────┘
        climb when you pass the graduation criterion — not before

The level ladder. Each rung is built from the one below it: you can't narrow a frequency (Level 3) until you can name its band (Level 1), and you can't judge how much a thing was boosted (Level 4) until you can tell boost from cut (Level 2). Climb in order. Skipping rungs just means relearning them later with worse habits.

Each level below gives you the same six things: the Goal, The Drill step by step, the Tools, a Self-Test to run on yourself, the Graduation Criterion that tells you you're ready to climb, and an honestly hedged time-to-graduate. The time estimates assume you're keeping Rule 1 — daily reps. They're typical, not promises; some people move faster, some slower, and "slower" is not a verdict on your ears. It's a verdict on how many reps you've done, which is the only thing that's ever true about ear training.

Level 1 — Band ID

Goal. Name which of the six bands from Chapter 4 — sub, lows, low-mids, mids, high-mids, air — is being boosted. First on pink noise, where there's nothing to distract you, then on a full mix, where there's everything to distract you.

The Drill. 1. Put a parametric EQ on a pink-noise source. (Generate pink noise from your DAW's test-tone utility, or grab a free pink-noise file — Appendix I.) Pink noise has roughly equal energy per octave, so a boost in any band is audible and unmasked. This is your training-wheels signal. 2. Set one bell boost, +6 dB, fairly wide (Q around 1), centered in the rough middle of one band: ~40 Hz (sub), ~120 Hz (lows), ~300 Hz (low-mids), ~1 kHz (mids), ~4 kHz (high-mids), ~10 kHz (air). 3. Bypass and engage the EQ a few times. Listen to what the boost does to the character of the noise — sub thickens the floor, low-mids add a chesty "woof," high-mids add an edgy "ssss-ish" rasp, air adds hiss-sheen. 4. Now randomize. Have a friend move the boost to a random band while you look away — or use a trainer that does it for you — and call the band out loud before you reveal anything. Then check. 5. Once pink noise is easy, switch the source to a full mix you know well and repeat. This is much harder: the mix already has energy everywhere, so a boost shifts a balance rather than revealing a frequency in a vacuum. Start with big +6 dB moves; shrink toward +3 dB as you improve.

Tools. Parametric EQ, pink-noise source, reference mix, a randomizer (a friend, or a trainer app).

Self-Test. Twenty randomized boosts — ten on pink noise, ten on a full mix. Score each: right band, adjacent band (off by one neighborhood), or wrong. Log the score.

Graduation Criterion. You can name the boosted band correctly on a full mix at +3 dB, around eight times in ten, on a system you know. "Adjacent" counts as a miss here — adjacent is what Level 1 exists to kill.

Time to graduate. For many people, somewhere in the first couple of weeks of daily reps the bands lock in on pink noise within days, and the full-mix version follows a week or two later. If you're still guessing at chance after a few weeks, you're almost certainly skipping the "out loud, before you look" step — go back to Rule 3.

Level 2 — Boost vs Cut Direction

Goal. Hear not just where the move is but which direction it went — was something boosted or cut — and roughly where. This is the skill that turns "something's off in the low-mids" into "something's missing in the low-mids," which is a completely different fix.

The Drill. 1. Same setup, full mix this time (you graduated pink noise). Set the EQ to a moderate move — ±4 dB, medium Q — in one band. 2. Have it randomized two ways at once: random band and random direction (boost or cut). You look away; the move gets made. 3. Call both before revealing: "cut, in the low-mids." Direction first, because direction is the new skill; the band you already own. 4. The tell to train on: a boost adds a character (more mud, more edge, more air); a cut opens a hole (less mud — the mix breathes; less air — the mix dulls; less low-mid — the mix thins and hollows). Boosts make the mix feel like more; cuts make it feel like cleaner or emptier depending on whether the band was crowded. 5. Bias the practice toward cuts. Cuts are subtler than boosts and they're the move you'll make far more often in real mixing — Chapter 22 is mostly a subtraction chapter. If you can hear cuts, boosts are free.

Tools. Parametric EQ, reference mix, two-axis randomizer (band + direction).

Self-Test. Twenty randomized moves, half boosts and half cuts. Score band-and-direction together — both right, or it's a miss. Log it.

Graduation Criterion. You call direction correctly around eight times in ten on a full mix at ±4 dB, and you still get the band right most of the time while doing it. Cuts and boosts should feel equally legible.

Time to graduate. Often a week or two on top of Level 1 for many people, with the honest caveat that cuts in a crowded band (a low-mid cut in a busy chorus) stay tricky longer than boosts. That's expected — it's also exactly the move that pays the most.

Level 3 — Frequency Precision

Goal. Stop saying "the low-mids" and start saying "around 300" — then "around 250 or 315." Narrow the guess from a band (an octave-plus wide) to an octave, then to a third of an octave. This is the level that makes you fast at the EQ, because you reach for the right knob instead of sweeping.

The Drill. 1. Full mix, single boost, moderate Q, +4 to +6 dB. Randomize the frequency within a band — not just the band. 2. First pass — octave precision. Call the octave the boost lives in. The reference octave centers, in round numbers, are roughly: 60, 125, 250, 500, 1 k, 2 k, 4 k, 8 k, 16 k. Name the nearest one. ("That's the 2 k octave.") 3. Verify with the analyzer or the answer key, then — this is the part that builds the skill — match it yourself: pull up a second EQ instance and try to dial the same boost by ear, A/B against the original at matched loudness, and adjust until they sound identical. You're not just identifying the frequency; you're learning to reproduce it, which is what etches it in. 4. Second pass — third-octave precision. Once octaves are reliable, narrow to thirds: within the 250 octave, is it nearer 200, 250, or 315? This is genuinely hard and it's the edge of what most working engineers can do reliably; don't expect tennis-ball precision, expect "I'm within a third most of the time." 5. Optional graduation accelerator: the "sweep and name" game. Open a track, boost narrow, and slowly sweep the boost while you say the frequencies out loud as it passes through them, watching the analyzer confirm. This calibrates your internal map of where the numbers live.

Tools. Parametric EQ (one to set the boost, one to match it), analyzer for verification, reference mix.

Self-Test. Twenty boosts. Score two ways: octave-correct (nearest of the nine centers) and third-correct (within a third-octave). Log both numbers — watch the octave score climb first, the third score follow.

Graduation Criterion. Octave-correct around eight in ten, and third-correct around six in ten, on a full mix. Six-in-ten on thirds is a real, useful ear — don't hold out for ten.

Time to graduate. This is the level where the few-weeks "click" usually shows up for many people: the octaves snap into place and suddenly frequencies have addresses. Expect a few weeks of daily reps to get octaves solid; thirds keep improving for months, honestly, and that's fine — you graduate on the octave standard and let thirds ripen at the higher levels.

The click is not a finish line — it's a door. Most people who feel it at Level 3 report that the rest of the levels suddenly make sense too, because once frequencies have locations, dynamics and space and color have somewhere to happen. Keep the daily reps going right through the door.

Level 4 — Q & Gain Amount

Goal. Hear the shape and size of a move, not just its location: is the bell wide or narrow (low Q or high Q), and how many dB — a gentle 2, an obvious 4, a heavy-handed 8? This is the level that protects you from your own EQ, because most beginner over-EQing is invisible to a beginner's ears.

The Drill. 1. Two sub-drills, run them separately. Gain first. Fix the frequency and Q, randomize only the gain: ±2, ±4, ±6, ±8 dB. Call the amount. Train the tell that ±2 is "I'm not totally sure anything changed," ±4 is "yes, clearly," ±6 is "that's a lot," ±8 is "that's a move." Matched loudness matters double here — a fat boost raises overall level, and if you don't compensate you'll rate everything as bigger than it is. 2. Q second. Fix frequency and gain, randomize Q from wide (≈0.7, a broad tonal tilt) to surgical (≈8, a narrow spike). Train the tell: wide Q changes the character of a whole region (warmer, brighter, the mix tilts), narrow Q affects a sliver (a single ring gone, a resonance notched, the rest untouched). Boost-and-sweep a high-Q bell versus a low-Q one to feel the difference between "a searchlight" and "a floodlight." 3. Combine: randomize gain and Q together, call both. Then add frequency back in from Level 3 and you're calling the complete move — "narrow cut, about 3 dB, around 3 k." That's a real engineer's sentence. 4. Reverse rep, the most useful one: have a friend make your EQ move worse — push a boost you made from +3 to +7, or widen a surgical cut into a broad one — and you say what they changed and by how much. This is the skill that catches over-processing in the wild.

Tools. Parametric EQ, reference mix, randomizer that controls gain and Q independently.

Self-Test. Two scores, fifteen reps each: gain-amount (within one step) and Q (wide / medium / narrow, correct bucket). Then ten combined reps scored as all-or-nothing. Log all three.

Graduation Criterion. Gain within one step around seven in ten; Q bucket correct around seven in ten; and on the combined reps you can produce a full "shape, amount, location" sentence that's mostly right around half the time. Half on the full sentence is genuinely good — it's three judgments at once.

Time to graduate. A few weeks of daily reps for many people once Level 3 is solid, with gain landing faster than Q (Q is the subtlest perceptual judgment in the EQ levels). If you've got the click from Level 3, this level rides on it.

Level 5 — Dynamics

Goal. Hear compression — not whether a track is compressed (everything is) but how: fast versus slow attack, roughly how much gain reduction, and the specific sound of the move going wrong, which Chapter 23 named the pumping line. Dynamics are invisible until you're trained, and then they're everywhere.

The Drill. 1. Use a drum loop or a full mix and a compressor with visible gain-reduction metering — but cover the meter for the calling, reveal it to check. (Rule 3 applies to compression too.) 2. Attack first, the headline skill. Set a moderate ratio and enough threshold for obvious gain reduction, then randomize attack between fast (≈1 ms) and slow (≈30–50 ms), holding release steady. Call it. The tell, straight from Chapter 23: a fast attack clamps the transient — the drum's click and punch get softened, the loop sounds rounder, flatter, more "held." A slow attack lets the transient through before clamping — the punch survives, even gets emphasized relative to the now-squashed body, so the loop sounds punchier and snappier. Counterintuitive and worth burning in: slow attack = more punch. Beginners always guess backwards. Train until you don't. 3. Amount of gain reduction second. Cover the meter, randomize threshold so GR lands at roughly 2, 4, or 8 dB, and call the amount by sound before revealing the meter. The tell: 2 dB is "gentle leveling, barely there," 4 dB is "clearly controlled, the dynamics are tamed," 8 dB is "squashed, the life is getting pressed out." Then check the meter. 4. The pumping line. Set a fast-ish attack, heavy GR, and a release timed wrong against the tempo, so the compressor audibly recovers between hits — the whole mix breathes, swelling up in the gaps and ducking on the hits. Learn that sound cold, because in the wild it's either a deliberate effect (the EDM pump) or an accident strangling your mix, and you need to hear which. Then sweep the release from too-fast (audible pumping/distortion on the low end) through musical (it "locks" to the groove) to too-slow (the compressor never resets, dynamics gone flat). Call where the pumping starts and stops.

Tools. Compressor with GR metering, drum loop and a full mix, a way to hide the meter for calling, randomizer for attack/threshold/release.

Self-Test. Three scores: attack (fast/slow, fifteen reps), GR amount (within one step, fifteen reps), and a pumping-detection rep — ten passes where you call "pumping / clean," half rigged to pump. Log all three.

Graduation Criterion. Attack direction around eight in ten (this one should get reliable — it's binary and load-bearing), GR amount within a step around seven in ten, and pumping called correctly around eight in ten. The attack score is the real gate; if you can't hear attack, you can't hear compression.

Time to graduate. Honestly, a few weeks to a couple of months for many people — dynamics take longer than EQ for most, because the cues are about time rather than tone and time is harder to hear in isolation. Attack direction often clicks suddenly (one day it just flips and stays flipped); GR amount and release nuance ripen slowly. Be patient here; this is the level that separates people who use compressors from people who hear them.

Level 6 — Space & Width

Goal. Hear the space a sound is sitting in and how wide it is: reverb type and size and pre-delay, the difference between a delay and a reverb, whether something's mono or wide, and the Haas trick that fakes width from a tiny delay (Chapters 24 and 25). This is depth-and-width perception, the dimensions beyond loud-and-bright.

The Drill. 1. Reverb type and size first. Put a reverb on a dry vocal or snare and audition the families from Chapter 24: a tight room (~0.4 s), a plate (~1.8 s, dense and bright), a long hall (lush, slow). Learn the characters — room = "in a space, close walls," plate = "smooth shimmer tail, no obvious early reflections," hall = "big, slow, distant," spring = "boingy, retro." Then randomize and call the type, then the rough size (short/medium/long). The size tell is the tail — how long the wash hangs after the sound stops. 2. Pre-delay second. Fix the reverb, randomize pre-delay between 0 ms and ~60 ms. The tell, exactly as Chapter 24 framed it: long pre-delay separates the dry sound from its reverb — the vocal stays up-front and dry-sounding while the tail blooms behind it a beat later, where zero pre-delay glues the reverb onto the sound so it feels washed and distant. Call short versus long pre-delay. This is the move that keeps a vocal wet and present, so it's worth hearing. 3. Delay vs reverb. Train the distinction cold: a delay produces discrete, repeating copies (you can count them — slap, eighth, dotted-eighth), a reverb produces a continuous wash (you can't count anything, it's a smear of thousands of reflections). Randomize one or the other on a vocal throw and call which. Then the combo case: a tempo-synced delay feeding a reverb, which is most pro vocal space — learn to hear both layers at once. 4. Mono vs wide. Use the mono fold-down from Chapter 4's DAW Lab. Play elements and call mono versus wide before folding to mono to check — the fold-down is your answer key. The tell: wide elements collapse, shift, or partially vanish in mono; true-mono elements are untouched. This is also your mono-compatibility ear, which Chapter 25 cares about a lot. 5. The Haas trick. Take a mono sound, duplicate it, and delay one copy by a tiny amount (5–35 ms) hard-panned opposite — and hear it spread across the stereo field without any actual stereo content. Sweep the delay: under ~5 ms it's comb-filtering tonal weirdness, ~10–30 ms it's lush width, past ~40 ms it starts becoming an audible slap echo. Call the Haas widening when you hear it — and crucially, fold it to mono and hear it partially cancel, which is exactly why Chapter 25 warns you to check.

Tools. Reverb, delay, a mono fold-down (utility or stock plugin — Appendix E), a dry vocal/snare source, randomizer.

Self-Test. Four scores: reverb type (room/plate/hall, fifteen reps), pre-delay (short/long, ten reps), delay-vs-reverb (ten reps), mono-vs-wide (ten reps, fold-down as answer key). Log them.

Graduation Criterion. Reverb type around seven in ten, pre-delay direction around seven in ten, delay-vs-reverb around nine in ten (this one should be easy once you know the "can I count the repeats?" test), and mono-vs-wide around eight in ten. You should be able to hear the Haas trick working and hear it weaken in mono — that pair is the graduation tell.

Time to graduate. A few weeks to a couple of months for many people. Delay-vs-reverb clicks fast; reverb type and pre-delay take longer because the differences are textural and easy to mask in a busy mix. Practice on solo'd sources first, then graduate to hearing space inside a full mix, which is much harder and the real-world version.

Level 7 — Color & Distortion

Goal. Hear saturation — how much, and whether the harmonics it's adding are flattering (the "expensive," "glued," "warm" thing from Chapter 26) or crossing the line into mud and fizz. This is the subtlest level, the one that's last because it's least, and it's the difference between saturation as seasoning and saturation as a mistake.

The Drill. 1. Amount first. Put a saturation plugin on a bass or a drum bus and randomize the drive from "barely on" through "obvious" to "cooking." Call the amount. The tell builds on everything below: light saturation thickens and warms without obviously distorting — the source sounds richer, fuller, a touch louder-feeling even matched (those new harmonics read as "more there"). Heavy saturation audibly distorts — fuzz, grit, the transients smear. Matched loudness is non-negotiable here, because saturation's whole trick is making things feel bigger, and unmatched you'll rate every drive setting as an improvement. This is the level the matched-loudness law was built for. 2. Harmonics vs mud second — the judgment that matters. Saturation adds harmonic content, and where it lands decides whether it helps. Train the two outcomes: useful saturation adds harmonics that help the source translate — the classic case from Chapter 26 is the 808 whose added harmonics in the 700 Hz–1.5 kHz region let it survive a phone speaker that can't reproduce the fundamental. Harmful saturation piles harmonics into the low-mids (the mud zone, 200–400 Hz) and the harsh high-mids, thickening the blanket and adding fizz. So the drill: saturate a bass, then A/B at matched loudness asking one question — did this become more audible and defined, or just thicker and dirtier? Then check on a phone speaker (the real test) and against the analyzer (watch where the new harmonics actually landed). 3. Odd vs even, optionally. Chapter 26's deep cut: even harmonics (tube-ish) tend to read as warm, fat, musical; odd harmonics (transistor-ish, hard clipping) read as edgy, aggressive, buzzy. You don't strictly need to call these to graduate, but if you can hear "warm-fat" versus "edgy-buzzy" you're hearing the flavor of distortion, not just the amount — which is the top of this skill. 4. The integration rep, because this is the last level: take a full mix and run all your levels at once on a single mystery change — a friend alters one thing (a band EQ'd, a compressor's attack flipped, a reverb swapped, a bus saturated) and you diagnose what category it even is first, then the specifics. That's not a drill anymore. That's mixing.

Tools. Saturation plugin, analyzer, a phone/earbud speaker for the translation check, bass and drum-bus sources, randomizer.

Self-Test. Two scores: saturation amount (light/medium/heavy, fifteen reps) and the harmonics-vs-mud judgment (ten reps — "helped translation / just muddied," half rigged each way, phone-checked). Log both.

Graduation Criterion. Amount around seven in ten, and the harmonics-vs-mud call around eight in ten — that judgment is the one that matters more than the precise drive number, because it's the one that protects your low-mids. If you can tell "this saturation earned its place" from "this saturation is mud with extra steps," you've graduated the program.

Time to graduate. A few weeks to a couple of months for many people, riding on the EQ levels (you can't hear "harmonics landed in the mud zone" until Level 3 gave 200–400 Hz an address). Realistically, this level keeps deepening as long as you mix — but you've graduated when the harmonics-vs-mud judgment is reliable, and that's enough to mix with confidence.

How the Levels Map to the Chapters' Listening Labs

This program and the book's chapters are two halves of one thing: this appendix is the drill, each chapter's Listening Lab is the application. The Listening Lab takes the skill you're grinding here in isolation and puts it to work on real music, in context, where it actually has to perform. Run the level alongside the chapters that lean on it, and each makes the other land.

The level-to-Listening-Lab map. The "Built in" chapter teaches the vocabulary; the "Applied in" Listening Labs make you use it on records. Do them together and the drill stops feeling like homework and starts feeling like the reason the chapter made sense.

The deepest integration is Chapter 30's Listening Lab, which is where all seven levels report for duty at once: diagnosing a broken mix means hearing the band (1), the direction it's wrong (2), the frequency (3), how badly (4), whether the dynamics are crushed (5), whether the space is wrong (6), and whether saturation muddied it (7) — and then ranking the fixes. Chapter 30 is the final exam for this entire appendix. When its symptom table runs at the speed of hearing for you, you've finished the course and started the practice.

A Weekly Schedule Template

Here's a template, not a prescription — a shape you bend to your own life. The whole engine is Rule 1: short, daily, kept. Fill in the level you're currently on; the columns just rotate the focus so no single sub-skill plateaus from neglect while another races ahead.

A weekly ear-training template. Total time: roughly an hour and a half across the week, in daily ten-to-fifteen-minute touches — the spacing is the active ingredient, not the total. Bend the days to your schedule; keep the dailiness.

Two notes on the template. First, the Wednesday reference listen and the Saturday application are not breaks from the program — they're where isolated drills become real hearing, and skipping them turns you into someone who aces the trainer and freezes at a real mix. Keep them. Second, the Friday "level below" rep matters more than it looks: skills decay without use, and the point of climbing the ladder is to keep the whole ladder, not just stand on the top rung. A quick maintenance rep on a graduated level costs five minutes and keeps months of work from softening.

The Lifetime Extension: From This Program to the Ten-Year Roadmap

This program graduates you through Level 7 in a matter of months if you keep the daily reps — and that's not the end, it's the first rung of a much longer ladder, the one Chapter 40 lays out as the producer's ten-year roadmap. The seven levels here build what that roadmap calls stage one — symptom vocabulary: you can name what you hear in shareable terms, you know the addresses, you can convert feelings into a fix list. That's further than most producers ever get. It's also just the beginning.

   THIS PROGRAM                          THE TEN-YEAR ARC
   (months, daily reps)                  (Chapter 40's roadmap)

   Levels 1–7 ──────────────────────►  STAGE 1: SYMPTOM VOCABULARY
   band · direction · frequency ·       name what you hear
   Q/gain · dynamics · space ·          ◄── you graduate here
   color
                                        STAGE 2: INSTANT DIAGNOSIS
   keep doing the daily reps,           symptom → cause → stage in
   keep closing feedback loops    ───►  one listen; you hear the
   on real music, keep releasing        PROCESSING itself

                                        STAGE 3: DECISION HEARING
   the references become a board   ───►  you hear WHY — intent,
   of advisors; recreate-then-          architecture, what's MISSING
   deviate (Ch 40) on tracks
   above your level                     STAGE 4: FORWARD HEARING
                                  ───►  arrangements before they
   the practice never ends; the         exist — the finished record
   tool stays the same: target,         audible in the demo, the
   feedback, edge                       part heard before it's played

From the program to the roadmap. The seven levels are stage one of four. The later stages aren't taught — they're accrued, by running this same practice (a target, a feedback check, difficulty at your edge) on real music for years. The tool never changes. Only the depth does.

Here's the honest shape of the arc, hedged because calendars vary wildly and anyone who gives you firm dates is selling something. Year one — roughly where this program lands you — is symptom vocabulary: you hear that something's wrong and you can name it ("the 300s are stacked," "that vocal's too dry for the section"). Around year three, for many people who keep the practice alive, comes instant diagnosis: the symptom table internalizes until it runs at the speed of hearing, and a new layer opens underneath — you start hearing the processing itself, the compressor's release breathing against the tempo, the pre-delay gap behind a vocal, the saturation on a bass you'd once have just called "warm." This is Level 5 and Level 7's skills grown into reflexes. By something like year ten — and this one's the loosest estimate of all, because it depends entirely on how many real reps and real releases you stacked — comes the thing that sounds like myth until you watch someone do it: hearing arrangements before they exist. The producer who hears a voice memo and says "this wants brushes, a high pad, nothing else until the second chorus" and is right isn't psychic. They've compressed a decade of pattern-matching into instinct. They hear the finished record inside the demo, and production becomes excavation.

None of that arrives from this appendix alone. It arrives from running this appendix's method — guess, check, adjust, at the edge of your ability — on real music, across real years, closing the feedback loop with real releases (Chapter 40's fourteen-day rule is loop closure, not hustle). The drills here are the on-ramp. The roadmap is the highway. And the tool never changes: a specific target, immediate feedback, difficulty at your edge. That's deliberate practice, and it's the only thing that has ever turned ears into instruments.

Two honest footnotes to keep on the wall. The stages overlap and regress — you'll have stage-three days and stage-one days, especially in unfamiliar genres, and fatigue sends everyone back a rung by midnight. And the arc has no finish line; nobody graduates hearing. The veteran who's mixed for decades still does the daily reps, still gets surprised, still learns something from a teenager's bounce. That's not the cost of the practice. That's the practice. Stay the student — the student is the one who gets to keep playing.

Print the level ladder. Start tonight, on Level 1, with pink noise and one boosted band you have to name out loud before you look. Ten minutes. Then write down the date and the score, because in a few weeks you're going to want to see how far the line has climbed — and it will have, because the only thing ear training has ever required is the thing nobody can buy for you: the reps, kept.